bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023.

The copyright holder for this preprint

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Neuronal wiring diagram of an adult brain

Sven Dorkenwald1,2, Arie Matsliah1, Amy R Sterling1,3, Philipp Schlegel4,5, Szi-chieh Yu1, Claire E.
McKellar1, Albert Lin1,6, Marta Costa5, Katharina Eichler5, Yijie Yin5, Will Silversmith1, Casey
Schneider-Mizell7, Chris S. Jordan1, Derrick Brittain7, Akhilesh Halageri1, Kai Kuehner1,
Oluwaseun Ogedengbe1, Ryan Morey1, Jay Gager1, Krzysztof Kruk3, Eric Perlman8, Runzhe
Yang1,2, David Deutsch1,9, Doug Bland1, Marissa Sorek1,3, Ran Lu1, Thomas Macrina1,2, Kisuk
Lee1,10, J. Alexander Bae1,11, Shang Mu1, Barak Nehoran1,2, Eric Mitchell1, Sergiy Popovych1,2,
Jingpeng Wu1, Zhen Jia1, Manuel Castro1, Nico Kemnitz1, Dodam Ih1, Alexander Shakeel
Bates4,5,12,13, Nils Eckstein14, Jan Funke14, Forrest Collman7, Davi D. Bock15, Gregory S.X.E.
Jefferis4,5, H. Sebastian Seung1,2*, Mala Murthy1*, the FlyWire Consortium
1
Princeton Neuroscience Institute, Princeton University, Princeton, USA
2
Computer Science Department, Princeton University, Princeton, USA
3
Eyewire, Boston, USA
4
Neurobiology Division, MRC Laboratory of Molecular Biology, Cambridge, UK
5
Drosophila Connectomics Group, Department of Zoology, University of Cambridge, Cambridge, UK
6
Center for the Physics of Biological Function, Princeton University, Princeton, USA
7
Allen Institute for Brain Science, Seattle, USA
8
Yikes LLC, Baltimore, USA
9
Department of Neurobiology, University of Haifa, Haifa, Israel
10
Brain & Cognitive Sciences Department, Massachusetts Institute of Technology, Cambridge, USA
11
Electrical and Computer Engineering Department, Princeton University, Princeton, USA
12
Harvard Medical School, Boston, USA
13
Centre for Neural Circuits and Behaviour, The University of Oxford, Oxford, UK
14
Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, USA
15
Department of Neurological Sciences, University of Vermont, Burlington, USA

*Correspondence to sseung@princeton.edu and mmurthy@princeton.edu

Abstract
Connections between neurons can be mapped by acquiring and analyzing electron
microscopic (EM) brain images. In recent years, this approach has been applied to chunks of
brains to reconstruct local connectivity maps that are highly informative, yet inadequate for
understanding brain function more globally. Here, we present the first neuronal wiring
diagram of a whole adult brain, containing 5×107 chemical synapses between ~130,000
neurons reconstructed from a female Drosophila melanogaster. The resource also
incorporates annotations of cell classes and types, nerves, hemilineages, and predictions of
neurotransmitter identities. Data products are available by download, programmatic access,
and interactive browsing and made interoperable with other fly data resources. We show
how to derive a projectome, a map of projections between regions, from the connectome.
We demonstrate the tracing of synaptic pathways and the analysis of information flow from
inputs (sensory and ascending neurons) to outputs (motor, endocrine, and descending
neurons), across both hemispheres, and between the central brain and the optic lobes.
Tracing from a subset of photoreceptors all the way to descending motor pathways illustrates
how structure can uncover putative circuit mechanisms underlying sensorimotor behaviors.
The technologies and open ecosystem of the FlyWire Consortium set the stage for future
large-scale connectome projects in other species.

1
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Nomenclature
segmentation product of automated pipeline

proofreading the process of correcting errors in the automated segmentation

reconstruction segmented + proofread → the final product

synapse one synaptic link between a pre and a postsynaptic site; a

presynaptic site is usually part of multiple synapses

connection the combination of all synapses between two neurons

Introduction
While rudimentary nervous systems existed in more ancient animals (Arendt, Tosches, and
Marlow 2016), brains evolved perhaps half a billion years ago (Ma et al. 2012), and are
essential for the generation of sophisticated behaviors. It is widely accepted that dividing a
brain into regions is helpful for understanding brain function (Mesulam 1998). Wiring
diagrams at the level of neurons and synapses have been controversial (Sporns, Tononi,
and Kötter 2005; Costandi 2012; Jabr 2012). Skepticism flourished (Sporns, Tononi, and
Kötter 2005; Costandi 2012; Jabr 2012) largely due to a lack of technologies that could
reconstruct such wiring diagrams (Lichtman and Denk 2011; Denk, Briggman, and
Helmstaedter 2012). The situation began to change in the 2000s (Denk and Horstmann
2004; Lichtman and Sanes 2008), due to the efforts of a small community of researchers.
Here we report a significant milestone attained by these efforts, the first neuronal wiring
diagram of a whole adult brain.

The brain of Drosophila melanogaster may seem tiny, but its 105 neurons and 108 synapses
enable a fly to see, smell, hear, walk, and, of course, fly. Flies engage in dynamic social
interactions (Coen et al. 2014), navigate over distances (Fisher 2022), and form long-term
memories (Cognigni, Felsenberg, and Waddell 2018). Portions of fly brains have been
reconstructed from electron microscopic (EM) images, which are sharp enough to reveal the
fine branches of neurons and the synapses that connect them. The resulting wiring diagrams
of neural circuits have provided crucial insights into how the brain generates social
(Schretter et al. 2020; Deutsch et al. 2020), memory-related (C.-H. Li and Yang 2020) or
navigation (Hulse et al. 2020) behaviors. Wiring diagrams of other fly brain regions have
been mapped and related to visual (S.-Y. Takemura et al. 2013; S. Takemura et al. 2017),
auditory (Baker et al. 2022), and olfactory (F. Li et al. 2020; Alexander S. Bates et al. 2020;
P. Schlegel et al. 2021) functions. Similarities with mammalian wiring diagrams (Borst and
Helmstaedter 2015) are striking.

The above wiring diagrams and many others from mammals (MICrONS Consortium et al.
2021; Shapson-Coe et al. 2021; Loomba et al. 2022; Turner et al. 2022; Nguyen et al. 2023)
have come from pieces of brain. But recordings of Drosophila neural activity have revealed
nearly brain-wide encoding of sensory (Pacheco et al. 2021) and motor (Brezovec et al.
2022; Schaffer et al. 2021; Aimon et al. 2022) variables. These studies and others in
vertebrates highlight that understanding how the brain processes sensory information or

2
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 1. A connectomic reconstruction of a whole fly brain. (a) All neuron morphologies
reconstructed with FlyWire. All neurons in the central brain and both optic lobes were segmented and
proofread. Note: image and dataset are mirror inverted relative to the native fly brain. (b) An overview
of many of the FlyWire resources which are being made available. FlyWire leverages existing
resources for EM imagery by Zheng et al. (Zheng et al. 2018), synapse predictions by Buhmann et al.
(Buhmann et al. 2021; Heinrich et al. 2018) and neurotransmitter predictions by Eckstein et al.
(Eckstein et al. 2023). Annotations of the FlyWire dataset such as hemilineages, nerves, and
hierarchical classes are established in our companion paper by Schlegel et al. (c) FlyWire uses CAVE
(in prep) for proofreading, data management, and analysis backend. The data can be accessed
programmatically through the CAVEclient, navis and natverse (Alexander Shakeel Bates et al. 2020),
and through the browser in Codex, Catmaid Spaces (in prep) and braincircuits.io (Gerhard et al., in
prep). Static exports of the data are also available. (d) The Drosophila brain can be divided into
spatially defined regions based on neuropils (Ito et al. 2014) (Ext. Data Fig. 1-1). Neuropils for the
lamina are not shown. (e) Synaptic boutons in the fly brain are often polyadic such that there are
multiple postsynaptic partners per presynaptic bouton. Each link between a pre- and a postsynaptic
location is a synapse. (f) Neuron tracts, trachea, neuropil, cell bodies can be readily identified from the
EM data which was acquired by Zheng et al. (Zheng et al. 2018). Scale bar: 10 μm

drives behavior will require understanding global information flow at the scale of the entire
brain.

The closest antecedent to our whole brain is the reconstruction of a fly “hemibrain” (Scheffer
et al. 2020), a resource that has already become indispensable to Drosophila researchers (F.
Li et al. 2020; P. Schlegel et al. 2021; Klapoetke et al. 2022; Wang et al. 2021). It is
estimated to contain about 20,000 neurons that are “uncropped,” i.e., minimally truncated by
the borders of the imaged volume, and 14 million synapses between them. Our
reconstruction of an entire adult brain contains 127,978 neurons (Fig. 1a), and 53 million

3
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

synapses between them. These and many other data products (Fig. 1b) are available for
download, programmatic access, and interactive browsing and made interoperable with
other fly data resources through a growing ecosystem of software tools (Fig. 1c). The
primary portal to the data is FlyWire Codex (codex.flywire.ai, manuscript in prep), which
makes the information visualizable and queryable.

The wiring diagram from our whole brain reconstruction is complete enough to deserve the
name “connectome.” It is a clear leap beyond C. elegans (300 neurons, <104 synapses)
(White et al. 1986; Varshney et al. 2011; Cook et al. 2019) and the 1st instar larva of
Drosophila (3,000 neurons, 5×105 synapses) (Winding et al. 2023). Our connectome
advances beyond the hemibrain in ways that are not simply numerical. It encompasses the
subesophageal zone (SEZ) of the central brain, important for diverse functions such as
gustation and mechanosensation (see companion paper Shiu et al. (Shiu et al. 2023) as well
as Eichler et al.(Eichler et al. 2023)), and containing many of the processes of neurons that
descend from the brain to the ventral nerve cord to drive motor behaviors. It includes
annotations for nearly all sexually-dimorphic neurons, analyzed in a companion paper
(Deutsch et al. in prep). Our reconstruction of both optic lobes goes far beyond existing
maps of columnar visual circuitry (S. Takemura et al. 2017; Shinomiya et al. 2019, 2022).
Connections between the optic lobes and central brain are included, as explored by a
companion paper (Kind et al., in prep). Also included are neurons that extend into the brain
through the nerves and neck connective, which are essential for tracing sensorimotor
pathways, as illustrated by the present paper and companion papers (list in prep).

Schlegel et al. have compared our wiring diagram with the hemibrain where they overlap and
showed that cell type counts and large strong connections were largely in agreement. This
means that the combined effects of natural variability across individuals and “noise” due to
imperfect reconstruction tend to be modest, so our wiring diagram of a single brain should be
useful for studying any normal Drosophila individual. That being said, there are known
male-female differences (Cachero et al. 2010). In addition, Schlegel et al. report high
variability for principal neurons of the mushroom body, a brain structure required for olfactory
learning and memory. Some mushroom body connectivity patterns have even been found to
be near random (Caron et al. 2013; Murthy, Fiete, and Laurent 2008), though deviations
from randomness have since been identified (Zheng et al. 2022). In short, Drosophila wiring
diagrams are useful because of their stereotypy, yet also open the door to studies of
connectome variation.

Our reconstruction utilized image acquisition and analysis techniques that are distinct from
those used for the hemibrain (Methods and Discussion). However, we have built directly on
the hemibrain in an important way. The companion paper by Schlegel et al. annotated cell
types of central brain neurons, principally by matching them with hemibrain neurons. This
approach was enabled by a growing ecosystem of software tools serving interoperability
between different fly data sources (Fig. 1c). Because annotations of cell types are essential
for scientific discovery, Schlegel et al. (Philipp Schlegel et al. 2023) should be cited along
with the present manuscript by those who use the FlyWire resource. Annotations in the SEZ
and optic lobes, largely absent from the hemibrain, were contributed by Drosophila labs in
the FlyWire Consortium, as well as by citizen scientists. Synapse predictions (Buhmann et
al. 2021; Heinrich et al. 2018) and estimates of neurotransmitter identities (Eckstein et al.
2023) were also contributed by the community.

4
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

In addition to describing the FlyWire resource, this manuscript also presents analyses that
illustrate how the data products can be used. Additional whole-brain network analyses are
provided in a companion paper (Lin et al., in prep). From the connectome with its huge
numbers of neurons and synapses, we derive a projectome, a reduced map of projections
between 78 fly brain regions known as neuropils (Fig. 1d, Ext. Data Fig. 1-1). We trace
synaptic pathways and analyze information flow from the inputs to the outputs of the brain,
across both hemispheres, and between the central brain and the optic lobes. In particular,
the organization of excitation and inhibition in pathways from photoreceptors in the ocelli to
descending motor neurons immediately suggests hypotheses about circuit mechanisms of
behavior.

Results

Reconstruction of a whole fly brain at electron microscopic resolution

Images of an entire adult female fly brain (Fig. 1e, f) were previously acquired by serial
section transmission EM, and released into the public domain by Zheng et al. (Zheng et al.
2018). We previously realigned the EM images (Popovych et al. 2022), automatically
segmented all neurons in the images (Macrina et al. 2021), created a computational system
that allows interactive proofreading of the segmentation, and assembled an online
community known as FlyWire (Dorkenwald, McKellar, et al. 2022). During the initial phase,
much proofreading was done by a distributed community of Drosophila labs in the FlyWire
Consortium, and focused on neurons of interest to these labs. During the later phase, the
remaining neurons were mainly proofread by two centralized teams at Princeton and
Cambridge, with significant contributions from citizen scientists worldwide. The recruitment
and training of proofreaders and their workflows are described in the Methods.

Chemical synapses were automatically detected in the images as pairs of

presynapse-postsynapse locations (Buhmann et al. 2021; Heinrich et al. 2018). The whole
brain contains 0.0188 mm3 of neuropil volume and ~130 million synapses. This works out to
6.9 synapses/µm3, much denser than the <1 synapse/µm3 reported for mammalian cortex
(Schüz and Palm 1989; Dorkenwald et al. 2017). The central brain and left and right optic
lobes contain 0.0103, 0.0042, and 0.0043 mm3 of neuropil volume, respectively, with
synapse counts in approximately the same proportion. Synapses were combined with
proofread neurons to yield the connectome, using the Connectome Annotation Versioning
Engine (CAVE, manuscript in prep).

We already showed that FlyWire proofreading can yield accurate results (Dorkenwald,
McKellar, et al. 2022) through comparison with light microscopic reconstructions of neurons
that are known to be highly stereotyped across individual flies. A second method is to
subject neurons to an additional round of proofreading (J. S. Kim et al. 2014; Scheffer et al.
2020), which was previously shown to yield few changes (Dorkenwald, McKellar, et al.
2022). Because proofreading workflows and personnel have changed over time, and
accuracy can vary across brain regions, we repeated this evaluation by subjecting 826
neurons from the central brain to a second round of proofreading. Relative to the second

5
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

round, our first round of proofreading achieved an average F1-Score of 99.2% by volume
(Ext. Data Fig. 1-2 a,b).

A third validation method is to quantify how many of the automatically detected synapses are
attached to proofread segments, as opposed to being isolated in tiny “orphan” segments
(Buhmann et al. 2021; Heinrich et al. 2018). We found high attachment rates of presynapses
(92.3% or ~120,100,000 presynapses attached) while attachment rates of postsynapses
were lower (43.9% or ~57,200,000 postsynapses attached) due to less proofreading and
reattachment of twigs which contain most of the postsynapses (Dorkenwald, McKellar, et al.
2022) (Ext. Data Fig. 1-2 c,d). Attachment rates were generally in agreement between the
two hemispheres of FlyWire and with the hemibrain (Ext. Data Fig. 1-2 e,f,g) and varied by
neuropil (Ext. Data Fig. 1-3). The bottom line is that accuracy of our connectome is state of
the art. As with the hemibrain (Scheffer et al. 2020), false negative synapses are the
dominant kind of error but false positives exist as well - for this reason all analyses we
present below (and connections indicated in Codex) use a threshold of 5 synapses to
determine a connection between two neurons). Assuming that such errors are statistically
independent, accuracy is expected to be high for detection of connections involving multiple
synapses (Scheffer et al. 2020; Schneider-Mizell et al. 2016; Meinertzhagen 2018).

FlyWire’s reconstruction remains open for proofreading and annotations and new versions of
the resource will be released in future. This allows for the correction of remaining errors as
they are discovered and further rounds of validation to be performed. Additionally, as
explained below, proofreading of photoreceptor axons in the compound eyes is still ongoing.
The first public release (called version 630) has been extensively validated for neurons in the
central brain. All neurons in the optic lobe were proofread but additional validation will likely
identify and correct minor reconstruction errors.

Intrinsic neurons of the brain

Of the 127,978 proofread neurons in FlyWire, 114,423 are fully contained within the brain
(including both central brain and optic lobes, but excluding afferent and efferent neurons,
with projections into and out of the brain, respectively; Fig. 2a,b). These intrinsic neurons
(Fig. 2c left) belong to the brain only, in contrast to other neurons that are shared by the
brain with other structures. Intrinsic neurons of the brain make up three quarters of the adult
fly nervous system (Methods), indicating a high degree of centralization in the brain. The
large fraction is related to the fact that the brain is substantially larger than the ventral nerve
cord (VNC) (Phelps et al. 2021; S.-Y. Takemura et al. 2023; Marin et al. 2023). Intrinsic
neurons amount to 84% of brain neurons. Their predominance means that the brain primarily
communicates with itself, and only secondarily with the outside world.

The nervous system of the larval fly is less centralized; intrinsic neurons of the brain make
up one quarter to one third of its nervous system (Winding et al. 2023). The closest structure
to a brain in C. elegans is the nerve ring (Brittin et al. 2021), which is co-located with multiple
sensory organs in the worm’s head. The nerve ring contains no intrinsic neurons, as all
neurons in the nerve ring also extend neurites into the rest of the nervous system. The
absence of intrinsic neurons is consistent with the convention that the nerve ring is not
commonly called a brain.

6
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 2. Neuron categories. (a) We grouped neurons in the fly brain by “flow”: intrinsic, afferent,
efferent. Each flow class is further divided into “super classes'' based on location and function. Neuron
annotations are described in more detail in our companion paper by (Philipp Schlegel et al. 2023).
The first public release is missing ~8,000 retinula cells in the compound eyes and four eyelets in one
hemisphere which are indicated by hatched bars. (b) Using these neuron annotations, we created an
aggregated synapse graph between the super classes in the fly brain. (c) Renderings of all neurons in
each super class. (d) There are eight nerves into each hemisphere in addition to the ocellar nerve and
the cervical connective (CV). All neurons traversing the nerves have been reconstructed and
accounted for. (e) Sensory neurons can be subdivided by the sensory modality they respond to. In
FlyWire, almost all sensory neurons have been typed by modality. The counts for the medial ocelli
were omitted and are shown in Fig. 7b. (f) Renderings of all non-visual sensory neurons. Scale bar:
100 µm

While the above statistics are based on neuron numbers, they are conceptually related to
volume-based measures of encephalization used in studies of brain evolution (Jerison
1955). For comparison, the rat brain occupies 65% of its central nervous system by volume
(Swanson 1995). Our neuron-based measure of encephalization cannot yet be computed for
rodents, but this will become possible as connectomics continues to scale.

7
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Afferent and efferent neurons

Brain neurons that are not intrinsic can be divided into two categories, depending on the
locations of their cell bodies. For afferent (sensory, ascending) neurons, the cell body is
outside the brain, while for efferent (descending, motor, endocrine) neurons, the cell body is
contained in the brain. Overall, it is generally accurate to think of an afferent neuron as a
brain input, and an efferent neuron as a brain output. The relation to information flow is
actually more subtle, however, as many fly neurites carry some mixture of presynapses and
postsynapses on both dendrites and axons (Eckstein et al. 2023; Winding et al. 2023;
Schneider-Mizell et al. 2016).

Schlegel et al. exhaustively identified all afferent and efferent neurons contained in cross
sections of nerves and the neck connective running between the brain and VNC (Fig. 2d).
Almost 95% of these neurons were in the neck connective, antennal nerve, and
maxillary-labial nerve. Although afferents are truncated in our reconstruction, Schlegel et al.
(Philipp Schlegel et al. 2023) along with other community members (Eichler et al. 2023; H.
Kim et al. 2020) were able to determine the sensory organs corresponding to 5,362 of the
5,495 non-visual sensory neurons (Fig. 2e,f). Non-visual sensory neurons enter the brain
through nerves (Fig. 2d) that mostly terminate in the antennal lobe or the SEZ (we define the
SEZ as containing the following neuropils: SAD, GNG, AMMC, and PRW (Sterne et al.
2021); see Ext. Data Fig. 1-1 for neuropil definitions). The antennal lobe (AL) is the first relay
center for processing of olfactory information, and many of the olfactory receptor neuron
(ORN) inputs to the AL were reconstructed in the hemibrain as well. The SEZ receives more
diverse inputs, including the projections of both mechanoreceptor and gustatory receptor
neurons - these projections were not contained in the hemibrain. The nerves contained few
efferent neurons, among which were head motor neurons (N=100) or endocrine neurons
(N=80) (Fig. 2a,b,c). A large fraction of efferent neurons have branches in the SEZ, including
most of the 100 motor neurons.

Visual afferents are by far the most numerous kind of sensory input, and enter the brain
directly rather than through nerves. This is the last class of neuron that remains to be fully
proofread. There are photoreceptor axons coming from the compound eyes (~12,800 of
which N=3,943 have already been proofread in both eyes), ocelli (N=270), and eyelets (8 of
which N=4 have been proofread).

The neurons traversing the neck connective were grouped into 1,303 efferent (descending)
and 2,364 afferent (ascending) neurons (Fig. 2a,b,c). In a companion paper, Eichler et al. (in
prep) typed these neurons and matched them to reconstructions from two separate EM
datasets of a VNC (Phelps et al. 2021; Azevedo et al. 2022; Cheong et al. 2023).

Optic lobes and central brain

Of the 114,423 intrinsic neurons, 32,422 are fully contained in the central brain, and 73,655
are fully contained in the optic lobes and ocellar ganglia (this number excludes the
photoreceptors, which are sensory afferent neurons, see above). Given that the visual areas
dominate the count, it seems safe to say that Drosophila is a highly visual animal. The optic
lobes, which are largely absent from the first larval instar, are a major reason that the adult
fly brain so dominates its nervous system.

8
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

The optic lobes and ocellar ganglia also contain 7,851 neurons that project into the central
brain, so called visual projection neurons (VPNs). We provide a more detailed analysis of
connections in the ocellar ganglion in Fig. 7. Many VPNs are columnar types that tile the
visual field. VPNs target specific neuropils (e.g., AOTU, PLP, and PVLP) or optic glomeruli
(Wu et al. 2016; Otsuna and Ito 2006) in the central brain. The influence of VPNs can be
very strong; 879 central neurons receive more than half their synapses from VPNs.

The hemibrain already characterized many VPN types along with their outputs in the central
brain (Scheffer et al. 2020). Our whole brain reconstruction reveals many other aspects of
VPN connectivity, such as their inputs in the medulla, lobula, or lobula plate - Schlegel et al.
have typed these neurons based on their axonal projections in the brain and comparing to
reconstructions in the hemibrain (Scheffer et al. 2020). In addition to feedforward targeting of
central neurons, VPNs make 20% of their synapses onto other VPNs, and 21% onto optic
lobe neurons. Companion papers investigate the visual projections to the central complex
(Kim et al, in prep) and the mushroom body (Heckman and Clowney, in prep).

There are 494 neurons that project from the central brain to the optic lobes (Otsuna and Ito
2006). We call these visual centrifugal neurons (VCNs), and they are distinct from previously
defined types of visual centrifugal neurons that are fully contained in the optic lobe, and their
functions are mostly unknown. VCNs are 15× less numerous than VPNs. Nevertheless, half
of all optic lobe neurons receive 5 or more synapses from VCNs, showing that much early
visual processing incorporates feedback from the central brain. Centrifugal inputs to the
retina are found in many vertebrate species, including humans (Repérant et al. 2006).

Many VCNs arborize broadly in the optic lobe, appearing to cover the entire visual field.
Some VCNs, however, cover only a subset of columns within a portion of the visual field. A
few optic lobe neurons receive as many as 50% of their synapses from VCNs. These belong
to the class of peptidergic neurons involved in circadian rhythmicity, which are detailed in a
companion paper (Zandawala et al., in prep). Tm5c is a columnar type (necessary for
Drosophila’s preference for UV over visible light (Karuppudurai et al. 2014)) with more than
10% of its input from VCNs.

A lamina wide-field neuron (Lawf2) can receive more than 10% of its input from VCNs, and a
major input source is octopaminergic (OA-AL2b2). It was previously shown that gain
modulation of Lawf2 neurons increases during flight (Tuthill et al. 2014), and this effect is
mimicked by bath application of octopamine. Transcriptomic studies showed that Lawf2
neurons express octopamine receptors at high levels (Davis et al. 2020).

Neuron super-classes
The neuron classes introduced above are organized into a hierarchy, as explained in our
companion paper (Philipp Schlegel et al. 2023). The three “flow” classes (afferent, intrinsic,
efferent) are divided into nine “super-classes” mentioned above (Fig. 2a). A simplified
representation of the connectome as a graph in which nodes are super-classes is shown in
Fig. 2b. Node sizes reflect neuron number, and link widths indicate connection number. This
is the first of several simplified representations that we will introduce to tame the complexity
of the connectome. Before continuing in this vein, we pause to discuss the properties of
individual neurons and synapses.

9
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 3. Neuron and connection sizes. (a) The synapse-rich (synapses in blue) neuropil is
surrounded by a layer of nuclei (random colors) located at the outside of the brain as well as between
the optic lobes (purple) and the central brain (blue). (b) An LPsP neuron can be divided into
morphologically distinct regions. Synapses (purple and blue) are found on the neuronal twigs and only
rarely on the backbone. (c) We selected seven diverse neurons as a reference for the following
panels. (d) The morphology of a neuron can be reduced to a skeleton from which the path length can
be measured. The histograms show the distribution of path length and volume (the sum of all internal
voxels) for all neurons. The triangles on top of the distributions indicate the measurements of the
neurons in (b). (e) Connections in the fly brain are usually multisynaptic as in this example of neurons
connecting with 71 synapses. (f) The number of connections with a given number of synapses and a
fitted truncated power law distribution. (g) In degree and out degree of intrinsic neurons in the fly brain
are linearly correlated (R=0.76). (h) The number of synapses per neuron varies between neurons by
over a magnitude and the number of incoming and outgoing synapses is linearly correlated (R=0.80).
Only intrinsic neurons were included in this plot. Scale bars: 50 μm (b, c), 10 µm (b-insets)

Neurons and glia

A basic property of the fly brain is that cell bodies are spatially segregated from neurites. Cell
bodies reside near the surface (“rind”) of the brain (Fig. 3a), surrounding a synapse-rich
interior that mainly consists of entangled neurons and glia, fiber bundles or tracts, as well as
tubules of the tracheal system (Fig. 1f, Ext. Data Figure 1-4a, Colodner et al, in prep).

A typical non-sensory Drosophila neuron is unipolar and consists of a primary neurite that
leaves the cell body (soma), enters the neuropil, and branches into secondary and
higher-order neurites (Fig. 3b). Secondary neurites can sometimes be classified as axons if

10
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

presynapses clearly dominate, or dendrites if postsynapses clearly dominate (Eckstein et al.

2023; Winding et al. 2023; Schneider-Mizell et al. 2016). Such an axon-dendrite distinction
was made, for example, when defining visual projection and centrifugal neurons above.

But some mixture of presynapses and postsynapses is generally found on all non-primary
neurites (Fig. 3b). In addition, the soma of insect neurons is separated from the main
processes via a typically thin neural filament (primary neurite, Fig. 3b). Given this structure,
the concept that signals pass from dendrites to soma to axon, which is often a good
approximation for mammalian neurons, does not apply for non-sensory neurons in the fly.

Neurons vary greatly in size and shape (Fig. 3c). We computed skeletons for all
reconstructed neurons (Fig. 3d) to measure neuronal path lengths. The median path length
of a neuronal arbor was 656 µm (Fig. 3d). It has been argued that branched arbors are
optimal for achieving a high degree of connectivity with other neurons (Chklovskii 2004).
Neurons with short path lengths are exceptions, and can be found in both the optic lobes
and central brain. Path length and volume both varied over two orders of magnitude (Fig. 3d,
path length percentiles: 0.1%: 0.059 mm, 99.9%: 19.211 mm, volume percentiles: 0.1%: 80
µm3, 99.9%: 459 µm3). In total, the brain contains ~146 m of neuronal path length.

Sizes vary significantly between different cell superclasses (Ext. Data Fig. 3-1a-f). Optic lobe
neurons are on average much shorter than central brain neurons (0.70 mm vs 2.15 mm on
average) and take up a smaller volume (0.0066 mm3 vs 0.0086 mm3 total neuronal volume),
which is why the optic lobes dominate the brain by neuron number but not by volume or
synapse count. Visual centrifugal neurons are among the largest in the brain, and larger on
average than visual projection neurons (5.05 mm vs 1.56 mm on average). While we
measured much shorter path lengths and volumes for afferent neurons because only part of
their axonal arbors is contained within the brain (Ext. Data Fig. 3-1b,e), arbors of efferents,
motor and descending neurons which also have some of their arbor outside the brain, were
among the largest we measured (Ext. Data Fig. 3-1c,f).

A small fraction of brain volume is glial cells, which are categorized in six types (Kremer et
al. 2017; Yildirim et al. 2019). We estimated that 13% of the cell bodies in the EM dataset
are non-neuronal or glial (Mu et al. 2021). Only a few astrocyte-like glia have been proofread
(Ext. Data Fig. 1-4b). Sheet-like fragments of ensheathing glia are readily found near fiber
bundles in the automated reconstruction. Further proofreading of glia could be prioritized in
the future if there is community demand.

Synapses and connections

Our connectome includes only chemical synapses; the identification of electrical synapses
awaits a future EM dataset with higher resolution (see Discussion). We use the term
“synapse” to mean chemical synapse. A Drosophila synapse is generally polyadic, meaning
that a single presynapse communicates with multiple target postsynapses (Fig. 1e). FlyWire
represents a polyadic synapse as multiple synapses, each of which is a pair of presynaptic
and postsynaptic locations (Buhmann et al. 2021). Polyadic synapses are common in other
invertebrate species, such as C. elegans, and exist in some mammalian brain structures
(e.g. retina).

11
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

We define a connection from neuron A to neuron B as the set of synapses from A to B. A

connection typically contains multiple synapses, and the number can be large (Fig. 3 e,f).
Connections with less than 10 synapses are typical, but a single connection can comprise
>100 synapses (N=14,969) or even >1,000 synapses (N=27). The strongest connection was
from a visual centrifugal neuron (LT39) onto a wide field lobula neuron (mALC2), and
contained over 2300 synapses.

These numbers are much larger than the report of a maximum of 41 synapses connecting a
pair of C. elegans neurons (Cook et al. 2019). To model such a distribution with a long tail,
(Scheffer et al. 2020) used a power law with exponential cutoff (Fig. 3g). Our fit found
comparable parameters, but the fit to our whole-brain distribution of connection strengths
was not as good as their fit to the hemibrain distribution. A similar power law is also a
reasonable fit to the distribution of connection strengths in C. elegans.

Setting a threshold of ≥5 synapses for determining a (strong) connection is likely to be

adequate for avoiding false positives in the dataset, but not missing connections (see
Methods). There are 2,613,129 such connections between the 124,891 identified neurons.
There are several reasons to focus on strong connections. First, a connection with many
synapses is expected to be strong in a physiological sense, other things being equal.
Second, strong connections are likely to be more reproducible across individuals (Hall and
Russell 1991; Witvliet et al. 2021; Philipp Schlegel et al. 2023)(Philipp Schlegel et al.
2023)(Hall and Russell 1991; Witvliet et al. 2021; Philipp Schlegel et al. 2023). Third, higher
accuracy (both precision and recall) of automatic detection is expected for strong
connections, assuming that errors are statistically independent (Schneider-Mizell et al. 2016;
Scheffer et al. 2020).

One of the most basic properties of a node in any network is its degree, the number of nodes
to which it is linked. To characterize the degree distribution in the Drosophila connectome,
we focused on intrinsic neurons (N=114,423) because, unlike afferent and efferent neurons,
they do not suffer from undercounting of connections due to truncation.

For any neuron, in-degree is defined as its number of presynaptic partners (input neurons),
and out-degree is defined as its number of postsynaptic partners (output neurons). The
median in-degree and out-degree of intrinsic neurons are 11 and 13 (Fig. 3g), respectively,
with the restriction mentioned above to connections involving five or more synapses. These
median values do not seem dramatically different from the median in-degree and out-degree
of 10 and 19 for neurons in the C. elegans hermaphrodite, considering that the latter
contains several hundred times fewer neurons than Drosophila.

The neuron in the Drosophila brain with maximum degree is a visual GABAergic interneuron
(CT1), with 6329 postsynaptic partners and 4999 presynaptic partners. CT1 arborizes
exclusively in the medulla neuropil of the optic lobe - indeed, most neuropils of the
Drosophila brain contain one or a few large GABAergic neurons private to that neuropil, with
high in-degree and out-degree (see Lin et al., in prep, for more analysis on connectivity
motifs in FlyWire); these neurons are considered to be important for local feedback gain
control (Prisco et al. 2021; Hong and Wilson 2015). In a C. elegans hermaphrodite (Cook et
al. 2019), the neuron with maximum degree is a command interneuron for backward
locomotion (AVAL), with 110 postsynaptic partners and 64 presynaptic partners. The

12
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

existence of neurons with much higher degree is a marked way in which the Drosophila
connectome differs from that of C. elegans. That being said, the degree of AVAL is large in a
relative sense because it is a large fraction of the total C. elegans neuron number (302).

The number of synapses established by a neuron is correlated with its total neurite path
length (R=0.80 (pre), R=0.89 (post), Ext. Data Fig. 3-1g). Presynapse and postsynapse
counts are similarly correlated per neuron (R=0.80, Fig. 3h). We asked whether large
neurons tend to use their many synapses to create stronger connections with individual
neurons versus more connections with many different neurons. The total number of
synapses established by a neuron was much better correlated with its in and out degrees
(R=0.93, R=0.93 respectively) than its average connection strength (R=0.26, R=0.31
respectively, Ext. Data Fig. 3-1h,i). It remains to be tested whether the additional partners
are from the same or different cell types.

Connections and neurons are not necessarily the functional units of neural computation. For
certain large fly neurons, the arbors are composed of multiple compartments that function
somewhat independently (Meier and Borst 2019; Amin et al. 2020). Perhaps these
subcellular compartments, rather than whole cells, should be regarded as nodes of the
connectome. Then CT1 would be replaced by many nodes with lower degrees. And the
connection from LT39 to mALC2 would be replaced by many connections with fewer
synapses between compartments of these neurons. A connectome of neuronal
compartments can in principle be studied using our resource, which includes the location of
every synapse.

Neurotransmitter identity
A statistical prediction of the small molecule neurotransmitter (GABA, glutamate,
acetylcholine, serotonin, dopamine, and octopamine) secreted by each neuron is available.
A number of validations suggest that the predictions are highly accurate in aggregate
(Eckstein et al. 2023), though for any given synapse the prediction could be wrong. We
assume that every neuron secretes a single small molecule neurotransmitter and combine
the predictions for all outgoing synapses to an estimate which we assign to all outgoing
synapses of a neuron, i.e. we assume neurons obey Dale’s law, although it is known that
co-transmission does occur in the fly brain (Sherer et al. 2020; Mao and Davis 2009;
Waddell et al. 2000; Croset, Treiber, and Waddell 2018).

GABAergic and glutamatergic neurons had much higher degrees than cholinergic neurons
(Ext. Data Fig. 3-1j). Across all neuron categories, we found that GABAergic neurons were
on average longer than glutamatergic and cholinergic neurons (Ext. Data Fig. 3-1k).

As a rule, we will assume that cholinergic neurons are excitatory and GABAergic and
glutamatergic neurons are inhibitory(Molina-Obando et al. 2019; McCarthy et al. 2011; Lu et
al. 2022; Liu and Wilson 2013). A companion paper identifies all GABAergic and
glutamatergic neurons that are bidirectionally coupled with large numbers of cholinergic
neurons (Lin et al., in prep). This reciprocal inhibitory-excitatory motif is widespread
throughout the fly brain (Lin et al. 2014; Scheffer et al. 2020).

13
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 4. Neuropil projections and analysis of crossing neurons. (a) Whole brain
neuropil-neuropil connectivity matrix. The main matrix was generated from intrinsic neurons, and
afferent and efferent neuron classes are shown on the side. Incoming synapses onto afferent neurons
and outgoing synapses from efferent neurons were not considered for this matrix. See Ext. Data Fig.
4-1 for neurotransmitter specific matrices. (b) Cartoon describing the generation of the matrix in (a).
Each neuron’s connectivity is mapped onto synaptic projections between different neuropils. (c)
shows examples from the matrix with each render corresponding to one row or column in the matrix
and (d) shows examples from the matrix with each render corresponding to one square in the matrix.
(e) Most neurons have pre- and postsynaptic locations in less than four neuropils. (f) Renderings
(subset of 3,000 each) and input and output fractions of neurons projecting to (N=11916) and from
(N=7528) the SEZ. The SEZ is roughly composed of five neuropils (the AMMC has a left and right
homologue). Average input and output fractions were computed by summing the row and column
values of the SEZ neuropils in the super class specific projection matrices. (g) Fraction of contralateral
synapses for each central brain neuron. (h) Fraction of ipsilateral, bilateral, contralateral, neurons
projecting to and from the center neuropils per super class. (i) Morphology of ipsilateral, bilateral, and
contralateral neurons from one hemisphere of all intrinsic super classes (up to 3,000 neurons per
plot). Scale bars: 100 µm

14
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

From connectome to projectome

For mammals, tracer injection studies have mapped the axonal projections between brain
regions of mouse (Zingg et al. 2014; Oh et al. 2014; Harris et al. 2019) and macaque
(Felleman and Van Essen 1991; Markov et al. 2014). In fly, large numbers of light
microscopic reconstructions of single neurons have been aggregated to map projections
between brain regions (Meissner et al. 2023; Chiang et al. 2011; Shih et al. 2015). Such
maps have been called projectomes (Kasthuri and Lichtman 2007) or mesoscale
connectomes (Sporns, Tononi, and Kötter 2005). In such techniques, the sampling of axons
is difficult to control, which means that accurate quantification of projection strength is
challenging.

Here we compute a projectome from a synapse-level connectome (Fig. 4a, Ext. Data Fig.
4-1). The interior of the fly brain has been subdivided into hierarchical neuropil regions (Ito et
al. 2014) (Ext. Fig. 1-1, Fig. 1d). Our fly projectome is defined as a map of projections
between these neuropil regions. Because cell bodies are spatially separated from neuropils,
a fly neuron cannot typically be assigned to a single brain region. This is unlike the situation
for a mammalian neuron, which is conventionally assigned to the region containing its cell
body. A typical fly neuron belongs to multiple neuropils.

The projectome is a neuropil-neuropil matrix computed as follows. Each intrinsic neuron

contributes to the projections between neuropils where it has pre- and postsynaptic sites.
We weighted neuron projections by the product of the respective number of synapses and
normalized the result for every neuron such that the matrix sums to the total number of
intrinsic neurons. Each column corresponds to all the neurons projecting to a neuropil and
each row to all neurons projecting out of it (Fig. 4b). Each square then represents the
summed fractional weight of all neurons projecting between two neuropils (Fig. 4c,d). We
added afferent and efferent neurons to the matrix by calculating the sum of the weighted
neuron projections per super class to and from all neuropils respectively.

While each neuropil is connected to many others, most neurons have synaptic sites in only a
few neuropils (Fig. 4e). We repeated this process for each fast neurotransmitter type (Ext.
Fig. 4-1). Some neuropil-neuropil connections exist strongly for one neurotransmitter but not
others. For example, the neuropils making up the central complex (FP, EB, PB, NO) and the
mushroom body (MB-CA, MB-PED, MB-VL, MB-ML) are largely tied together by excitatory
connections.

We observed a strong symmetry between projections in the left and right hemisphere as well
as with the central neuropils located on the midline (Ext. Data Fig. 4-2a,b); this highlights the
strong similarity between the two sides of the brain. We observed that contralateral
projections (projections from one side of the brain to the other) were generally weaker than
projections to the same or ipsilateral neuropil (Ext. Data Fig. 4-2c).

The SEZ (Fig. 4f) is the ventral portion of the central brain, and has been shown to
contribute to a variety of behaviors (Sterne et al. 2021). It is almost wholly unrepresented in
the hemibrain reconstruction (Scheffer et al. 2020), and is also unreconstructed in the larval
brain (Winding et al. 2023). The five neuropils in the SEZ (left and right AMMC, GNG, SAD,
and PRW; Fig. 4f) amount to 17.8% of central brain neuropil volume (0.0018 mm3 of 0.0103

15
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

mm3); they contain afferents mostly from non-visual sensory neurons (mechanosensory and
taste) and ascending neurons, as well as a large number of efferents (motor, endocrine, and
descending neurons - in fact, descending neurons receive on average 69% of their inputs in
one of the five SEZ neuropils). The SEZ is thus important for information flow to and from the
brain. Judging from the projectome (Fig. 4a), the SEZ neuropils interact with almost all parts
of the brain. Notable exceptions are the central complex (EB, FB, PB, and NO) and the
mushroom body (MB), suggesting less crosstalk between those circuits and neurons in the
SEZ (explored in more detail in Fig. 6).

Hemispheric organization
Our reconstruction includes both left and right brain hemispheres. This is important for
tracing sensorimotor pathways that cross from one side to the other, and more generally for
understanding interactions between the two hemispheres. The projectome (Fig. 4a) already
reveals that most projections (88%) are ipsilateral or between neuropils on the same side of
the brain.

The low fraction of non-ipsilateral neurons is primarily due to their scarceness in the optic
lobes. Only 157 neurons (0.2%) in the optic lobes cross hemispheres, and cross the central
brain without making synapses there (Supplemental Information 2) - these neurons are
considered to be “fully contained” in the optic lobes because our definition depends only on
synapse locations. These neurons mediate direct interactions between the two optic lobes,
and their rarity suggests that these interactions represent a smaller fraction of the
computations that occur within the optic lobes. Integration of information from both eyes may
rely more on the abundant crossing connections between the central brain targets (AOTU,
PLP, PVLP) of VPNs.

A higher proportion (40%) of central brain neurons are non-ipsilateral, largely owing to
central neuropils, like those of the central complex and SEZ. To classify non-ipsilateral
neurons, we started by examining the spatial distributions of their postsynapses (inputs). We
divided the neuropils into three categories. Left and Right included the neuropils that come in
mirror-symmetric pairs. Center included the seven remaining neuropils that are located on
the midline. For each neuron, we computed the proportions of its postsynapses in Left,
Right, and Center neuropils (Ext. Fig. 4-3). Each neuron was assigned to the dominant
category, and near-ties were rare. The exceptions are symmetric neurons with cell bodies at
the midline of the brain (Ext. Data Fig. 4-4, N=106).

Next, we asked how many neurons of Left and Right categories have presynapses (outputs)
in the other hemisphere. Similar to the analysis of the 1st instar larval connectome (Winding
et al. 2023), we found that neurons projecting to the other hemisphere can be grouped into
bilateral neurons, those with outputs in both hemispheres, and contralateral neurons which
almost exclusively had presynapses in the other hemisphere (Fig. 4g-i). Notably, many more
visual centrifugal neurons projected to the contralateral hemisphere than visual projection
neurons, and both visual centrifugal neurons and neurons of the central brain contain a large
fraction of bilateral neurons (Fig. 4h) - as stated earlier, this analysis again revealed the
dominance of ipsilateral connections in the brain. While mixing between the hemispheres is
more rare, mixing between sensory modalities within a hemisphere is common (see Fig. 6
below).

16
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 5: Optic lobes. (a) Rendering of a subset of the neurons in the fly brain. A cut through the
optic lobe is highlighted. (b) All 779 Mi1 neurons in the right optic lobe. (c) A single Mi1 neuron, (d) all
neurons crossing through the column in c as defined by a cylinder in the medulla with 1 µm radius
through it, and (e) all neurons sharing a connection with the single Mi1 neuron shown in (c) (≥ 5
synapses) - 3 large neurons (CT1, OA-AL2b2, Dm17) were excluded for the visualization. (f) The two
LPi1-2 neurons in the right lobula plate (neuropil shown in background). Scale bars: 50 µm (b,c,d,e,f),
10 µm (b-inset)

Many types of fly neurons are known to exhibit striking stereotypy across individuals, and
also across both hemispheres of the same individual. A companion paper shows
quantitatively using FlyWire and hemibrain data that these two kinds of stereotypy are similar
in degree (Philipp Schlegel et al. 2023).

Optic lobes: columns and beyond

So far we have mentioned neurons that connect the optic lobes with each other, or with the
central brain. The intricate circuitry within each optic lobe is also included in FlyWire’s
connectome. Photoreceptor axons terminate in the lamina and medulla, neuropils of the
optic lobes (Fig. 5a,b). Each eye contains approximately 800 ommatidia that map to columns
in the lamina arranged in a hexagonal lattice (Fig. 5b). This structure repeats in subsequent
neuropils from lamina to medulla to lobula to lobula plate. The neuropils have been finely
subdivided into layers that are perpendicular to the columns (Fischbach and Dittrich 1989).
The 2D visual field is mapped onto each layer. Any given cell type prefers to synapse in
some subset of the layers. Cell types vary greatly in size. Uni-columnar cell types are the
smallest (Fig. 5b,c). At the other extreme are large cells that span almost all columns (Fig.

17
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

5d). In between there are many multi-columnar cell types that are still being classified (Fig.
5e).

Mi1 is a true “tiling” type, i.e., its arbors cover the visual field with little or no overlap, and
have similar size and shape (Fig. 5b). Dm12 arbors overlap with each other, but the spatial
arrangement is still regular. These and other distal medullary cell types were previously
characterized by multicolor light microscopy (Nern, Pfeiffer, and Rubin 2015). Our EM
reconstructions reveal even more detailed information about the spatial patterning of these
types (e.g., co-fasciculation of neurites of neighboring Dm12 cells). More importantly,
FlyWire’s reconstruction encompasses all multi-columnar cell types, including those outside
the medulla. Judging from the many examples we have studied throughout the optic lobe, it
seems that regular coverage of the visual field without gaps is a defining criterion for most
cell types, similar to mammalian retina (Bae et al. 2018). There are, however, exceptional
cell types that cover the visual field in an irregular manner. For example, there are exactly
two LPi1-2 cells per optic lobe (Shinomiya et al. 2022). The shapes of each pair are
complementary, as if they were created by cutting the visual field into two pieces with a
jigsaw (Fig. 5f); this tiling was not evident when reconstructing only a portion of an optic lobe
(Shinomiya et al. 2022).

Much of the existing research on widefield visual motion processing has relied on the
simplifying idea that the computations are mostly in columnar circuits, and the columnar
outputs are finally integrated by large tangential cells in the lobula plate. This research has
been aided by wiring diagrams containing connections between cells in the same column or
neighboring columns (S.-Y. Takemura et al. 2013; S. Takemura et al. 2017; Shinomiya et al.
2019). An absence of information across columns, has necessitated treating each column as
identical in simulations of the optic lobe (Lappalainen et al. 2023). FlyWire’s connectome
contains not only the columnar neurons (Fig. 5b), but also all neurons that extend across
columns (Fig. 5d,e). These neurons are both excitatory and inhibitory, and can support
interactions between even distant columns. This opens up the possibility of a much richer
understanding of optic lobe computations, and this is explored in a companion paper on hue
selectivity (Christenson et al. in prep).

Some columnar cell types are known to exhibit spatial gradients in connectivity (Dombrovski
et al. 2023), and our reconstruction makes it possible to investigate such gradients for any
columnar cell type in the optic lobe. Similar gradients have also been studied in mammalian
retina (Yu et al. 2018), and such continuous variation is an interesting complement to the
conventional notion that cell types are discrete.

Analysis of information flow

While afferent and efferent neurons make up a numerically small proportion of the brain
(estimated 14.7% and 1.1% respectively), they are important because they connect the brain
to the outside world. Examining connections to these neurons is useful when attempting to
predict the functions of intrinsic neurons from the connectome. For example, one might try to
identify the shortest path in the connectome from an afferent (input) neuron that leads to a
given intrinsic neuron. The sensory modality of the afferent neuron could provide a clue as to
the function of the intrinsic neuron. This approach, while intuitive, ignores connection
strengths and multiplicities of parallel pathways. We therefore use a probabilistic model (P.

18
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 6. Information flow through the Drosophila central brain (a) We applied the information
flow model for connectomes by Schlegel et al. (P. Schlegel et al. 2021) to the connectome of the
central brain neurons. Neurons are traversed probabilistically according to the ratio of incoming
synapses from neurons that are in the traversed. The information flow calculations were seeded with
the afferent classes of neurons (including the sensory categories). (b) We rounded the traversal
distances to assign neurons to layers. For gustatory neurons, we show a subset of the neurons (up to
1,000) that are reached in each layer. (c) For each sensory modality we used the traversal distances
to establish a neuron ranking. Each panel shows the distributions of neurons of each super class
within the sensory modality specific rankings (see Ext. Data Fig. 6-1a for the complete set). (d) We
assign neurons to neurotransmitter types and show their distribution within the traversal rankings
similar to (c). The arrows highlight the sequence of GABA - glutamate peaks found for almost all
sensory modalities (see Ext. Data Fig. 6-1b for the complete set). (e) We UMAP projected the matrix
of traversal distances to obtain a 2d representation of each neuron in the central brain. Neurons from
the same class co-locate (see also Ext. Data Fig. 6-2) (f) Neurons in the UMAP plot are colored by the
rank order in which they are reached from a given seed neuron set. Red neurons are reached earlier
than blue neurons (see Ext. Data Fig. 6-1c for the complete set).

Schlegel et al. 2021) to estimate information flow in the connectome, starting from a set of
seed neurons (Fig. 6a; see Methods).

The likelihood of a neuron being traversed increases with the fraction of inputs from already
traversed neurons and caps out at an input fraction of 30%. We ran the traversal model for

19
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

every subset of afferent neurons as seeds (N=12 input modalities to the central brain, Fig.
6b, Fig. 2e, Supplemental Information 3, see Methods for full list). We then measured
information flow from these starting neurons to all intrinsic and efferent neurons of the
central brain (for this analysis, we ignore circuitry within the optic lobes, and consider VCNs
(visual centrifugal neurons) as efferents of the central brain). We then ranked all neurons by
their traversal distance from each set of starting neurons and normalized the order to
percentiles. For instance, a neuron at the 20th percentile had a lower rank than 80% of
neurons. This allowed us to determine how early information from each afferent modality
reached various targets, including the descending neurons, endocrine neurons, motor
neurons and visual centrifugal neurons (Fig. 6c, Ext. Data Fig. 6-1a). As expected, endocrine
neurons are closest to the gustatory sensory neurons while motor and descending neurons
were reached early for mechanosensory and visual afferents (Ext. Data Fig. 6-1a).

Do the afferent cell classes target inhibitory neurons early or late? We found that putative
inhibitory neurons (neurons predicted to express GABA and glutamate) were
overrepresented in the set of early neurons (Fig. 6d). Surprisingly, we identified a sequence
of GABAergic and glutamatergic peaks in the sequence of neurons targeted that was
replicated for almost all afferent modalities (Ext. Data Fig. 6-1b).

To visualize information flow in a common space, we treated the traversal distances starting
from each seed population as a neuron embedding and built a UMAP projection from all of
these embeddings (Fig. 6e). Within the map, we found that neurons of the same cell class
(e.g. two groups of Kenyon cells, all mushroom body output neurons, all antennal lobe local
neurons, and all central complex neurons) are clustered. Next, we displayed traversal order
on top of the UMAP plot to compare traversal orders starting from different modalities. We
find that every neuron in the central brain can be reached by starting from any modality - this
“small world” property of the network is covered in more detail in a companion paper (Lin et
al., in prep). Comparing orders revealed that almost all neurons in the central brain are
reached early starting from some modality, with the exception of neurons in the central
complex (Fig. 6f, Ext. Data Fig. 6-2), highlighting that the central complex is dominated by
internal computations (Hulse et al. 2020). Kenyon cells were contained in two clusters - one
of which is targeted very early from olfactory receptor neurons and the other targeted early
by visual projection neurons (Vogt et al. 2016).

Our information flow analysis provides a compressed representation of the connectome, but
ignores signs of connections and the biophysics of neurons and synapses, and therefore
terms like “early” and “late” should not be interpreted as true latencies to sensory stimulation.
A companion paper (Shiu et al. 2023) builds a leaky integrate-and-fire model of Drosophila
brain dynamics, using the connectome and including connection weights (number of
synapses) and putative connection signs (excitatory or inhibitory).

Cell types and other annotations

Neurons in Drosophila are considered to be identifiable across hemispheres and individuals
(Luan et al. 2006; Pfeiffer et al. 2010), enabling cell type classification of all neurons in
FlyWire - such classification is useful for generating testable hypotheses about circuit
function from the connectome. FlyWire community members, experts in diverse regions of
the fly brain, have shared 91,649 annotations of 59,548 neurons (Supplemental Information

20
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

4), including the majority of sexually-dimorphic neurons (Deutsch et al., in prep), sensory
neurons (Eichler et al. 2023), as well as a diversity of cell types in the optic lobes and SEZ
(Fig. 2f), two brain regions not covered in the hemibrain connectome. Each neuron in
FlyWire is given a unique name based on the neuropil it receives and sends most of its
synapses. Curation of these annotations continues, and we invite further community efforts
to identify cell types, and these can be contributed through Codex (codex.flywire.ai).

In addition, matching between cell types identified in the hemibrain (Scheffer et al. 2020) and
both hemispheres of FlyWire provides additional annotations for neurons contained in both
datasets. Our companion paper (Philipp Schlegel et al. 2023) provides cell type annotations
for 26,150 neurons via such matching.

All cell annotations can be queried in Codex. Some of these have already been mentioned,
such as the “flow” annotations of intrinsic vs. afferent vs. efferent, super-class annotations of
Fig. 2, neurotransmitter predictions, left-right annotations for cell body location, in addition to
lineages, or groups of neurons derived from a single neuroblast (Schlegel et al. 2023).

Ocellar circuit structure and function: linking sensory inputs to motor outputs
The completeness of the FlyWire connectome enables tracing complete pathways from
sensory inputs to motor outputs - here we demonstrate this capability by examining circuits
that emanate from the ocellar ganglion and leveraging cell type information. In addition to the
large compound eyes, flying insects have smaller visual sensory organs (Hofbauer and
Buchner 1989), including the three ocelli on the dorsal surface of the head cuticle (Fig. 7a).
The ocelli are under-focused eyes, projecting a blurry image of light level changes in the UV
and blue color spectrum (Hu, Reichert, and Stark 1978; Stark, Sapp, and Carlson 1989);
these eyes are thought to be useful for flight control and orientation relative to the horizon
(Stange et al. 2002). Importantly, while the role of the ocelli has been hypothesized (e.g.,
light level differences between the eyes when the fly is shifted off axis should quickly drive
righting motions of the head, wings, and body to stabilize gaze and re-orient the body), little
is known about the circuitry downstream of this sensory organ that would mediate this
function.

Photoreceptor axons (N=270) from the three ocelli innervate three distinct regions of the
ocellar ganglion separated by glial sheets (Fig. 7a, b). The ocellar ganglion additionally
contains 62 neurons that we categorized into four broad groups (Fig. 7c, Ext. Data Fig.
7-1a): local neurons (N=15), two types of interneurons, divided based on their arborizations
and caliber (OCG01 (N=12), OCG02 (N=8)), descending neurons (DNp28, N=2), and
centrifugal or feedback neurons (N=25). Ocellar local neurons are small (116 outgoing
synapses, 449 µm path length on average) and connect sparsely with photoreceptors from
all ocelli.

Twelve OCG01 interneurons and two descending neurons (DNp28, one per lateral ocellus)
represent the main pathway from the ocellar ganglion to the central brain. DNp28 projects to
the intermediate, haltere, wing, and neck tectula of the ventral nerve cord (Cheong et al.
2023; S.-Y. Takemura et al. 2023). In each ocellus, half of the OCG01s were inferred to
express glutamate (likely inhibitory), and the other half acetylcholine (excitatory). There are
four OCG01s per ocellus (Fig. 7d). OCG01s tile the ocellar ganglion, indicating their

21
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Figure 7. Ocellar circuits and their integration with visual projection neurons. (a) Overview of
the three ocelli (left, medial, right) which are positioned on the top of the head. Photoreceptors from
each ocellus project to a specific subregion of the ocellar ganglion which are separated by glia
(marked with black lines on the EM). (b) Renderings of the axons of the photoreceptors and their
counts, and (c) OCG01, OCG02 and DNp28 neurons with arbors. “Information flow” from pre- and
postsynapses is indicated by arrows along the arbors. (d) Connectivity matrix of connections between
photoreceptors and ocellar projection neurons, including two descending neurons (DNp28). (e)
Comparison of number of glutamatergic and cholinergic synapses from ocellar projection neurons
onto downstream neurons colored by super class (R=.65, p<1e-21). (f) Summary of the observed
connectivity between ocellar projection neurons, visual projection neurons and descending neurons.
Scale bar: 100 µm

receptive fields tile the visual fields of the ocelli (Ext. Data Fig. 7-1 b,c). OCG02 axons are
much thinner than the OCG01s, and likely transmit signals slower. Two OCG02 subgroups
(a, b) innervate similar neuropils to the OCG01s (IPS, SPS), and OCG02c neurons target
the PLP, a brain region that also receives input from visual projection neurons from the
compound eyes (Wu et al. 2016).

Neurons downstream from OCG01s in the IPS, SPS, and GNG receive inhibitory input from
the ipsilateral ocellus and excitatory input from the contralateral ocellus (Fig. 7d, right), and
the amount of synaptic input from each ocellus is tightly correlated (Fig. 7e, R=0.65,
p<1e-21) - this balance is likely to be a key ingredient in how signals are integrated (the
descending circuits are activated by a signal difference between the eyes). We found that 15
different descending neurons (DNs) each receive over 200 synapses from the OCG01
neurons. For example, two DNs in each hemisphere received over 30% of their synaptic
inputs in the brain from ocellar projection neurons: DNp20/DNOVS1 (left: 57%, right: 44%),
DNp22/DNOVS2 (left: 36%, right: 33%). DNOVS1 and other descending neurons with strong
input from OCG01s generally receive strong input from ipsilateral visual projection neurons
as well (Ext. Data Fig. 7-1d). For example, DNOVS1 is also activated by rotational optic flow
fields across the compound eye, and projects to the neck motor system (Suver et al. 2016;
Haag, Wertz, and Borst 2007). A handful of glutamatergic (putative inhibitory) visual

22
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

projection neurons sparsely innervate descending neurons in both hemispheres. As the

ocelli transmit mainly information about light levels, the dense integration with motion
direction signals from the compound eyes was not previously appreciated, but should aid in
precision adjustments of head and body movements for gaze stabilization and flight control
(A. J. Kim et al. 2017).

There is also extensive feedback from the brain directly to the ocellar ganglion via 25 ocellar
centrifugal neurons (OCC). We found striking targeting specificity of two OCC subgroups
(OCC01a, b) which synapse onto all OCG01 and DNp28 neurons with strong connections
compared with their overall synaptic budget (Ext. Data Fig. 7-1e). The OCC01s receive input
in a wide range of neuropils, notably the SEZ, as well as IPS and SPS, the same neuropils
that receive inputs from the OCG projection neurons (Ext. Data Fig. 7-1f). It remains to be
determined what role the OCCs play in gating visual information and potentially driving the
OCGs in the absence of photoreceptor activity.

Based on the summary wiring diagram of Fig. 7f, we hypothesize how the pathways from the
ocelli to descending neurons function. As in a Braitenberg vehicle for phototaxis (Braitenberg
1984), excitation and inhibition are organized so that the head and body of the fly should roll
around the anteroposterior axis to orient the ocelli towards light. In this example, the
whole-brain connectome, extending from brain inputs to outputs, uncovers new pathways
and facilitates the generation of putative circuit mechanisms of sensorimotor behavior.

Discussion
By reconstructing a complete brain wiring diagram, FlyWire enables many kinds of studies
that were not previously possible using wiring diagrams of portions of the fly brain. The optic
lobes and the SEZ are two prominent regions mostly missing from the hemibrain, the
previous state of the art. Both sides of the brain are included, enabling the tracing of
pathways that cross the midline. Due to the presence of afferent and efferent neurons, one
can trace pathways from sensory inputs to intrinsic neurons to brain outputs (motor,
endocrine, and descending neurons). This was done in a global fashion using the
information flow model, and more specifically to uncover the structure and hypothesize a
circuit mechanism for behaviors supported by the ocelli. Our companion papers provide
additional global analyses of the connectome (Lin et al., in prep) and studies of specific
families of pathways.

Connectome annotation
Connectome annotation with structural and functional information is an important emerging
field, analogous to genome annotation. Annotations are important because they make the
connectome usable for hypothesis generation about circuit function. We carried out a
hierarchical and systematic annotation of all neurons in the connectome as detailed in our
companion paper (Philipp Schlegel et al. 2023), describing over 4000 robustly identifiable
cell types. We also collected a large number of annotations from the community (57% of all
neurons have an annotation label) leveraging a broad knowledge base - further curation of
these labels will help to refine them.

23
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Comparative connectomics
For the first time, one can now compare entire connectomes of different species, starting
with Drosophila melanogaster and C. elegans, as touched on by the present manuscript, and
explored in more depth by Lin et al (Lin et al., in prep). One can also compare connectomes
of the same species at different developmental stages (Winding et al. 2023). While ours is
still the only adult fly connectome, it can be compared with the hemibrain reconstruction
where they overlap, to detect wiring differences between adults of the same species (Philipp
Schlegel et al. 2023).

Connectomes, transcriptomes, and brain development

Transcriptomics with single cell resolution is being applied to mammalian brains (BRAIN
Initiative Cell Census Network (BICCN) 2021), and to the Drosophila brain as well.
Transcriptomic atlases of the central adult brain (Croset, Treiber, and Waddell 2018; Davie et
al. 2018) and optic lobes (Kurmangaliyev et al. 2020; Özel et al. 2021) are appearing.
Comparing connectomes with transcriptomes is already proving useful for studying
molecular mechanisms of development (Kovács, Barabási, and Barabási 2020; Yoo et al.
2023; Alexander Shakeel Bates et al. 2019). Clearly more fly connectomes at multiple
developmental stages are needed.

Brain simulation
Connectome-based brain simulation was one of the original motivations for connectomics
(Seung 2012). A neural network simulation of visual motion detection based on the wiring
diagram of columnar circuits in the optic lobe has been created (Lappalainen et al. 2023).
Such a connectome-based approach can at last be scaled up to an entire brain, as shown by
Shiu et al. (Shiu et al. 2023).

Block face versus serial section EM

The hemibrain was reconstructed (Scheffer et al. 2020) from images acquired by FIB-SEM
(Knott et al. 2008; Xu et al. 2017; Hayworth et al. 2020), a form of block face EM (Denk and
Horstmann 2004). FlyWire is based on transmission EM images of serial sections (ssTEM)
that were manually cut and collected (Zheng et al. 2018), an evolution of the approach that
was used for the C. elegans connectome (White et al. 1986). In the end, both block face and
serial section EM have turned out to be viable for fly connectomes. Both approaches yield
similar accuracy (Ext. Fig. 1-2, Ext. Fig. 1-3). Hybrid methods that combine both imaging
approaches are also being developed (Hayworth et al. 2020).

Artificial and human intelligence

Owing to the use of artificial intelligence (AI), the hemibrain and FlyWire have yielded
connectomes that are orders of magnitude larger than those of C. elegans (Cook et al. 2019)
or the larval fly (Winding et al. 2023). The hemibrain images were automatically segmented
using flood-filling convolutional nets (Januszewski et al. 2018), whereas FlyWire used the
older approach of boundary-detecting convolutional nets (Jain et al. 2007; Turaga et al.
2010). FlyWire also required another kind of AI, alignment of serial section images using
convolutional nets (Popovych et al. 2022). The improved alignment was crucial for making
ssTEM as amenable as block face to automated reconstruction (Lee et al. 2019). In spite of

24
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

enormous progress in AI, both the hemibrain (Scheffer et al. 2020) and FlyWire (Methods)
required an estimated 50 and 30 person-years of human effort for proofreading the
automated segmentation respectively (see Methods). This is because AI has reduced the
amount of human labor required per unit brain volume, but EM image volumes have
increased even faster. Further reduction in human proofreading is necessary for
reconstructing many fly connectomes to study variation, or to scale up to whole mammalian
brains.

Imaging smaller
The EM images used by FlyWire were acquired at a resolution of 4×4×40 nm3. Sharpening
this resolution would presumably enable accurate attachment of twigs to backbones, which
is currently the main factor limiting the accuracy of reconstructing synaptic connectivity.
Higher resolution might also enable the reconstruction of electrical synapses, which are
included in the C. elegans connectome. Increasing resolution by 2× in all three dimensions
would increase the data volume by 8×. Handling much larger data volumes should be
possible as methods for acquiring and analyzing EM images are progressing rapidly.

Imaging larger
Imaging a larger volume would open up other interesting opportunities. Imaging a whole fly
CNS would enable the mapping of all pathways linking the brain and VNC. The volume of
the whole CNS is not much larger than that of the brain. In the meantime, it is possible to
establish correspondences between FlyWire and FANC, a reconstruction of a separate VNC
(Phelps et al. 2021; Azevedo et al. 2022). The first C. elegans connectome was obtained
similarly as a mosaic drawn from multiple worms (White et al. 1986). Imaging an entire fly,
both CNS and body, would enable the addition of sensory organs and muscles to the
reconstruction. This also has precedent in the C. elegans connectome, which includes
neuromuscular junctions, and the first instar Drosophila larva for which a whole-animal EM
dataset was recently published (Schoofs et al. 2023).

FlyWire and other related technologies have already been applied to millimeter-scale chunks
of mammalian brain (MICrONS Consortium et al. 2021; Shapson-Coe et al. 2021), which are
>50× larger in volume than a fly brain. The U.S. National Institutes of Health is planning a
ten year project to reconstruct a whole mouse brain from an exabyte of EM images and a
report from the Wellcome trust recently examined the road to a whole mouse brain
connectome (Jefferis et al. 2023).

Openness
The 1996 Bermuda Principles mandated daily release of Human Genome Project sequences
into the public domain (Collins, Morgan, and Patrinos 2003). We believe that openness is
also important for large-scale connectomics projects, particularly because these projects are
expensive, require coordinated effort, and take several years to complete - sharing
connectomes only after proofreading and annotation are completed prevents scientific
discovery that can occur while the connectome is being completed. Shortly after its
inception, FlyWire has been open to any Drosophila researcher. As a result, hundreds of
scientists and proofreaders from over 50 labs joined FlyWire with over 200 of them
contributing over 100 edits (Supplemental Table 1) and 86 contributing ten or more

25
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

annotations (Supplemental Table 2). As a result, there are multiple studies that used
completed portions of FlyWire’s connectome as proofreading proceeded (Zhao et al. 2022;
Mabuchi et al. 2023a; Eichler et al. 2023; Shiu et al. 2022; Baker et al. 2022; Zheng et al.
2022; Deutsch et al. 2020; Task et al. 2022; Chou et al. 2022; Israel et al. 2022; Kind et al.
2021; Sterne et al. 2021; P. Schlegel et al. 2021; Mabuchi et al. 2023b). Openness has also
enabled FlyWire to move faster by incorporating data sources from the community. The EM
data on which FlyWire is built was shared in 2018 by Bock and colleagues (Zheng et al.
2018). FlyWire’s synapse data was previously published by Buhmann et al. (Buhmann et al.
2021) who incorporated synapse segmentations from Heinrich et al. (Heinrich et al. 2018),
neurotransmitter labels for every synapse were made available ahead of publication by
Eckstein et al. (Eckstein et al. 2023), numerous annotations were contributed by Schlegel et
al., and over 90K (and counting) cell annotations have been shared by the community.

FlyWire Consortium
Doug Bland1, Krzysztof Kruk3, Zairene Lenizo16, Alexander Shakeel Bates4,5,12,13, Nseraf3, Austin
T Burke1, Katharina Eichler5, Nashra Hadjerol16, Kyle Patrick Willie1, Ryan Willie1, Yijie Yin5, John
Anthony Ocho16, Sven Dorkenwald1,2, Joshua Bañez16, Arti Yadav4, Shirleyjoy Serona16, Rey
Adrian Candilada16, Dustin Garner17, Philipp Schlegel4,5, Jet Ivan Dolorosa16, Ariel Dagohoy16,
Remer Tancontian16, Mendell Lopez16, Regine Salem16, Griffin Badalamente5, annkri (Anne
Kristiansen)3, Kendrick Joules Vinson16, Nelsie Panes16, Laia Serratosa Capdevila4, Anjali
Pandey4, Darrel Jay Akiatan16, Ben Silverman1, Dharini Sapkal4, Shaina Mae Monungolh16, Jay
Gager1, Varun Sane5, Miguel Albero16, AzureJay (Jaime Skelton)3, Márcia dos Santos5, David
Deutsch1,9, Zeba Vohra4, Kaiyu Wang14, Emil Kind18, Chitra Nair4, Dhwani Patel4, Imaan F. M.
Tamimi5, Michelle Darapan Pantujan16, James Hebditch1, Alexandre Javier4, Rashmita Rana4,
Bhargavi Parmar4, Merlin Moore1, Mark Lloyd Pielago16, Allien Mae Gogo16, Markus William
Pleijzier4, Mark Larson19, Joseph Hsu4, Thomas Stocks3, Jacquilyn Laude16, Itisha Joshi4, Chereb
Martinez16, Dhara Kakadiya4, John David Asis16, Amalia Braun20, Clyde Angelo Lim16, Alvin Josh
Mandahay16, Marchan Manaytay16, Marina Gkantia5, Kaushik Parmar4, Quinn Vanderbeck12,
Claire E. McKellar1, Philip Lenard Ampo16, Daril Bautista16, Irene Salgarella4, Christopher Dunne5,
John Clyde Saguimpa16, Eva Munnelly5, Chan Hyuk Kang21, Jansen Seguido16, Jinmook Kim21,
Gizem Sancer22, Lucia Kmecova23, Christa Baker1, Jenna Joroff12, Steven Calle23, Cathy Pilapil16,
Yashvi Patel4, Olivia Sato19, Siqi Fang4, Paul Brooks24, Mai Bui25, JousterL (Matthew
Lichtenberger)3, edmark tamboboy16, Katie Molloy19, Alexis E Santana-Cruz23, Janice Salocot16,
Celia David1, Kfay3, Seongbong Yu21, Arzoo Diwan4, Farzaan Salman26, Szi-chieh Yu1, Monika
Patel4, TR773, Sarah Morejohn1, Sebastian Molina-Obando27, Sanna Koskela14, Tansy Yang14,
bl4ckscor3 (Daniel Lehmann)3, Sangeeta Sisodiya4, Selden Koolman1, Philip K. Shiu28, Sky
Cho25, Brian Reicher19, Marlon Blanquart4, Marissa Sorek1,3, Lucy Houghton17, Hyungjun Choi21,
Matt Collie19, Joanna Eckhardt1, Benjamin Gorko17, Li Guo17, Zhihao Zheng1, Alisa Poh29, Marina
Lin25, István Taisz4, Wes Murfin52, Peter Gibb12, Nils Reinhard30, Nidhi Patel4, Sandeep Kumar1,
Minsik Yun31, Megan Wang1, Devon Jones1, Lucas Encarnacion-Rivera32, Annalena Oswald27,
Akanksha Jadia4, Leonie Walter18, Nik Drummond4, Ibrahim Tastekin33, Yuta Mabuchi34, Xin
Zhong18, Fernando J Figueroa Santiago23, Urja Verma4, Nick Byrne19, Edda Kunze18, Thomas
Crahan17, Hewhoamareismyself (Ryan Margossian)3, Haein Kim34, Iliyan Georgiev3, Fabianna
Szorenyi23, Benjamin Bargeron35, Tomke Stuerner4, Damian Demarest36, Atsuko Adachi37, Burak
Gür27, Andrearwen3, a5hm0r3, Robert Turnbull4, Andrea Sandoval28, Diego A. Pacheco12, Haley
Croke38, Alexander Thomson14, Jonas Chojetzki27, Connor Laughland14, Suchetana B. Dutta18,
Paula Guiomar Alarcón de Antón18, Binglin Huang17, Patricia Pujols23, Isabel Haber19, Amanda
González-Segarra28, Albert Lin1,6, Daniel T. Choe39, Veronika Lukyanova40, Marta Costa5, Maria

26
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ioannidou27, Zequan Liu41, Tatsuo Okubo12, Miriam A. Flynn14, Gianna Vitelli35, Meghan
Laturney28, Carolina Manyari-Diaz35, Shuo Cao42, Hyunsoo Yim21, Anh Duc Le38, Kate Maier35,
Seungyun Yu21, Yeonju Nam21, Mavil3, Eleni Samara20, Nino Mancini35, Amanda Abusaif28,
Audrey Francis43, Jesse Gayk4, Sommer S. Huntress44, Raquel Barajas33, Mindy Kim19, Xinyue
Cui34, Amy R Sterling1,3, Anna Li12, Gabriella R. Sterne28, Lena Lörsch27, Keehyun Park21, Alan
Mathew4, Taewan Kim21, Guan-ting Wu45, Serene Dhawan46, Margarida Brotas33, Cheng-hao
Zhang45, Shanice Bailey4, Alexander Del Toro28, Arie Matsliah1, Kisuk Lee1,10, Thomas Macrina1,2,
Casey Schneider-Mizell7, Mert Erginkaya33, Sergiy Popovych1,2, Oluwaseun Ogedengbe1,
Runzhe Yang1,2, Akhilesh Halageri1, Will Silversmith1, Ryan Morey1, Eric Mitchell1, Andrew
Champion4, Dodam Ih1, Nico Kemnitz1, Manuel Castro1, Zhen Jia1, Jingpeng Wu1, Stephan
Gerhard47, Shang Mu1, Barak Nehoran1,2, Eric Perlman8, J. Alexander Bae1,11, Chris S. Jordan1,
Ran Lu1, Derrick Brittain7, Kai Kuehner1, Nils Eckstein14, David J. Anderson42, Rudy Behnia37,
Salil S. Bidaye35, Davi D. Bock15, Alexander Borst20, Eugenia Chiappe33, Forrest Collman7,
Kenneth J. Colodner44, Andrew Dacks26, Barry Dickson14, Álvaro Sanz Díez37, Jan Funke14,
Denise Garcia38, Stefanie Hampel23, Volker Hartenstein48, Bassem Hassan18, Charlotte
Helfrich-Forster30, Wolf Huetteroth49, Gregory S.X.E. Jefferis4,5, Jinseop Kim21, Sung Soo Kim17,
Young-Joon Kim31, Wei-Chung Lee12, Feng Li14, Gerit A. Linneweber18, Gaby Maimon43, Richard
Mann37, Mala Murthy1, Michael Pankratz36, Lucia Prieto-Godino46, Jenny Read40, Michael
Reiser14, Katie von Reyn38, Carlos Ribeiro33, Kristin Scott28, Andrew M. Seeds23, Mareike
Selcho49, H. Sebastian Seung1,2, Marion Silies27, Julie Simpson17, Mathias F. Wernet18, Rachel I.
Wilson12, Fred W Wolf50, Zepeng Yao51, Nilay Yapici34, Meet Zandawala30

1
Princeton Neuroscience Institute, Princeton University, Princeton, USA
2
Computer Science Department, Princeton University, Princeton, USA
3
Eyewire, Boston, USA
4
Neurobiology Division, MRC Laboratory of Molecular Biology, Cambridge, UK
5
Drosophila Connectomics Group, Department of Zoology, University of Cambridge, Cambridge, UK
6
Center for the Physics of Biological Function, Princeton University, Princeton, USA
7
Allen Institute for Brain Science, Seattle, USA
8
Yikes LLC, Baltimore, USA
9
Department of Neurobiology, University of Haifa, Haifa, Israel
10
Brain & Cognitive Sciences Department, Massachusetts Institute of Technology, Cambridge, USA
11
Electrical and Computer Engineering Department, Princeton University, Princeton, USA
12
Harvard Medical School, Boston, USA
13
Centre for Neural Circuits and Behaviour, The University of Oxford, Oxford, UK
14
Janelia Research Campus, Howard Hughes Medical Institute, Ashburn, USA
15
Department of Neurological Sciences, University of Vermont, Burlington, USA
16
SixEleven, Davao City, Philippines
17
University of California, Santa Barbara, USA
18
Freie Universität Berlin, Berlin, Germany
19
Harvard, Boston, USA
20
Department Circuits-Computation-Models, Max Planck Institute for Biological Intelligence, Planegg, Germany
21
Sungkyunkwan University, Seoul, South Korea
22
Department of Neuroscience, Yale University, New Haven, USA
23
Institute of Neurobiology, University of Puerto Rico Medical Sciences Campus, San Juan, Puerto Rico
24
Department of Zoology, University of Cambridge, UK
25
Program in Neuroscience and Behavior, Mount Holyoke College, South Hadley, USA
26
Department of Biology, West Virginia University, Morgantown, USA
27
Johannes-Gutenberg University Mainz, Mainz, Germany
28
University of California, Berkeley, USA
29
University of Queensland, Brisbane, Australia
30
Julius-Maximilians-Universität Würzburg, Würzburg, Germany
31
Gwangju Institute of Science and Technology, Gwangju, South Korea
32
Stanford University School of Medicine, Stanford, USA
33
Champalimaud Foundation, Lisbon, Portugal

27
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

34
Cornell University, Ithaca, USA
35
Max Planck Florida Institute for Neuroscience, Jupiter, USA
36
University of Bonn, Bonn, Germany
37
Zuckerman Institute, Columbia University, New York, USA
38
Drexel, Philadelphia, USA
39
Seoul National University, Seoul, South Korea
40
Newcastle University, Newcastle, UK
41
RWTH Aachen University, Aachen, Germany
42
Caltech, Pasadena, USA
43
Rockefeller University, New York, USA
44
Mount Holyoke College, South Hadley, USA
45
National Hualien Senior High School, Hualien, Taiwan
46
The Francis Crick Institute, London, UK
47
Aware LLC, Zurich, Switzerland
48
University of California, Los Angeles, USA
49
Institute of Biology, Leipzig University, Leipzig, Germany
50
University of California, Merced, USA
51
University of Florida, Gainesville, USA
52
Retired MD-PhD, Fort Collins, USA

Acknowledgements
We thank John Wiggins, G. McGrath, and Dave Barlieb for computer system administration
and M. Husseini for project administration. We are grateful to J. Maitin-Shepard for
Neuroglancer. Mala Murthy and Sebastian Seung acknowledge support from the National
Institutes of Health (NIH) BRAIN Initiative RF1 MH117815, RF1 MH129268 and U24
NS126935, from the Princeton Neuroscience Institute, as well as assistance from Google.
Davi Bock was supported by NIH NIMH BRAIN Initiative grant 1RF1MH120679-01 and a
Neuronex2 award (NSF 2014862). Gregory S.X.E. Jefferis and Davi Bock were supported by
Wellcome Trust Collaborative Award (203261/Z/16/Z). Gregory S.X.E. Jefferis was
supported by Wellcome Trust Collaborative Award 220343/Z/20/Z, Neuronex2 award (MRC
MC_EX_MR/T046279/1) and received core support from the MRC (MC-U105188491).
Ibrahim Tastekin was supported with a Marie Skłodowska-Curie postdoctoral fellowship
(H2020-WF-01-2018-867459 to Ibrahim Tastekin) and by the Portuguese Research Council
(Grant PTDC/MED-NEU/4001/2021). Andrew Seeds and Stefanie Hampel were supported
by National Institute Of Neurological Disorders And Stroke of the National Institutes of Health
under Award Number RF1NS121911. Derrick Brittain, Casey Schneider-Mizell, and Forrest
Collman thank the Allen Institute for Brain Science founder, P. G. Allen, for his vision,
encouragement and support. This work was also supported by the Intelligence Advanced
Research Projects Activity via Department of Interior/Interior Business Center contract no.
D16PC0005 to H.S.S. The US Government is authorized to reproduce and distribute reprints
for Governmental purposes notwithstanding any copyright annotation thereon. The views
and conclusions contained herein are those of the authors and should not
be interpreted as necessarily representing the official policies or endorsements, either
expressed or implied, of Intelligence Advanced Research Projects Activity, Department of
Interior/Interior Business Center or the US Government.

Contributions
Members of the FlyWire consortium contributed proofreading and annotations (see
Supplemental Tables 1, 2). SGerhard provided braincircuits.io. TM and NK realigned the

28
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

dataset with methods developed by EM, BN and TM and infrastructure developed by SP, ZJ.
JAB, SM wrote code for masking defects and misalignments. KL trained the convolutional
net for boundary detection, using ground-truth data realigned by DI. JW used the
convolutional net to generate an affinity map that was segmented by RL. NK, MAC, OO, AH,
CSJ, KKuehner and ARS adapted and improved Neuroglancer for proofreading and
annotations. JG, KKruk, AM, SD, FC and CSM created interactive analysis and annotation
tools for the community. AM created Codex with help from ARS, SD, KKuehner and RM.
ARS and AM created the website. ARS, CEM and MS onboarded community members and
tested new proofreaders. ARS, MS, CSJ and CEM designed tutorials. CEM, ARS and MS
provided community support. SD, FC, CSM, CSJ, AH, DBrittain and WMS built and
maintained CAVE for FlyWire and managed user access. SD, PS, AM and EP curated the
data and made it available for download. EP and DDB provided a coordinate mapping
service. ASB, NE, GSXEJ and JF provided neurotransmitter information. SCY, CEM, MC,
KE, YY and PS trained and managed proofreaders. SD, SCY, PS and GSXEJ led the
targeted proofreading effort. SD, PS, AM, AChampion and KKuehner maintained the
proofreading management platforms. SD evaluated the proofreading accuracy. SD, AL,
HSS, DD and RY analyzed the data. SD, DBland and SCY annotated and analyzed the
ocellar circuit. SD, HSS, MM, AL, PS and ARS wrote the manuscript with feedback from
ASB, WHuetteroth, GSXEJ and contributions from all authors. HSS, MM, GSXEJ, DDB
sponsored large-scale proofreading. GSXEJ, DDB led the Cambridge effort. MM, HSS led
the overall effort.

Competing interests
T. Macrina, K. Lee, S. Popovych, D. Ih, N. Kemnitz, and H. S. Seung declare financial
interests in Zetta AI.

29
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Methods

Neuropils
Meshes for individual neuropils were based on work by Ito et al. (Ito et al. 2014). More
specifically, we took meshes previously generated from a full brain segmentation of the
JFRC2 template brain which are also used by the Virtual Fly Brain project (see also
https://natverse.org/nat.flybrains/reference/JFRC2NP.surf.html). These meshes were moved
from JFRC2 into FlyWire (FAFB14.1) space through a series of non-rigid transforms. In
addition, we also generated two neuropil meshes for the laminae and for the ocellar
ganglion. For these, the FlyWire synapse cloud was voxelized with 2 µm isotropic resolution,
meshed using the marching cube algorithm using Python and manually post-processed in
Blender 3d.

We calculated a volume for each neuropil using its mesh. In the aggregated volumes
presented in the paper we assigned the lamina, medulla, accessory medulla, lobula, lobula
plate to the optic lobe. The remaining neuropils but the ocellar ganglion were assigned to the
central brain.

Neuropil synapse assignments

We assigned synapses to neuropils based on their presynaptic location. We used ncollpyde
(https://pypi.org/project/ncollpyde/) to calculate if the location was within a neuropil mesh and
assigned the synapse accordingly. Some synapses remained unassigned after this step
because the neuropils only resemble rough outlines of the underlying data. We then
assigned all remaining synapses to the closest neuropil if the synapse was within 10 µm
from it. The remaining synapses were left unassigned.

Correction of left-right inversion

Our reconstruction used the FAFB EM dataset (Zheng et al. 2018). A number of consortium
members (A. Bates, P. Kandimalla, S. Noselli) alerted us that the FAFB imagery seemed
left-right inverted based on cell types innervating the asymmetric body (Lapraz et al. 2023).
Eventually a left-right inversion during FAFB imaging was confirmed. All side annotations in
figures, in Codex and elsewhere are based on the true biological side. For technical reasons
we were unable to invert the underlying FAFB image data and therefore continue to show
images and reconstructions in the same orientation as (Zheng et al. 2018) although we now
know in such frontal views the fly’s left is on the viewer’s left. For full details of this issue
including approaches to display FAFB and other brain data with the correct chirality, please
see our companion paper (Schlegel et al., in prep).

Proofreading system
FlyWire uses the Connectome Annotation Versioning Engine (CAVE) for hosting the
proofreadable segmentation and all of its annotations. CAVE’s proofreading system is the
PyChunkedGraph which has been described in detail elsewhere (Dorkenwald, Turner, et al.
2022; Dorkenwald, McKellar, et al. 2022).

30
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Proofreading annotations
Any user in FlyWire was able to mark a cell as complete, indicating that a cell was good for
analysis. However, such annotations did not prevent future proofreading of a cell as
commonly smaller branches were added later on. We created an annotation table for these
completion markings. Each completion marking was defined by a point in space and the cell
segment that overlapped with this point at any given time during proofreading was
associated with the annotation. We created a webservice allowing users to submit
completion markings for any cell. For convenience, we added an interface to this surface
directly into Neuroglancer such that users can submit completion information for cells right
after proofreading (Supplemental Information 1). When users submitted completion
annotations we also recorded the current state of the cell. We encouraged users to submit
new completion markings for a cell that they edited to indicate that edits were intentional.
Recording the status of a cell at submission allowed us to calculate volumetric changes to a
cell through further proofreading and flag cells for review if they received substantial
changes without new completion markings.

Onboarding proofreaders
Proofreaders came from several distinct labor pools: community members, citizen scientists
from Eyewire (Flyers), and professional proofreading teams at Princeton and Cambridge. All
proofreaders completed the built-in interactive tutorial and directed to Self-Guided
Proofreading Training. For practice and learning purposes, the Sandbox, a complete replica
of the FlyWire data, allowed new users to freely make edits and explore without affecting the
actual “Production“ dataset. When ready, an Onboarding Coordinator tested the new
proofreader before giving access to the Production dataset (Dorkenwald et al., 2022). Later
onboarding called for users to send demonstration Sandbox edits that were reviewed by the
Onboarding Coordinator. A new class of view-only users was introduced in early 2023,
allowing researchers early data access for analysis purposes. All early access users
attended a live onboarding session in Zoom prior to being granted edit or view access.

Training the professional proofreading team

The professional proofreading team received additional proofreading training. Correct
proofreading relies on a diverse array of 2D and 3D visual cues. Proofreaders learned about
3D morphology, resulting from false merger or false split without the knowledge of knowing
what types of cells they are. Proofreaders studied various types of ultrastructures as the
ultrastructures provide valuable 2D cues and serve as reliable guides for accurate tracing.
Before professional proofreaders were admitted into Production, each of them practiced on
average >200 cells in a testing dataset where additional feedback was given. In this dataset,
we determined the accuracy of test cells by comparing them to ground-truth reconstructions.
To improve proofreading quality, peer learning was highly encouraged.

Recruitment of citizen scientists

The top 100 players from Eyewire, a gamified EM reconstruction platform that crowdsources
reconstructions in mouse retina and zebrafish hindbrain (Kim et al. 2014), received an
invitation to beta test proofreading in FlyWire. A new set of user onboarding and training
materials were created for citizen scientists, including: a blog, forum, and public Google
docs. We created bite-sized introduction videos, a comprehensive “FlyWire 101” resource,

31
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

as well as an Optic Lobe Cell Guide to aid users in understanding the unique morphology of
flies. A virtual Citizen Science Symposium introduced players to the project, after which the
self-dubbed “Flyers” began creating their own resources, such as a new comprehensive
visual guide to cell types, conducting literature reviews, and even developing helpful FlyWire
plugins. As of publication, FlyWire has 12 add-on apps ranging from a batch processor to
cell naming helper (https://blog.flywire.ai/2022/08/11/flywire-addons/).

Proofreading strategy to complete the connectome

As previously described (Dorkenwald, McKellar, et al. 2022), proofreading of the
connectome was focused on the microtubule-rich ‘backbones’ of neurons. Microtubule-free
‘twigs’ were only added if discovered incidentally or sought out specifically by members of
the community. After proofreading, users marked neuronal segments as ‘complete’ indicating
that neurons were ready for analysis but further changes remained possible. While
Drosophila neuroscientist members of the FlyWire community generally contributed
proofreading for their neurons of interest, the bulk of the segments was proofread by
professional proofreaders in the following way: first we proofread all segments with an
automatically detected nucleus in the central brain (Mu et al. 2021) by extending it as much
as possible and removing all false mergers (pieces of other neurons or glia attached), and
second, going through the remaining segments in descending order of their synapse count
(pre+post) up to a predefined size threshold of 100 synapses.

Quality Assurance
To assess quality, a group of expert centralized proofreaders conducted a review of 3106
segments in the central brain. These specific neurons were chosen based on certain criteria
such as significant change since being marked complete and small overall volume. An
additional 826 random neurons were included in the review pool as well. Proofreaders were
unaware which neurons were added for quality measurement and which ones because they
were flagged by a metric. We compared the 826 neurons before and after the review and
found that the initial reconstruction scored an average F1-Score of 99.2% by volume (Ext.
Data Fig. 1-2a,b).

Quantification of proofreading effort

Any quantification of the total proofreading time that was required to create the FlyWire
resource is a rough estimate because of the distributed nature of the community, the
interlacing of analysis and proofreading and the variability in how proofreading was
performed. The first public release, version 630, required 2,712,769 edits. We measured
proofreading times during early proofreading rounds that included proofreading of whole
cells in the central brain. We collected timings and number of edits for 29,135 independent
proofreading tasks after removing outliers with more than 500 edits. From this data we were
able to calculate an average time per edit. However, we observed that proofreading times
per edit were much higher for proofreading tasks that required few edits (<5). That meant
that our measurements were not representative for the second round of proofreading which
went over segments with > 100 synapses. These usually required 1-5 edits. We adjusted for
that by computing estimates for proofreading speeds of both rounds by limiting the
calculations to a subset of the timed tasks: (round 1) The average time per edit in our
proofreading time dataset, (round 2) the average time of tasks with 1-5 edits. We average

32
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

these times for an overall proofreading time because the number of tasks in each category
were similar. The result was an average time of 79s per edit which adds up to an estimate of
29.8 person-years assuming a 2000h work year.

Completion rates
We adopted the completion rate calculations from the hemibrain (Scheffer et al. 2020). Every
presynaptic and postsynaptic location was assigned to a segment. Using the neuropil
assignments, we then calculated the fraction of presynapses that were assigned to
segments marked as proofread for each neuropil and analogous for postsynaptic location.

Comparison with the hemibrain

We retrieved the latest completion rates and synapse numbers for the hemibrain from
neuprint (v1.2.1). In some cases, neuropil comparisons were not directly possible because of
redefined regions in the hemibrain dataset. We excluded these regions from the comparison.

Crowdsourced annotation
FlyWire’s large community and diversity of expertise allowed us to crowdsource the
identification of neurons. There is no limit to the number of annotations a neuron can
receive. A standardized format is encouraged but not required. One user might first report
that a neuron is a descending interneuron, while another might add that it is the Giant Fiber
descending neuron, and another might add all its synonyms and citations from the literature.
Contributors’ names are visible so they can be consulted if there is disagreement. The
disadvantage to this approach is that there isn’t one precise name for every neuron, but the
advantage is a richness of information and dialog. The annotations are not meant to be a
finished, static list, but a continually growing, living data source. These annotations were
solicited from the FlyWire community through Town Halls, email announcements, interest
groups in the FlyWire Forum, online instructions, and by personal contact from the
Community Manager. Citizen scientists also contributed annotations, after receiving training
on particular cell types by experts.

Neuron categorizations
Neuron categorization, sensory modality annotations and nerve assignments are described
in detail in Schlegel et al. In brief, neurons were assigned to one of three “flow” classes:
afferent (to the brain), intrinsic (within the brain), and efferent (out of the brain). Intrinsic
neurons had their entire arbor within the FlyWire dataset. This included cells that projected
to and from the subesophageal zone (SEZ). Next, each flow class was divided into “super”
classes in the following way. afferent: sensory, ascending. intrinsic: central, optic, visual
projection (from the optic lobes to the central brain), visual centrifugal (from the central brain
to the optic lobes). efferent: endocrine, descending, motor.

Skeletonization and path length calculation

We generated skeletons for all neurons marked as proofread using skeletor
(https://github.com/navis-org/skeletor) which implements multiple skeletonization algorithms
such as TEASAR (Sato et al. 2000). In brief, neuron meshes from the exported

33
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

segmentation (LOD 1) were downloaded and skeletonized using the “wavefront” method in
skeletor. These raw skeletons were then further processed (e.g. to remove false twigs and
heal breaks) and produce downsampled versions using navis
(https://github.com/navis-org/navis). A modified version of this skeletonization pipeline is
implemented in fafbseg (https://github.com/navis-org/fafbseg-py).

Synaptic connections
We imported the automatically predicted synapses from Buhmann et al. (Buhmann et al.
2021) which we combined with the predictions by Heinrich et al. to assign scores to all
synapses (Heinrich et al. 2018) to improve precision. We removed synapses from the
imported list if they fulfilled any of the following criteria: (1) either the pre- or postsynaptic
location remained unassigned to a segment (proofread or unproofread), (2) It had a score
≤50.

Connection threshold
For all the analyses presented in this paper, save for synapse distributions, we employed a
consistent threshold of >4. Our decision to use a synapse threshold on connections was due
partly to the fact that synapses in the FlyWire dataset were not manually proofread. For
these analyses, many of which demonstrate the high interconnectivity of the fly brain, we
chose a conservative threshold to ensure that considered connections are real. Use of a
threshold is also in keeping with previous work analyzing wiring diagrams in Drosophila
(Scheffer et al. 2020). Thus, we are likely undercounting the number of true connections.
The distribution of synapse counts (FIgure 3f) does not display any bimodality that could be
used to set the threshold. Therefore, the choice of 5 synapses per connection is a
reasonable but arbitrary one. In the companion paper analyzing the network properties of the
FlyWire connectome, it is found that statistical properties of the whole-brain network, such as
reciprocity and clustering coefficient, are robust to our choice of threshold (Lin et al., in prep).
The FlyWire data is available without an imposed threshold, so users can choose their own
appropriate threshold for their specific use case.

Neuropil projectome construction

Under the simplifying assumptions that information flow through the neuron can be
approximated by the fraction of synapses in a given region, and that inputs and outputs can
be treated independently, we can construct a matrix representing the projections of a single
neuron between neuropils. The fractional inputs of a given neuron are a 1 x N vector
containing the fraction of incoming synapses the neuron has in each of the N neuropils, and
the fractional outputs are a similar vector containing the fraction of outgoing synapses in
each of the N neuropils. We multiply these vectors against each other to generate the N x N
matrix of the neuron's fractional weights. Summing these matrices across all intrinsic
neurons produces a matrix of neuropil-to-neuropil connectivity (Figure 4a). In this
projectome, all neurons contribute an equal total weight of one.

34
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Dominant input side

We assigned neuropils to the left and right hemispheres or the center if the neuropil has no
homologue. We then counted how many postsynapses each neuron had in each of these
three regions and assigned it to the one with the largest count.

Contralateral and bilateral neuron analysis

For each neuron, we calculated the fraction of presynapses in the left and right hemisphere.
The hemisphere opposite its dominant input side was named the contralateral hemisphere.
We excluded neurons that had either most of their presynapses or most of their
postsynapses in the center region.

Rank analysis & Information Flow

We used the information flow algorithm implemented by Schlegel et al. (P. Schlegel et al.
2021) (https://github.com/navis-org/navis) to calculate a rank for each neuron starting with a
set of seed neurons. The algorithm traverses the synapse graph of neurons probabilistically.
The likelihood of a neuron being added to the traversed set increased linearly with the
fraction of synapses it receives from already traversed neurons up to 30% and was
guaranteed above this threshold. We repeated the rank calculation for all sets of afferent
neurons as seed as well as the whole set of sensory neurons. The groups we used are:

olfactory receptor neurons, gustatory receptor neurons, mechanosensory Johnston’s Organ

neurons, head and neck bristle mechanosensory neurons, mechanosensory taste peg
neurons, thermosensory neurons, hygrosensory neurons, visual projection neurons, visual
photoreceptors, ocellar photoreceptors and ascending neurons.

Additionally, we created input seeds by combining all listed modalities, all sensory
modalities, and all listed modalities with visual sensory groups excluded.

For each modality we then ordered the neurons according to their rank and assigned them a
percentile based on their location in the order. To compute a reduced dimensionality, we
treated the vector of all ranks (one for each modality) as neuron embedding and calculated
two dimensional embeddings using UMAP (McInnes, Healy, and Melville 2018) with the
following parameters: n_components=2, min_dist=0.35, metric="cosine", n_neighbors=50,
learning_rate=.1, n_epochs=1000.

35
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

References
Aimon, Sophie, Karen Y. Cheng, Julijana Gjorgjieva, and Ilona C. Grunwald Kadow. 2022.
“Combined Patterns of Activity of Major Neuronal Classes Underpin a Global Change in
Brain State during Spontaneous and Forced Walk inDrosophila.” bioRxiv.
https://doi.org/10.1101/2022.01.17.476660.
Amin, Hoger, Anthi A. Apostolopoulou, Raquel Suárez-Grimalt, Eleftheria Vrontou, and
Andrew C. Lin. 2020. “Localized Inhibition in the Drosophila Mushroom Body.” eLife 9
(September). https://doi.org/10.7554/eLife.56954.
Arendt, Detlev, Maria Antonietta Tosches, and Heather Marlow. 2016. “From Nerve Net to
Nerve Ring, Nerve Cord and Brain--Evolution of the Nervous System.” Nature Reviews.
Neuroscience 17 (1): 61–72.
Azevedo, Anthony, Ellen Lesser, Brandon Mark, Jasper Phelps, Leila Elabbady, Sumiya
Kuroda, Anne Sustar, et al. 2022. “Tools for Comprehensive Reconstruction and
Analysis of Drosophila Motor Circuits.” bioRxiv.
https://doi.org/10.1101/2022.12.15.520299.
Bae, J. Alexander, Shang Mu, Jinseop S. Kim, Nicholas L. Turner, Ignacio Tartavull, Nico
Kemnitz, Chris S. Jordan, et al. 2018. “Digital Museum of Retinal Ganglion Cells with
Dense Anatomy and Physiology.” Cell 173 (5): 1293–1306.e19.
Baker, Christa A., Claire McKellar, Rich Pang, Aljoscha Nern, Sven Dorkenwald, Diego A.
Pacheco, Nils Eckstein, Jan Funke, Barry J. Dickson, and Mala Murthy. 2022. “Neural
Network Organization for Courtship-Song Feature Detection in Drosophila.” Current
Biology: CB 32 (15): 3317–33.e7.
Bates, Alexander Shakeel, Jasper Janssens, Gregory Sxe Jefferis, and Stein Aerts. 2019.
“Neuronal Cell Types in the Fly: Single-Cell Anatomy Meets Single-Cell Genomics.”
Current Opinion in Neurobiology 56 (June): 125–34.
Bates, Alexander Shakeel, James D. Manton, Sridhar R. Jagannathan, Marta Costa, Philipp
Schlegel, Torsten Rohlfing, and Gregory Sxe Jefferis. 2020. “The Natverse, a Versatile
Toolbox for Combining and Analysing Neuroanatomical Data.” eLife.
https://doi.org/10.7554/elife.53350.
Bates, Alexander S., Philipp Schlegel, Ruairi J. V. Roberts, Nikolas Drummond, Imaan F. M.
Tamimi, Robert Turnbull, Xincheng Zhao, et al. 2020. “Complete Connectomic
Reconstruction of Olfactory Projection Neurons in the Fly Brain.” Current Biology: CB 30
(16): 3183–99.e6.
Borst, Alexander, and Moritz Helmstaedter. 2015. “Common Circuit Design in Fly and
Mammalian Motion Vision.” Nature Neuroscience 18 (8): 1067–76.
BRAIN Initiative Cell Census Network (BICCN). 2021. “A Multimodal Cell Census and Atlas
of the Mammalian Primary Motor Cortex.” Nature 598 (7879): 86–102.
Braitenberg, Valentino. 1984. Vehicles: Experiments in Synthetic Psychology. Cambridge,
MA: MIT Press.
Brezovec, Luke E., Andrew B. Berger, Shaul Druckmann, and Thomas R. Clandinin. 2022.
“Mapping the Neural Dynamics of Locomotion across the Drosophila Brain.” bioRxiv.
https://doi.org/10.1101/2022.03.20.485047.
Brittin, Christopher A., Steven J. Cook, David H. Hall, Scott W. Emmons, and Netta Cohen.
2021. “A Multi-Scale Brain Map Derived from Whole-Brain Volumetric Reconstructions.”
Nature 591 (7848): 105–10.
Buhmann, Julia, Arlo Sheridan, Caroline Malin-Mayor, Philipp Schlegel, Stephan Gerhard,
Tom Kazimiers, Renate Krause, et al. 2021. “Automatic Detection of Synaptic Partners
in a Whole-Brain Drosophila Electron Microscopy Data Set.” Nature Methods 18 (7):
771–74.
Cachero, S., A. D. Ostrovsky, Y. Y. Jai, and B. J. Dickson. 2010. “Sexual Dimorphism in the
Fly Brain.” Current Biology: CB.
https://www.sciencedirect.com/science/article/pii/S0960982210009474.
Caron, Sophie J. C., Vanessa Ruta, L. F. Abbott, and Richard Axel. 2013. “Random

36
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Convergence of Olfactory Inputs in the Drosophila Mushroom Body.” Nature 497 (7447):
113–17.
Cheong, H. S. J., K. Eichler, T. Stuerner, S. K. Asinof, A. S. Champion, E. C. Marin, T. B.
Oram, et al. 2023. “Transforming Descending Input into Behavior: The Organization of
Premotor Circuits in the Drosophila Male Adult Nerve Cord Connectome.” bioRxiv.
https://doi.org/10.1101/2023.06.07.543976.
Chiang, Ann-Shyn, Chih-Yung Lin, Chao-Chun Chuang, Hsiu-Ming Chang, Chang-Huain
Hsieh, Chang-Wei Yeh, Chi-Tin Shih, et al. 2011. “Three-Dimensional Reconstruction of
Brain-Wide Wiring Networks in Drosophila at Single-Cell Resolution.” Current Biology:
CB 21 (1): 1–11.
Chklovskii, Dmitri B. 2004. “Synaptic Connectivity and Neuronal Morphology: Two Sides of
the Same Coin.” Neuron 43 (5): 609–17.
Chou, Ya-Hui, Chi-Jen Yang, Hao-Wei Huang, Nan-Fu Liou, Michael Raphael Panganiban,
David Luginbuhl, Yijie Yin, et al. 2022. “Mating-Driven Variability in Olfactory Local
Interneuron Wiring.” Science Advances 8 (7): eabm7723.
Coen, Philip, Jan Clemens, Andrew J. Weinstein, Diego A. Pacheco, Yi Deng, and Mala
Murthy. 2014. “Dynamic Sensory Cues Shape Song Structure in Drosophila.” Nature
507 (7491): 233–37.
Cognigni, Paola, Johannes Felsenberg, and Scott Waddell. 2018. “Do the Right Thing:
Neural Network Mechanisms of Memory Formation, Expression and Update in
Drosophila.” Current Opinion in Neurobiology 49 (April): 51–58.
Collins, Francis S., Michael Morgan, and Aristides Patrinos. 2003. “The Human Genome
Project: Lessons from Large-Scale Biology.” Science 300 (5617): 286–90.
Cook, Steven J., Travis A. Jarrell, Christopher A. Brittin, Yi Wang, Adam E. Bloniarz, Maksim
A. Yakovlev, Ken C. Q. Nguyen, et al. 2019. “Whole-Animal Connectomes of Both
Caenorhabditis Elegans Sexes.” Nature 571 (7763): 63–71.
Costandi, Mo. 2012. “Anti-Connectome-Ism.” The Guardian, September 21, 2012.
https://www.theguardian.com/science/neurophilosophy/2012/sep/21/connectome-review
.
Croset, Vincent, Christoph D. Treiber, and Scott Waddell. 2018. “Cellular Diversity in the
Drosophila Midbrain Revealed by Single-Cell Transcriptomics.” eLife 7 (April).
https://doi.org/10.7554/eLife.34550.
Davie, Kristofer, Jasper Janssens, Duygu Koldere, Maxime De Waegeneer, Uli Pech, Łukasz
Kreft, Sara Aibar, et al. 2018. “A Single-Cell Transcriptome Atlas of the Aging Drosophila
Brain.” Cell 174 (4): 982–98.e20.
Davis, Fred P., Aljoscha Nern, Serge Picard, Michael B. Reiser, Gerald M. Rubin, Sean R.
Eddy, and Gilbert L. Henry. 2020. “A Genetic, Genomic, and Computational Resource
for Exploring Neural Circuit Function.” eLife 9 (January).
https://doi.org/10.7554/eLife.50901.
Denk, Winfried, Kevin L. Briggman, and Moritz Helmstaedter. 2012. “Structural
Neurobiology: Missing Link to a Mechanistic Understanding of Neural Computation.”
Nature Reviews. Neuroscience 13 (5): 351–58.
Denk, Winfried, and Heinz Horstmann. 2004. “Serial Block-Face Scanning Electron
Microscopy to Reconstruct Three-Dimensional Tissue Nanostructure.” PLoS Biology 2
(11): e329.
Deutsch, David, Diego Pacheco, Lucas Encarnacion-Rivera, Talmo Pereira, Ramie Fathy,
Jan Clemens, Cyrille Girardin, et al. 2020. “The Neural Basis for a Persistent Internal
State in Drosophila Females.” eLife 9 (November). https://doi.org/10.7554/eLife.59502.
Dombrovski, Mark, Martin Y. Peek, Jin-Yong Park, Andrea Vaccari, Marissa Sumathipala,
Carmen Morrow, Patrick Breads, et al. 2023. “Synaptic Gradients Transform Object
Location to Action.” Nature 613 (7944): 534–42.
Dorkenwald, Sven, Claire E. McKellar, Thomas Macrina, Nico Kemnitz, Kisuk Lee, Ran Lu,
Jingpeng Wu, et al. 2022. “FlyWire: Online Community for Whole-Brain Connectomics.”
Nature Methods 19 (1): 119–28.
Dorkenwald, Sven, Philipp J. Schubert, Marius F. Killinger, Gregor Urban, Shawn Mikula,

37
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Fabian Svara, and Joergen Kornfeld. 2017. “Automated Synaptic Connectivity Inference
for Volume Electron Microscopy.” Nature Methods 14 (4): 435–42.
Dorkenwald, Sven, Nicholas L. Turner, Thomas Macrina, Kisuk Lee, Ran Lu, Jingpeng Wu,
Agnes L. Bodor, et al. 2022. “Binary and Analog Variation of Synapses between Cortical
Pyramidal Neurons.” eLife 11 (November). https://doi.org/10.7554/eLife.76120.
Eckstein, Nils, Alexander Shakeel Bates, Andrew Champion, Michelle Du, Yijie Yin, Philipp
Schlegel, Alicia Kun-Yang Lu, et al. 2023. “Neurotransmitter Classification from Electron
Microscopy Images at Synaptic Sites in Drosophila Melanogaster.” bioRxiv.
https://doi.org/10.1101/2020.06.12.148775.
Eichler, Katharina, Stefanie Hampel, Adrián Alejandro-García, Steven A. Calle-Schuler,
Alexis Santana-Cruz, Lucia Kmecova, Jonathan M. Blagburn, Eric D. Hoopfer, and
Andrew M. Seeds. 2023. “Somatotopic Organization among Parallel Sensory Pathways
That Promote a Grooming Sequence in Drosophila.” bioRxiv.
https://doi.org/10.1101/2023.02.11.528119.
Felleman, D. J., and D. C. Van Essen. 1991. “Distributed Hierarchical Processing in the
Primate Cerebral Cortex.” Cerebral Cortex 1 (1): 1–47.
Fischbach, K-F, and A. P. M. Dittrich. 1989. “The Optic Lobe of Drosophila Melanogaster. I. A
Golgi Analysis of Wild-Type Structure.” Cell and Tissue Research 258 (3).
https://doi.org/10.1007/bf00218858.
Fisher, Yvette E. 2022. “Flexible Navigational Computations in the Drosophila Central
Complex.” Current Opinion in Neurobiology 73 (April): 102514.
Haag, Juergen, Adrian Wertz, and Alexander Borst. 2007. “Integration of Lobula Plate
Output Signals by DNOVS1, an Identified Premotor Descending Neuron.” The Journal
of Neuroscience: The Official Journal of the Society for Neuroscience 27 (8):
1992–2000.
Hall, D. H., and R. L. Russell. 1991. “The Posterior Nervous System of the Nematode
Caenorhabditis Elegans: Serial Reconstruction of Identified Neurons and Complete
Pattern of Synaptic Interactions.” The Journal of Neuroscience: The Official Journal of
the Society for Neuroscience 11 (1): 1–22.
Harris, Julie A., Stefan Mihalas, Karla E. Hirokawa, Jennifer D. Whitesell, Hannah Choi, Amy
Bernard, Phillip Bohn, et al. 2019. “Hierarchical Organization of Cortical and Thalamic
Connectivity.” Nature 575 (7781): 195–202.
Hayworth, Kenneth J., David Peale, Michał Januszewski, Graham W. Knott, Zhiyuan Lu, C.
Shan Xu, and Harald F. Hess. 2020. “Gas Cluster Ion Beam SEM for Imaging of Large
Tissue Samples with 10 Nm Isotropic Resolution.” Nature Methods 17 (1): 68–71.
Heinrich, Larissa, Jan Funke, Constantin Pape, Juan Nunez-Iglesias, and Stephan Saalfeld.
2018. “Synaptic Cleft Segmentation in Non-Isotropic Volume Electron Microscopy of the
Complete Drosophila Brain.” In Medical Image Computing and Computer Assisted
Intervention – MICCAI 2018, 317–25. Springer International Publishing.
Hofbauer, A., and E. Buchner. 1989. “Does Drosophila Have Seven Eyes?” Die
Naturwissenschaften 76 (7): 335–36.
Hong, Elizabeth J., and Rachel I. Wilson. 2015. “Simultaneous Encoding of Odors by
Channels with Diverse Sensitivity to Inhibition.” Neuron 85 (3): 573–89.
Hu, Karin G., Heinrich Reichert, and William S. Stark. 1978. “Electrophysiological
Characterization ofDrosophila Ocelli.” Journal of Comparative Physiology 126 (1):
15–24.
Hulse, Brad K., Hannah Haberkern, Romain Franconville, Daniel B. Turner-Evans, Shinya
Takemura, Tanya Wolff, Marcella Noorman, et al. 2020. “A Connectome of the
Drosophila Central Complex Reveals Network Motifs Suitable for Flexible Navigation
and Context-Dependent Action Selection.” bioRxiv.
https://www.biorxiv.org/content/10.1101/2020.12.08.413955v2.full-text.
Israel, Shai, Eyal Rozenfeld, Denise Weber, Wolf Huetteroth, and Moshe Parnas. 2022.
“Olfactory Stimuli and Moonwalker SEZ Neurons Can Drive Backward Locomotion in
Drosophila.” Current Biology: CB 32 (5): 1131–49.e7.
Ito, Kei, Kazunori Shinomiya, Masayoshi Ito, J. Douglas Armstrong, George Boyan, Volker

38
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Hartenstein, Steffen Harzsch, et al. 2014. “A Systematic Nomenclature for the Insect
Brain.” Neuron 81 (4): 755–65.
Jabr, F. 2012. “The Connectome Debate: Is Mapping the Mind of a Worm Worth It.” Scientific
American, 2012.
Jain, V., J. F. Murray, F. Roth, S. Turaga, V. Zhigulin, K. L. Briggman, M. N. Helmstaedter, W.
Denk, and H. S. Seung. 2007. “Supervised Learning of Image Restoration with
Convolutional Networks.” In 2007 IEEE 11th International Conference on Computer
Vision, 1–8. ieeexplore.ieee.org.
Januszewski, Michał, Jörgen Kornfeld, Peter H. Li, Art Pope, Tim Blakely, Larry Lindsey,
Jeremy Maitin-Shepard, Mike Tyka, Winfried Denk, and Viren Jain. 2018.
“High-Precision Automated Reconstruction of Neurons with Flood-Filling Networks.”
Nature Methods 15 (8): 605–10.
Jefferis, Gregory, Lucy Collins, Carles Bosch, Marta Costa, and Philipp Schlegel. 2023.
“Scaling up Connectomics,” June.
https://cms.wellcome.org/sites/default/files/2023-06/Connectomics-scaling-up-connecto
mics_0.pdf.
Jerison, H. J. 1955. “Brain to Body Ratios and the Evolution of Intelligence.” Science 121
(3144): 447–49.
Karuppudurai, Thangavel, Tzu-Yang Lin, Chun-Yuan Ting, Randall Pursley, Krishna V.
Melnattur, Fengqiu Diao, Benjamin H. White, et al. 2014. “A Hard-Wired Glutamatergic
Circuit Pools and Relays UV Signals to Mediate Spectral Preference in Drosophila.”
Neuron 81 (3): 603–15.
Kasthuri, Narayanan, and Jeff W. Lichtman. 2007. “The Rise of the ‘Projectome.’” Nature
Methods. Springer Science and Business Media LLC.
Kim, Anmo J., Lisa M. Fenk, Cheng Lyu, and Gaby Maimon. 2017. “Quantitative Predictions
Orchestrate Visual Signaling in Drosophila.” Cell 168 (1-2): 280–94.e12.
Kim, Hyunsoo, Mihoko Horigome, Yuki Ishikawa, Feng Li, J. Scott Lauritzen, Gwyneth Card,
Davi D. Bock, and Azusa Kamikouchi. 2020. “Wiring Patterns from Auditory Sensory
Neurons to the Escape and Song-Relay Pathways in Fruit Flies.” The Journal of
Comparative Neurology 528 (12): 2068–98.
Kim, J. S., Matthew J. Greene, Aleksandar Zlateski, Kisuk Lee, Mark Richardson, Srinivas C.
Turaga, Michael Purcaro, et al. 2014. “Space-Time Wiring Specificity Supports Direction
Selectivity in the Retina.” Nature 509 (7500): 331–36.
Kind, Emil, Kit D. Longden, Aljoscha Nern, Arthur Zhao, Gizem Sancer, Miriam A. Flynn,
Connor W. Laughland, et al. 2021. “Synaptic Targets of Photoreceptors Specialized to
Detect Color and Skylight Polarization in Drosophila.” eLife 10 (December).
https://doi.org/10.7554/eLife.71858.
Klapoetke, Nathan C., Aljoscha Nern, Edward M. Rogers, Gerald M. Rubin, Michael B.
Reiser, and Gwyneth M. Card. 2022. “A Functionally Ordered Visual Feature Map in the
Drosophila Brain.” Neuron 110 (10): 1700–1711.e6.
Knott, Graham, Herschel Marchman, David Wall, and Ben Lich. 2008. “Serial Section
Scanning Electron Microscopy of Adult Brain Tissue Using Focused Ion Beam Milling.”
The Journal of Neuroscience: The Official Journal of the Society for Neuroscience 28
(12): 2959–64.
Kovács, István A., Dániel L. Barabási, and Albert-László Barabási. 2020. “Uncovering the
Genetic Blueprint of the C. Elegans Nervous System.” Proceedings of the National
Academy of Sciences of the United States of America 117 (52): 33570–77.
Kremer, Malte C., Christophe Jung, Sara Batelli, Gerald M. Rubin, and Ulrike Gaul. 2017.
“The Glia of the Adult Drosophila Nervous System.” Glia 65 (4): 606–38.
Kurmangaliyev, Yerbol Z., Juyoun Yoo, Javier Valdes-Aleman, Piero Sanfilippo, and S.
Lawrence Zipursky. 2020. “Transcriptional Programs of Circuit Assembly in the
Drosophila Visual System.” Neuron 108 (6): 1045–57.e6.
Lappalainen, Janne K., Fabian D. Tschopp, Sridhama Prakhya, Mason McGill, Aljoscha
Nern, Kazunori Shinomiya, Shin-Ya Takemura, Eyal Gruntman, Jakob H. Macke, and
Srinivas C. Turaga. 2023. “Connectome-Constrained Deep Mechanistic Networks

39
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Predict Neural Responses across the Fly Visual System at Single-Neuron Resolution.”
bioRxiv. https://doi.org/10.1101/2023.03.11.532232.
Lapraz, F., C. Boutres, C. Fixary-Schuster, B. R. De Queiroz, P. Y. Plaçais, D. Cerezo, F.
Besse, T. Préat, and S. Noselli. 2023. “Asymmetric Activity of NetrinB Controls Laterality
of the Drosophila Brain.” Nature Communications 14 (1): 1052.
Lee, Kisuk, Nicholas Turner, Thomas Macrina, Jingpeng Wu, Ran Lu, and H. Sebastian
Seung. 2019. “Convolutional Nets for Reconstructing Neural Circuits from Brain Images
Acquired by Serial Section Electron Microscopy.” Current Opinion in Neurobiology.
https://doi.org/10.1016/j.conb.2019.04.001.
Li, Cai-Hong, and Jun-Li Yang. 2020. “Wolfberry Extracts Inhibit Aβ1-42 Aggregation and
Rescue Memory Loss of AD Drosophila.” Food Science and Human Wellness 9 (1):
64–70.
Lichtman, Jeff W., and Winfried Denk. 2011. “The Big and the Small: Challenges of Imaging
the Brain’s Circuits.” Science 334 (6056): 618–23.
Lichtman, Jeff W., and Joshua R. Sanes. 2008. “Ome Sweet Ome: What Can the Genome
Tell Us about the Connectome?” Current Opinion in Neurobiology 18 (3): 346–53.
Li, Feng, Jack W. Lindsey, Elizabeth C. Marin, Nils Otto, Marisa Dreher, Georgia Dempsey,
Ildiko Stark, et al. 2020. “The Connectome of the Adult Drosophila Mushroom Body
Provides Insights into Function.” eLife 9 (December).
https://doi.org/10.7554/eLife.62576.
Lin, Andrew C., Alexei M. Bygrave, Alix de Calignon, Tzumin Lee, and Gero Miesenböck.
2014. “Sparse, Decorrelated Odor Coding in the Mushroom Body Enhances Learned
Odor Discrimination.” Nature Neuroscience 17 (4): 559–68.
Liu, Wendy W., and Rachel I. Wilson. 2013. “Glutamate Is an Inhibitory Neurotransmitter in
the Drosophila Olfactory System.” Proceedings of the National Academy of Sciences of
the United States of America 110 (25): 10294–99.
Loomba, Sahil, Jakob Straehle, Vijayan Gangadharan, Natalie Heike, Abdelrahman Khalifa,
Alessandro Motta, Niansheng Ju, et al. 2022. “Connectomic Comparison of Mouse and
Human Cortex.” Science 377 (6602): eabo0924.
Luan, Haojiang, Nathan C. Peabody, Charles R. Vinson, and Benjamin H. White. 2006.
“Refined Spatial Manipulation of Neuronal Function by Combinatorial Restriction of
Transgene Expression.” Neuron 52 (3): 425–36.
Lu, Jenny, Amir H. Behbahani, Lydia Hamburg, Elena A. Westeinde, Paul M. Dawson,
Cheng Lyu, Gaby Maimon, Michael H. Dickinson, Shaul Druckmann, and Rachel I.
Wilson. 2022. “Transforming Representations of Movement from Body- to World-Centric
Space.” Nature 601 (7891): 98–104.
Mabuchi, Yuta, Xinyue Cui, Lily Xie, Haein Kim, Tianxing Jiang, and Nilay Yapici. 2023a.
“GABA-Mediated Inhibition in Visual Feedback Neurons Fine-Tunes Drosophila Male
Courtship.” bioRxiv. https://doi.org/10.1101/2023.01.25.525544.
Mabuchi, Yuta, Xinyue Cui, Lily Xie, Haein Kim, Tianxing Jiang, and Nilay Yapici. 2023b.
“Visual Feedback Neurons Fine-Tune Drosophila Male Courtship via GABA-Mediated
Inhibition.” https://doi.org/10.2139/ssrn.4380792.
Macrina, T., K. Lee, R. Lu, N. L. Turner, J. Wu, and S. Popovych. 2021. “Petascale Neural
Circuit Reconstruction: Automated Methods.” bioRxiv.
https://www.biorxiv.org/content/10.1101/2021.08.04.455162.abstract.
Mao, Zhengmei, and Ronald L. Davis. 2009. “Eight Different Types of Dopaminergic
Neurons Innervate the Drosophila Mushroom Body Neuropil: Anatomical and
Physiological Heterogeneity.” Frontiers in Neural Circuits 3 (July): 5.
Marin, Elizabeth C., Billy J. Morris, Tomke Stuerner, Andrew S. Champion, Dominik
Krzeminski, Griffin Badalamente, Marina Gkantia, et al. 2023. “Systematic Annotation of
a Complete Adult Male Drosophila Nerve Cord Connectome Reveals Principles of
Functional Organisation.” bioRxiv. https://doi.org/10.1101/2023.06.05.543407.
Markov, Nikola T., Maria M. Ercsey-Ravasz, A. R. Ribeiro Gomes, Camille Lamy, Loic
Magrou, Julien Vezoli, Pierre Misery, et al. 2014. “A Weighted and Directed Interareal
Connectivity Matrix for Macaque Cerebral Cortex.” Cerebral Cortex 24 (1): 17–36.

40
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ma, Xiaoya, Xianguang Hou, Gregory D. Edgecombe, and Nicholas J. Strausfeld. 2012.
“Complex Brain and Optic Lobes in an Early Cambrian Arthropod.” Nature 490 (7419):
258–61.
McCarthy, Ellena V., Ying Wu, Tagide Decarvalho, Christian Brandt, Guan Cao, and Michael
N. Nitabach. 2011. “Synchronized Bilateral Synaptic Inputs to Drosophila Melanogaster
Neuropeptidergic Rest/arousal Neurons.” The Journal of Neuroscience: The Official
Journal of the Society for Neuroscience 31 (22): 8181–93.
McInnes, Leland, John Healy, and James Melville. 2018. “UMAP: Uniform Manifold
Approximation and Projection for Dimension Reduction.” arXiv [stat.ML]. arXiv.
http://arxiv.org/abs/1802.03426.
Meier, Matthias, and Alexander Borst. 2019. “Extreme Compartmentalization in a Drosophila
Amacrine Cell.” Current Biology: CB 29 (9): 1545–50.e2.
Meinertzhagen, Ian A. 2018. “Of What Use Is Connectomics? A Personal Perspective on the
Drosophila Connectome.” The Journal of Experimental Biology 221 (10).
https://doi.org/10.1242/jeb.164954.
Meissner, Geoffrey W., Aljoscha Nern, Zachary Dorman, Gina M. DePasquale, Kaitlyn
Forster, Theresa Gibney, Joanna H. Hausenfluck, et al. 2023. “A Searchable Image
Resource of Drosophila GAL4 Driver Expression Patterns with Single Neuron
Resolution.” eLife 12 (February). https://doi.org/10.7554/eLife.80660.
Mesulam, M. M. 1998. “From Sensation to Cognition.” Brain: A Journal of Neurology 121 ( Pt
6) (June): 1013–52.
MICrONS Consortium, J. Alexander Bae, Mahaly Baptiste, Agnes L. Bodor, Derrick Brittain,
Joann Buchanan, Daniel J. Bumbarger, et al. 2021. “Functional Connectomics Spanning
Multiple Areas of Mouse Visual Cortex.” bioRxiv.
https://doi.org/10.1101/2021.07.28.454025.
Molina-Obando, Sebastian, Juan Felipe Vargas-Fique, Miriam Henning, Burak Gür, T. Moritz
Schladt, Junaid Akhtar, Thomas K. Berger, and Marion Silies. 2019. “ON Selectivity in
the Drosophila Visual System Is a Multisynaptic Process Involving Both Glutamatergic
and GABAergic Inhibition.” eLife 8 (September). https://doi.org/10.7554/eLife.49373.
Murthy, Mala, Ila Fiete, and Gilles Laurent. 2008. “Testing Odor Response Stereotypy in the
Drosophila Mushroom Body.” Neuron 59 (6): 1009–23.
Mu, Shang, Szi-Chieh Yu, Nicholas L. Turner, Claire E. McKellar, Sven Dorkenwald, Forrest
Collman, Selden Koolman, et al. 2021. “3D Reconstruction of Cell Nuclei in a Full
Drosophila Brain.” bioRxiv. https://doi.org/10.1101/2021.11.04.467197.
Nern, A., B. D. Pfeiffer, and G. M. Rubin. 2015. “Optimized Tools for Multicolor Stochastic
Labeling Reveal Diverse Stereotyped Cell Arrangements in the Fly Visual System.”
Proceedings of the. https://www.pnas.org/content/112/22/E2967.short.
Nguyen, Tri M., Logan A. Thomas, Jeff L. Rhoades, Ilaria Ricchi, Xintong Cindy Yuan, Arlo
Sheridan, David G. C. Hildebrand, Jan Funke, Wade G. Regehr, and Wei-Chung Allen
Lee. 2023. “Structured Cerebellar Connectivity Supports Resilient Pattern Separation.”
Nature 613 (7944): 543–49.
Oh, Seung Wook, Julie A. Harris, Lydia Ng, Brent Winslow, Nicholas Cain, Stefan Mihalas,
Quanxin Wang, et al. 2014. “A Mesoscale Connectome of the Mouse Brain.” Nature 508
(7495): 207–14.
Otsuna, Hideo, and Kei Ito. 2006. “Systematic Analysis of the Visual Projection Neurons of
Drosophila Melanogaster. I. Lobula-Specific Pathways.” The Journal of Comparative
Neurology 497 (6): 928–58.
Özel, Mehmet Neset, Félix Simon, Shadi Jafari, Isabel Holguera, Yen-Chung Chen, Najate
Benhra, Rana Naja El-Danaf, et al. 2021. “Neuronal Diversity and Convergence in a
Visual System Developmental Atlas.” Nature 589 (7840): 88–95.
Pacheco, Diego A., Stephan Y. Thiberge, Eftychios Pnevmatikakis, and Mala Murthy. 2021.
“Auditory Activity Is Diverse and Widespread throughout the Central Brain of
Drosophila.” Nature Neuroscience 24 (1): 93–104.
Pfeiffer, Barret D., Teri-T B. Ngo, Karen L. Hibbard, Christine Murphy, Arnim Jenett, James
W. Truman, and Gerald M. Rubin. 2010. “Refinement of Tools for Targeted Gene

41
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Expression in Drosophila.” Genetics 186 (2): 735–55.

Phelps, Jasper S., David Grant Colburn Hildebrand, Brett J. Graham, Aaron T. Kuan, Logan
A. Thomas, Tri M. Nguyen, Julia Buhmann, et al. 2021. “Reconstruction of Motor Control
Circuits in Adult Drosophila Using Automated Transmission Electron Microscopy.” Cell
184 (3): 759–74.e18.
Popovych, Sergiy, Thomas Macrina, Nico Kemnitz, Manuel Castro, Barak Nehoran, Zhen
Jia, J. Alexander Bae, et al. 2022. “Petascale Pipeline for Precise Alignment of Images
from Serial Section Electron Microscopy.” bioRxiv.
https://doi.org/10.1101/2022.03.25.485816.
Prisco, Luigi, Stephan Hubertus Deimel, Hanna Yeliseyeva, André Fiala, and Gaia
Tavosanis. 2021. “The Anterior Paired Lateral Neuron Normalizes Odour-Evoked
Activity in the Drosophila Mushroom Body Calyx.” eLife 10 (December).
https://doi.org/10.7554/eLife.74172.
Repérant, J., R. Ward, D. Miceli, J. P. Rio, M. Médina, N. B. Kenigfest, and N. P. Vesselkin.
2006. “The Centrifugal Visual System of Vertebrates: A Comparative Analysis of Its
Functional Anatomical Organization.” Brain Research Reviews 52 (1): 1–57.
Sato, M., I. Bitter, M. A. Bender, A. E. Kaufman, and M. Nakajima. 2000. “TEASAR:
Tree-Structure Extraction Algorithm for Accurate and Robust Skeletons.” In .
https://doi.org/10.1109/pccga.2000.883951.
Schaffer, Evan S., Neeli Mishra, Matthew R. Whiteway, Wenze Li, Michelle B. Vancura,
Jason Freedman, Kripa B. Patel, et al. 2021. “Flygenvectors: The Spatial and Temporal
Structure of Neural Activity across the Fly Brain.” bioRxiv.
https://doi.org/10.1101/2021.09.25.461804.
Scheffer, Louis K., C. Shan Xu, Michal Januszewski, Zhiyuan Lu, Shin-Ya Takemura,
Kenneth J. Hayworth, Gary B. Huang, et al. 2020. “A Connectome and Analysis of the
Adult Drosophila Central Brain.” eLife 9 (September).
https://doi.org/10.7554/eLife.57443.
Schlegel, P., A. S. Bates, T. Stürner, and S. R. Jagannathan. 2021. “Information Flow, Cell
Types and Stereotypy in a Full Olfactory Connectome.” eLife.
https://elifesciences.org/articles/66018.
Schlegel, Philipp, Yijie Yin, Alexander S. Bates, Sven Dorkenwald, Katharina Eichler, Paul
Brooks, Daniel S. Han, et al. 2023. “A Consensus Cell Type Atlas from Multiple
Connectomes Reveals Principles of Circuit Stereotypy and Variation.” bioRxiv.
Schneider-Mizell, Casey M., Stephan Gerhard, Mark Longair, Tom Kazimiers, Feng Li,
Maarten F. Zwart, Andrew Champion, et al. 2016. “Quantitative Neuroanatomy for
Connectomics in Drosophila.” eLife 5 (March). https://doi.org/10.7554/eLife.12059.
Schoofs, Andreas, Anton Miroschnikow, Philipp Schlegel, Ingo Zinke, Casey M.
Schneider-Mizell, Albert Cardona, and Michael J. Pankratz. 2023. “Serotonergic
Reinforcement of a Complete Swallowing Circuit.” bioRxiv.
https://doi.org/10.1101/2023.05.26.542464.
Schretter, Catherine E., Yoshinori Aso, Alice A. Robie, Marisa Dreher, Michael-John Dolan,
Nan Chen, Masayoshi Ito, et al. 2020. “Cell Types and Neuronal Circuitry Underlying
Female Aggression in Drosophila.” eLife 9 (November).
https://doi.org/10.7554/eLife.58942.
Schüz, A., and G. Palm. 1989. “Density of Neurons and Synapses in the Cerebral Cortex of
the Mouse.” The Journal of Comparative Neurology 286 (4): 442–55.
Seung, Sebastian. 2012. Connectome: How the Brain’s Wiring Makes Us Who We Are.
HMH.
Shapson-Coe, A., M. Januszewski, D. R. Berger, and A. Pope. 2021. “A Connectomic Study
of a Petascale Fragment of Human Cerebral Cortex.” bioRxiv.
https://www.biorxiv.org/content/10.1101/2021.05.29.446289v1.abstract.
Sherer, Lewis M., Elizabeth Catudio Garrett, Hannah R. Morgan, Edmond D. Brewer, Lucy
A. Sirrs, Harold K. Shearin, Jessica L. Williams, Brian D. McCabe, R. Steven Stowers,
and Sarah J. Certel. 2020. “Octopamine Neuron Dependent Aggression Requires
dVGLUT from Dual-Transmitting Neurons.” PLoS Genetics 16 (2): e1008609.

42
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Shih, Chi-Tin, Olaf Sporns, Shou-Li Yuan, Ta-Shun Su, Yen-Jen Lin, Chao-Chun Chuang,
Ting-Yuan Wang, Chung-Chuang Lo, Ralph J. Greenspan, and Ann-Shyn Chiang. 2015.
“Connectomics-Based Analysis of Information Flow in the Drosophila Brain.” Current
Biology: CB 25 (10): 1249–58.
Shinomiya, Kazunori, Gary Huang, Zhiyuan Lu, Toufiq Parag, C. Shan Xu, Roxanne Aniceto,
Namra Ansari, et al. 2019. “Comparisons between the ON- and OFF-Edge Motion
Pathways in the Drosophila Brain.” eLife 8 (January).
https://doi.org/10.7554/eLife.40025.
Shinomiya, Kazunori, Aljoscha Nern, Ian A. Meinertzhagen, Stephen M. Plaza, and Michael
B. Reiser. 2022. “Neuronal Circuits Integrating Visual Motion Information in Drosophila
Melanogaster.” Current Biology: CB 32 (16): 3529–44.e2.
Shiu, Philip K., Gabriella R. Sterne, Stefanie Engert, Barry J. Dickson, and Kristin Scott.
2022. “Taste Quality and Hunger Interactions in a Feeding Sensorimotor Circuit.” eLife
11 (July). https://doi.org/10.7554/eLife.79887.
Shiu, Philip K., Gabriella R. Sterne, Nico Spiller, Romain Franconville, Andrea Sandoval,
Joie Zhou, Neha Simha, et al. 2023. “A Leaky Integrate-and-Fire Computational Model
Based on the Connectome of the Entire Adult Drosophila Brain Reveals Insights into
Sensorimotor Processing.” bioRxiv. https://doi.org/10.1101/2023.05.02.539144.
Sporns, Olaf, Giulio Tononi, and Rolf Kötter. 2005. “The Human Connectome: A Structural
Description of the Human Brain.” PLoS Computational Biology 1 (4): e42.
Stange, G., S. Stowe, J. S. Chahl, and A. Massaro. 2002. “Anisotropic Imaging in the
Dragonfly Median Ocellus: A Matched Filter for Horizon Detection.” Journal of
Comparative Physiology. A, Neuroethology, Sensory, Neural, and Behavioral Physiology
188 (6): 455–67.
Stark, W. S., R. Sapp, and S. D. Carlson. 1989. “Ultrastructure of the Ocellar Visual System
in Normal and Mutant Drosophila Melanogaster.” Journal of Neurogenetics 5 (2):
127–53.
Sterne, Gabriella R., Hideo Otsuna, Barry J. Dickson, and Kristin Scott. 2021. “Classification
and Genetic Targeting of Cell Types in the Primary Taste and Premotor Center of the
Adult Drosophila Brain.” eLife 10 (September). https://doi.org/10.7554/eLife.71679.
Suver, Marie P., Ainul Huda, Nicole Iwasaki, Steve Safarik, and Michael H. Dickinson. 2016.
“An Array of Descending Visual Interneurons Encoding Self-Motion in Drosophila.” The
Journal of Neuroscience: The Official Journal of the Society for Neuroscience 36 (46):
11768–80.
Swanson, L. W. 1995. “Mapping the Human Brain: Past, Present, and Future.” Trends in
Neurosciences 18 (11): 471–74.
Takemura, Shin-Ya, Arjun Bharioke, Zhiyuan Lu, Aljoscha Nern, Shiv Vitaladevuni, Patricia
K. Rivlin, William T. Katz, et al. 2013. “A Visual Motion Detection Circuit Suggested by
Drosophila Connectomics.” Nature 500 (7461): 175–81.
Takemura, Shin-Ya, Kenneth J. Hayworth, Gary B. Huang, Michal Januszewski, Zhiyuan Lu,
Elizabeth C. Marin, Stephan Preibisch, et al. 2023. “A Connectome of the Male
Drosophila Ventral Nerve Cord.” bioRxiv. https://doi.org/10.1101/2023.06.05.543757.
Takemura, S., A. Nern, D. B. Chklovskii, and L. K. Scheffer. 2017. “The Comprehensive
Connectome of a Neural Substrate for ‘ON’motion Detection in Drosophila.” eLife.
https://cdn.elifesciences.org/articles/24394/elife-24394-v2.pdf.
Task, Darya, Chun-Chieh Lin, Alina Vulpe, Ali Afify, Sydney Ballou, Maria Brbic, Philipp
Schlegel, et al. 2022. “Chemoreceptor Co-Expression in Drosophila Melanogaster
Olfactory Neurons.” eLife 11 (April). https://doi.org/10.7554/eLife.72599.
Turaga, Srinivas C., Joseph F. Murray, Viren Jain, Fabian Roth, Moritz Helmstaedter, Kevin
Briggman, Winfried Denk, and H. Sebastian Seung. 2010. “Convolutional Networks Can
Learn to Generate Affinity Graphs for Image Segmentation.” Neural Computation 22 (2):
511–38.
Turner, Nicholas L., Thomas Macrina, J. Alexander Bae, Runzhe Yang, Alyssa M. Wilson,
Casey Schneider-Mizell, Kisuk Lee, et al. 2022. “Reconstruction of Neocortex:
Organelles, Compartments, Cells, Circuits, and Activity.” Cell, February.

43
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

https://doi.org/10.1016/j.cell.2022.01.023.
Tuthill, John C., Aljoscha Nern, Gerald M. Rubin, and Michael B. Reiser. 2014. “Wide-Field
Feedback Neurons Dynamically Tune Early Visual Processing.” Neuron 82 (4): 887–95.
Varshney, Lav R., Beth L. Chen, Eric Paniagua, David H. Hall, and Dmitri B. Chklovskii.
2011. “Structural Properties of the Caenorhabditis Elegans Neuronal Network.” PLoS
Computational Biology 7 (2): e1001066.
Vogt, Katrin, Yoshinori Aso, Toshihide Hige, Stephan Knapek, Toshiharu Ichinose, Anja B.
Friedrich, Glenn C. Turner, Gerald M. Rubin, and Hiromu Tanimoto. 2016. “Direct Neural
Pathways Convey Distinct Visual Information to Drosophila Mushroom Bodies.” eLife 5
(April). https://doi.org/10.7554/eLife.14009.
Waddell, S., J. D. Armstrong, T. Kitamoto, K. Kaiser, and W. G. Quinn. 2000. “The Amnesiac
Gene Product Is Expressed in Two Neurons in the Drosophila Brain That Are Critical for
Memory.” Cell 103 (5): 805–13.
Wang, Kaiyu, Fei Wang, Nora Forknall, Tansy Yang, Christopher Patrick, Ruchi Parekh, and
Barry J. Dickson. 2021. “Neural Circuit Mechanisms of Sexual Receptivity in Drosophila
Females.” Nature 589 (7843): 577–81.
White, J. G., E. Southgate, J. N. Thomson, and S. Brenner. 1986. “The Structure of the
Nervous System of the Nematode Caenorhabditis Elegans.” Philosophical Transactions
of the Royal Society of London. Series B, Biological Sciences 314 (1165): 1–340.
Winding, Michael, Benjamin D. Pedigo, Christopher L. Barnes, Heather G. Patsolic,
Youngser Park, Tom Kazimiers, Akira Fushiki, et al. 2023. “The Connectome of an
Insect Brain.” Science 379 (6636): eadd9330.
Witvliet, Daniel, Ben Mulcahy, James K. Mitchell, Yaron Meirovitch, Daniel R. Berger,
Yuelong Wu, Yufang Liu, et al. 2021. “Connectomes across Development Reveal
Principles of Brain Maturation.” Nature 596 (7871): 257–61.
Wu, Ming, Aljoscha Nern, W. Ryan Williamson, Mai M. Morimoto, Michael B. Reiser,
Gwyneth M. Card, and Gerald M. Rubin. 2016. “Visual Projection Neurons in the
Drosophila Lobula Link Feature Detection to Distinct Behavioral Programs.” eLife 5
(December). https://doi.org/10.7554/eLife.21022.
Xu, C. Shan, Kenneth J. Hayworth, Zhiyuan Lu, Patricia Grob, Ahmed M. Hassan, José G.
García-Cerdán, Krishna K. Niyogi, Eva Nogales, Richard J. Weinberg, and Harald F.
Hess. 2017. “Enhanced FIB-SEM Systems for Large-Volume 3D Imaging.” eLife 6
(May). https://doi.org/10.7554/eLife.25916.
Yildirim, Kerem, Johanna Petri, Rita Kottmeier, and Christian Klämbt. 2019. “Drosophila Glia:
Few Cell Types and Many Conserved Functions.” Glia 67 (1): 5–26.
Yoo, Juyoun, Mark Dombrovski, Parmis Mirshahidi, Aljoscha Nern, Samuel A. LoCascio, S.
Lawrence Zipursky, and Yerbol Z. Kurmangaliyev. 2023. “Identifying Determinants of
Synaptic Specificity by Integrating Connectomes and Transcriptomes.” bioRxiv.
https://doi.org/10.1101/2023.04.03.534791.
Yu, Wan-Qing, Rana N. El-Danaf, Haruhisa Okawa, Justin M. Pacholec, Ulf Matti, Karin
Schwarz, Benjamin Odermatt, et al. 2018. “Synaptic Convergence Patterns onto Retinal
Ganglion Cells Are Preserved despite Topographic Variation in Pre- and Postsynaptic
Territories.” Cell Reports 25 (8): 2017–26.e3.
Zhao, Arthur, Eyal Gruntman, Aljoscha Nern, Nirmala A. Iyer, Edward M. Rogers, Sanna
Koskela, Igor Siwanowicz, et al. 2022. “Eye Structure Shapes Neuron Function in
Drosophila Motion Vision.” bioRxiv. https://doi.org/10.1101/2022.12.14.520178.
Zheng, Zhihao, J. Scott Lauritzen, Eric Perlman, Camenzind G. Robinson, Matthew Nichols,
Daniel Milkie, Omar Torrens, et al. 2018. “A Complete Electron Microscopy Volume of
the Brain of Adult Drosophila Melanogaster.” Cell 174 (3): 730–43.e22.
Zheng, Zhihao, Feng Li, Corey Fisher, Iqbal J. Ali, Nadiya Sharifi, Steven Calle-Schuler,
Joseph Hsu, et al. 2022. “Structured Sampling of Olfactory Input by the Fly Mushroom
Body.” Current Biology: CB 32 (15): 3334–49.e6.
Zingg, Brian, Houri Hintiryan, Lin Gou, Monica Y. Song, Maxwell Bay, Michael S. Bienkowski,
Nicholas N. Foster, et al. 2014. “Neural Networks of the Mouse Neocortex.” Cell 156 (5):
1096–1111.

44
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Extended Data Figures

Ext.-Figure 1-1. Neuropils of the fly brain.

45
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 1-2. Completeness and accuracy of FlyWire’s reconstruction. (a) shows the result of
our evaluation of proofread segments in the central brain. Experts attempted further proofreading of
826 neurons. We computed volumetric overlaps between the original and the final segment to
calculate precision, recall, and F1 Scores. (b) Examples (top: before, bottom: after) of the changes
made during further proofreading for a neuron scoring an F1-Score of 0.936. Arrows highlight
locations that changed. (c,d) For each neuropil, we quantified what fraction of the synapses within it
are pre- and postsynaptically attached to a proofread segment. (c) displays the distribution for
presynaptic attachment and (d) the distribution for postsynaptic attachment. (e, f, g) Comparisons
between FlyWire’s reconstruction and the hemibrain were made for overlapping neuropils. Dots
represent neuropils and are colored according to Ext. Data Fig. 1-1. (e) Comparison of the number of
automatically detected synapses. The axes are log-transformed. (f) Comparison of post-synaptic
completion rates and (g) pre-synaptic completion rate. The axes are truncated.

46
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 1-3. Completion rates by neuropil.

47
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 1-4. Trachea and glia cells. (a) Rendering of all trachea segments in the FlyWire
dataset. (b) Rendering of some reconstructed glia cells in the FlyWire dataset. At the time of writing,
only a subset of the glia cells, with bias towards the central brain, have been proofread and labeled.
Scale bar: 100 µm; insets: 10 µm.

48
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Extended Data Figure 3-1. Measurements of neuron size. Colored markers refer to neurons in Fig.
3b. (a) Neuron path lengths of intrinsic neurons, (b) afferent neurons, and (b) efferent neurons by
super class. (d) Volumes of intrinsic neurons, (e) afferent neurons, and (f) efferent neurons by super
class. (g) Comparisons of path lengths and number of incoming and outgoing synapses. (h) For
intrinsic neurons, comparisons of the in- and out-degrees with the number of incoming and outgoing
synapses. Every dot is a neuron. (i) Comparison of average connection strengths (synapses per
connection) with the number of synapses. Every dot is a neuron. (j) In- and out-degree distributions by
neurotransmitter type. (k) Neuron path lengths by neurotransmitter type.

49
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 4-1. Neuropil-neuropil projection maps. (a) Projection maps produced as in Fig. 4a
limited to connections from cholinergic, (b) GABAergic, and (c) glutamatergic neurons. (d) The
difference between the putative excitatory (acetylcholine) and the putative inhibitory (GABA,
glutamate) projection maps.

50
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 4-2. Neuropil-neuropil projections compared between hemispheres. Each dot is a
neuropil-neuropil projection in one hemisphere and the axes show the fractional weights as calculated
in Fig. 4a,b. Red dots are comparisons between the same neuropils in different hemispheres (e.g.
AMMC(L) -> VLP(L) vs AMMC(R) -> VLP(R). (a) Comparison of projections between neuropils in both
hemispheres and between hemispheres. (b) Comparisons of projections with the center neuropils. (c)
Comparisons of projections between ipsilateral and contralateral neuropil projections.

51
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 4-3. Input side analysis. We assigned postsynaptic locations to either the center region
or the left or right hemisphere. (a-g) For each super class, the fraction of synapses in the left vs right
hemisphere is shown for those neurons receiving most of their neurons laterally (top plot). The lower
plot shows the fraction of synapses in the center vs the lateral regions for all neurons. (h) Each
neuron was assigned to the side where it received most of its inputs.

52
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 4-4. Neurons on the midline with dendrites in both hemispheres. (a) All symmetric
neurons with a cell body on the midline (N=106). (b-e) examples of individual neurons. Scale bar: 100
µm, inset: 50 µm

53
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Ext. Figure 4-5. Renderings of neurons for each cross-hemisphere category (up to 3,000
neurons rendered per group). Scale bar: 100 µm

54
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Extended Data Figure 6-1. Percentile ranks for every modality. (a) For each sensory modality we
used the traversal distances to establish a neuron ranking. Each panel shows the distributions of
neurons of each super class within the sensory modality specific rankings. (b) Same as in (a) for the
fast neurotransmitters. (c) Neurons in the UMAP plot are colored by the rank order in which they are
reached from a given seed neuron set. Red neurons are reached earlier than blue neurons.

55
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Extended Data Figure 6-2. Rank-based UMAP projection and neuropils. (a) Every neuron in the
central brain was assigned to the neuropil where it received the most synapses. Every dot is then
colored by the assigned neuropil (Ext. Data Fig. 1-1). (b) Same as in a but limited to the central
complex neurons. Neurons in the central complex with an assigned neuropil other than the ones
shown are shown are colored black.

56
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Extended Data Figure 7-1. Ocellar circuit. (a) Renderings of all neurons (excluding the
photoreceptors) with arbors in the ocellar ganglion. “Information flow” from pre- and postsynapses is
indicated by arrows along the arbors. (b) Overview of the three ocelli (left, medial, right) which are
positioned on the top of the head. Photoreceptors from each ocellus project to a specific subregion of
the ocellar ganglion which are separated by glia (marked with black lines on the EM). (c) Top view of
the dendritic arbors within the ocellar ganglion of each DNp28 (brown) and OCG01 (blue: cholinergic,
green: glutamatergic). The render on the lower shows all 12 OCG01s and 2 DNp28s. Each other
render shows one neuron in color and all others in the background in gray for reference. (d)
Comparison of number of synapses from OCG01 neurons and visual projection neurons onto
descending neurons. (e) Connectivity matrix for connections between ocellar centrifugal neurons and
ocellar projection neurons. (f) Inputs to ocellar centrifugal neurons by neuropil. Scale bars: 100 µm
(a), 20 µm (c)

57
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Supplemental Information 1: Neuroglancer Interface and Lightbulb. (a) FlyWire's lightbulb

menu displays the proofreading and annotation status of every segment (red box). (b) It is
color coded for easy reference (yellow: cell has not been declared complete; purple:
complete but not labeled; green: proofread and labeled; black: out of date segmentation). (c)
Users can load cell identification directly within the FlyWire editor, perform basic connectivity
analysis, and view a cell's edit history.

58
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Supplemental Information 2: Bilateral optic lobe neurons. On the left: putative LC14 (top) and
putative LC14b (bottom). Scale bar: 100 µm

59
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Supplemental Information 3: Mechanosensory neuron subtypes. Scale bar: 100 µm

60
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Supplemental Information 4: Distribution of community annotations by cell type

61
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Name Lab Affiliation total edits

Doug Bland Mala Murthy Lab, Sebastian Seung Lab 247,366

Zairene Lenizo Mala Murthy Lab, Sebastian Seung Lab 122,932
Nseraf Flyers (citizen scientist) 110,082
Kyle Patrick Willie Mala Murthy Lab, Sebastian Seung Lab 83,635
Nashra Hadjerol Mala Murthy Lab, Sebastian Seung Lab 81,700
Austin T Burke Mala Murthy Lab, Sebastian Seung Lab 80,647
Ryan Willie Mala Murthy Lab, Sebastian Seung Lab 74,951
John Anthony Ocho Mala Murthy Lab, Sebastian Seung Lab 67,436
Arti Yadav Greg Jefferis Lab 59,394
Joshua Bañez Mala Murthy Lab, Sebastian Seung Lab 59,014
Shirleyjoy Serona Mala Murthy Lab, Sebastian Seung Lab 55,859
Yijie Yin Greg Jefferis Lab 50,805
Rey Adrian Candilada Mala Murthy Lab, Sebastian Seung Lab 50,492
Jet Ivan Dolorosa Mala Murthy Lab, Sebastian Seung Lab 46,454
Mendell Lopez Mala Murthy Lab, Sebastian Seung Lab 43,885
Ariel Dagohoy Mala Murthy Lab, Sebastian Seung Lab 43,124
Regine Salem Mala Murthy Lab, Sebastian Seung Lab 42,181
Griffin Badalamente Greg Jefferis Lab 41,195
Remer Tancontian Mala Murthy Lab, Sebastian Seung Lab 41,130
Nelsie Panes Mala Murthy Lab, Sebastian Seung Lab 39,801
Laia Serratosa Capdevila Greg Jefferis Lab, Rachel Wilson Lab 39,798
Kendrick Joules Vinson Mala Murthy Lab, Sebastian Seung Lab 38,646
Anjali Pandey Greg Jefferis Lab 37,344
Darrel Jay Akiatan Mala Murthy Lab, Sebastian Seung Lab 36,987
Dustin Garner Sung Soo Kim Lab 36,209
Ben Silverman Mala Murthy Lab, Sebastian Seung Lab 34,676
Dharini Sapkal Greg Jefferis Lab 31,244
Shaina Mae Monungolh Mala Murthy Lab, Sebastian Seung Lab 28,721
Jay Gager Mala Murthy Lab, Sebastian Seung Lab 28,317
Krzysztof Kruk Flyers (citizen scientist) 27,486
Miguel Albero Mala Murthy Lab, Sebastian Seung Lab 26,831
Katharina Eichler Greg Jefferis Lab, Seeds Hamepl Lab 24,481
Zeba Vohra Greg Jefferis Lab 24,431
Emil Kind Mathias Wernet Lab 24,052
Varun Sane Greg Jefferis Lab 23,762
annkri (Anne Kristiansen) Flyers (citizen scientist) 22,020
Chitra Nair Greg Jefferis Lab 21,696
Márcia dos Santos Greg Jefferis Lab 21,209
Dhwani Patel Greg Jefferis Lab 20,389
Imaan F. M. Tamimi Greg Jefferis Lab 19,937
Michelle Darapan Pantujan Mala Murthy Lab, Sebastian Seung Lab 18,846
James Hebditch Mala Murthy Lab, Sebastian Seung Lab 18,719

62
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Alexandre Javier Greg Jefferis Lab 17,813

Rashmita Rana Greg Jefferis Lab 17,547
Bhargavi Parmar Greg Jefferis Lab 17,133
Merlin Moore Mala Murthy Lab, Sebastian Seung Lab 16,590
Mark Lloyd Pielago Mala Murthy Lab, Sebastian Seung Lab 16,566
Allien Mae Gogo Mala Murthy Lab, Sebastian Seung Lab 16,243
Mark Larson Wei-Chung Lee Lab 15,692
Joseph Hsu Greg Jefferis Lab, Scott Waddell Lab 15,496
Jacquilyn Laude Mala Murthy Lab, Sebastian Seung Lab 14,764
Itisha Joshi Greg Jefferis Lab 14,717
Chereb Martinez Mala Murthy Lab, Sebastian Seung Lab 14,638
Dhara Kakadiya Greg Jefferis Lab 14,544
John David Asis Mala Murthy Lab, Sebastian Seung Lab 14,202
Clyde Angelo Lim Mala Murthy Lab, Sebastian Seung Lab 14,139
Alvin Josh Mandahay Mala Murthy Lab, Sebastian Seung Lab 14,112
Thomas Stocks Flyers (citizen scientist) 13,683
AzureJay (Jaime Skelton) Flyers (citizen scientist) 13,540
Marchan Manaytay Mala Murthy Lab, Sebastian Seung Lab 13,244
Kaushik Parmar Greg Jefferis Lab 13,136
Philipp Schlegel Greg Jefferis Lab 12,750
Philip Lenard Ampo Mala Murthy Lab, Sebastian Seung Lab 12,572
Daril Bautista Mala Murthy Lab, Sebastian Seung Lab 12,539
Irene Salgarella Greg Jefferis Lab 12,475
John Clyde Saguimpa Mala Murthy Lab, Sebastian Seung Lab 11,391
Chan Hyuk Kang Jinseop Kim Lab 11,129
Markus William Pleijzier Greg Jefferis Lab 10,498
Marina Gkantia Greg Jefferis Lab 10,140
Jansen Seguido Mala Murthy Lab, Sebastian Seung Lab 9,979
Jinmook Kim Jinseop Kim Lab 9,879
Quinn Vanderbeck Rachel Wilson Lab 8,845
Cathy Pilapil Mala Murthy Lab, Sebastian Seung Lab 8,738
Yashvi Patel Greg Jefferis Lab 8,485
Eva Munnelly Greg Jefferis Lab 8,102
Olivia Sato Wei-Chung Lee Lab 8,055
Siqi Fang Greg Jefferis Lab 7,981
Paul Brooks Greg Jefferis Lab 6,838
Claire E. McKellar Mala Murthy Lab, Sebastian Seung Lab 6,802
Christopher Dunne Greg Jefferis Lab 6,307
Mai Bui Ken Colodner Lab 6,228
JousterL (Matthew Lichtenberger) Flyers (citizen scientist) 5,883
edmark tamboboy Mala Murthy Lab, Sebastian Seung Lab 5,801
Mareike Selcho Mareike Selcho Lab 5,565
Lucia Kmecova Seeds Hampel Lab 5,539
Katie Molloy Wei-Chung Lee Lab 5,492

63
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Alexis E Santana-Cruz Seeds Hampel Lab 5,274

Janice Salocot Mala Murthy Lab, Sebastian Seung Lab 5,133
Steven Calle Seeds Hampel Lab 4,922
Kfay Flyers (citizen scientist) 4,886
Seongbong Yu Jinseop Kim Lab 4,832
Arzoo Diwan Greg Jefferis Lab 4,787
Monika Patel Greg Jefferis Lab 4,482
Gregory S.X.E. Jefferis Greg Jefferis Lab 4,472
Sarah Morejohn Mala Murthy Lab, Sebastian Seung Lab 4,090
Sanna Koskela Michael Reiser Lab 3,822
bl4ckscor3 (Daniel Lehmann) Flyers (citizen scientist) 3,735
Celia David Mala Murthy Lab, Sebastian Seung Lab 3,611
Sangeeta Sisodiya Greg Jefferis Lab 3,493
Tansy Yang Janelia Tracers 3,422
Selden Koolman Mala Murthy Lab, Sebastian Seung Lab 3,384
Christa Baker Mala Murthy Lab 3,381
Szi-chieh Yu Mala Murthy Lab, Sebastian Seung Lab 3,376
Gerit A. Linneweber Gerit Linneweber Lab 3,237
Amalia Braun Alexander Borst Lab 3,125
Sky Cho Ken Colodner Lab 2,972
Wolf Huetteroth Wolf Huetteroth Lab 2,830
Brian Reicher Wei-Chung Lee Lab 2,794
TR77 Flyers (citizen scientist) 2,775
Marlon Blanquart Greg Jefferis Lab 2,662
Farzaan Salman Andrew Dacks Lab 2,524
Hyungjun Choi Jinseop Kim Lab 2,373
Li Guo Julie Simpson Lab 2,095
Forrest Collman Allen Institute Connectomics 2,016
Marissa Sorek Mala Murthy Lab, Sebastian Seung Lab 2,007
Joanna Eckhardt Mala Murthy Lab 1,995
Alisa Poh Barry Dickson Lab 1,922
Marina Lin Ken Colodner Lab 1,920
Stefanie Hampel Seeds Hampel Lab 1,645
Wes Murfin Citizen Scientist 1,578
Peter Gibb Rachel Wilson Lab 1,448
Zhihao Zheng Sebastian Seung Lab 1,421
Nidhi Patel Greg Jefferis Lab 1,394
Lucy Houghton Sung Soo Kim Lab 1,357
Devon Jones Mala Murthy Lab, Sebastian Seung Lab 1,295
Annalena Oswald Marion Silies Lab 1,187
Lucas Encarnacion-Rivera Mala Murthy Lab 1,164
Akanksha Jadia Greg Jefferis Lab 1,141
Leonie Walter Mathias Wernet Lab 1,102
Nik Drummond Alexander Borst Lab 1,099

64
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Xin Zhong Mathias Wernet Lab 1,083

Benjamin Gorko Sung Soo Kim Lab 1,064
Fernando J Figueroa Santiago Seeds Hampel Lab 1,049
István Taisz Greg Jefferis Lab 1,043
Urja Verma Greg Jefferis Lab 1,033
Ibrahim Tastekin Carlos Ribeiro Lab 1,025
Sandeep Kumar Mala Murthy Lab 987
Yuta Mabuchi Nilay Yapici Lab 963
Nick Byrne Wei-Chung Lee Lab 951
Edda Kunze Gerit Linneweber Lab 907
Thomas Crahan Sung Soo Kim Lab 901
Hewhoamareismyself (Ryan Margossian) Flyers (citizen scientist) 874
Iliyan Georgiev Flyers (citizen scientist) 825
Fabianna Szorenyi Seeds Hampel Lab 817
Tomke Stuerner Greg Jefferis Lab 736
Atsuko Adachi Richard Mann Lab, Rudy Behnia Lab 695
Minsik Yun Young-Joon Kim Lab 625
Andrearwen Flyers (citizen scientist) 607
Robert Turnbull Greg Jefferis Lab 586
Alexander Thomson Janelia Tracers, Michael Reiser Lab 527
a5hm0r Flyers (citizen scientist) 516
Sebastian Molina-Obando Marion Silies Lab 470
Connor Laughland Janelia Tracers, Michael Reiser Lab 469
Suchetana B. Dutta Bassem Hassan Lab 458
Paula Guiomar Alarcón de Antón Mathias Wernet Lab 426
Patricia Pujols Seeds Hampel Lab 423
Binglin Huang Sung Soo Kim Lab 423
Kenneth J. Colodner Ken Colodner Lab 421
Isabel Haber Rachel Wilson Lab 392
Albert Lin Mala Murthy Lab 362
Alexander Shakeel Bates Greg Jefferis Lab, Rachel Wilson Lab 340
Daniel T. Choe Jinseop Kim Lab 340
Veronika Lukyanova Jenny Read Lab 337
Marta Costa Greg Jefferis Lab 334
Maria Ioannidou Marion Silies 332
Jonas Chojetzki Marion Silies 331
Zequan Liu Xueying "Snow" Wang (Lichtman Lab) 317
Haley Croke Katie von Reyn Lab 308
Gizem Sancer Mathias Wernet Lab 308
Tatsuo Okubo Rachel Wilson Lab 306
Miriam A. Flynn Janelia Tracers, Michael Reiser Lab 297
Meghan Laturney Kristin Scott Lab 274
Benjamin Bargeron Salil Bidaye Lab 273
Davi D. Bock Davi Bock Lab 255

65
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Hyunsoo Yim Jinseop Kim Lab 240

Anh Duc Le Denise Garcia Lab 237
Seungyun Yu Jinseop Kim Lab 224
Yeonju Nam Jinseop Kim Lab 221
Mavil Flyers (citizen scientist) 217

Eleni Samara 213

Audrey Francis Gaby Maimon Lab 196
Jesse Gayk Greg Jefferis Lab 195
Zepeng Yao Kristin Scott Lab 194
Sommer S. Huntress Ken Colodner Lab 192
Carolina Manyari-Diaz Salil Bidaye Lab 191
Raquel Barajas Carlos Ribeiro Lab 186
Mindy Kim Wei-Chung Lee Lab 185
Burak Gür Marion Silies Lab 182
Nils Reinhard Charlotte Helfrich-Forster Lab 177
Amanda Abusaif Kristin Scott Lab 176
Anna Li Rachel Wilson Lab 173
Sven Dorkenwald Sebastian Seung Lab 169
Fred W Wolf Fred Wolf Lab 163
Lena Lörsch Marion Silies Lab 159
Keehyun Park Jinseop Kim Lab 155
Xinyue Cui Nilay Yapici Lab 152
Haein Kim Nilay Yapici Lab 145
Alan Mathew Greg Jefferis Lab 141
Taewan Kim Jinseop Kim Lab 135
Guan-ting Wu National Hualien Senior High School 124
Margarida Brotas Eugenia Chiappe Lab 112
Cheng-hao Zhang National Hualien Senior High School 109
Philip K. Shiu Kristin Scott Lab 108
Shanice Bailey Greg Jefferis Lab 102

Supplementary Table 1: Number of proofreading edits by consortium members. Only members

with ≥100 edits are shown.

66
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Name Lab Affiliation total Labels

Volker Hartenstein Volker Hartenstein Lab 13,762

Alexander Shakeel Bates Greg Jefferis Lab, Rachel Wilson Lab 11,260

Krzysztof Kruk Flyers (citizen scientist) 11,138

Sven Dorkenwald Sebastian Seung Lab 6,375

Katharina Eichler Greg Jefferis Lab, Seeds Hampel Lab 6,366

Philipp Schlegel Greg Jefferis Lab 3,751

David Deutsch Mala Murthy Lab 2,549

Kaiyu Wang Barry Dickson Lab 2,443

Yijie Yin Greg Jefferis Lab 2,354

Stefanie Hampel Seeds Hampel Lab 2,129

annkri (Anne Kristiansen) Flyers (citizen scientist) 1,871

Dustin Garner Sung Soo Kim Lab 1,782

Wolf Huetteroth Wolf Huetteroth Lab 1,409

AzureJay (Jaime Skelton) Flyers (citizen scientist) 1,303

Amalia Braun Alexander Borst Lab 1,104

Austin T Burke Mala Murthy Lab, Sebastian Seung Lab 1,013

Gizem Sancer Mathias Wernet Lab 942

Jenna Joroff Wei-Chung Lee Lab 900

Gregory S.X.E. Jefferis Greg Jefferis Lab 716

Christa Baker Mala Murthy Lab 622

Claire E. McKellar Mala Murthy Lab, Sebastian Seung Lab 612

Markus William Pleijzier Greg Jefferis Lab 541

Christopher Dunne Greg Jefferis Lab 517

Márcia dos Santos Greg Jefferis Lab 448

Varun Sane Greg Jefferis Lab 442

Quinn Vanderbeck Rachel Wilson Lab 424

Lucia Kmecova Seeds Hampel Lab 412

Steven Calle Seeds Hampel Lab 408

Rey Adrian Candilada Mala Murthy Lab, Sebastian Seung Lab 364

Sebastian Molina-Obando Marion Silies Lab 347

Philip K. Shiu Kristin Scott Lab 321

Eva Munnelly Greg Jefferis Lab 311

Remer Tancontian Mala Murthy Lab, Sebastian Seung Lab 308

Doug Bland Mala Murthy Lab, Sebastian Seung Lab 306

Ariel Dagohoy Mala Murthy Lab, Sebastian Seung Lab 306

Joshua Bañez Mala Murthy Lab, Sebastian Seung Lab 301

Marina Gkantia Greg Jefferis Lab 300

Jet Ivan Dolorosa Mala Murthy Lab, Sebastian Seung Lab 280

Nashra Hadjerol Mala Murthy Lab, Sebastian Seung Lab 264

Zairene Lenizo Mala Murthy Lab, Sebastian Seung Lab 236

67
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Matt Collie Rachel Wilson Lab 223

Farzaan Salman Andrew Dacks Lab 219

Marion Silies Marion Silies 183

Kendrick Joules Vinson Mala Murthy Lab, Sebastian Seung Lab 175

John Anthony Ocho Mala Murthy Lab, Sebastian Seung Lab 166

Thomas Stocks Flyers (citizen scientist) 161

Kenneth J. Colodner Ken Colodner Lab 161

Gerit A. Linneweber Gerit Linneweber Lab 156

Celia David Mala Murthy Lab, Sebastian Seung Lab 151

TR77 Flyers (citizen scientist) 147

Megan Wang Mala Murthy Lab 130

Szi-chieh Yu Mala Murthy Lab, Sebastian Seung Lab 129

Lucy Houghton Sung Soo Kim Lab 123

Nils Reinhard Charlotte Helfrich-Forster Lab 123

Ben Silverman Mala Murthy Lab, Sebastian Seung Lab 121

Regine Salem Mala Murthy Lab, Sebastian Seung Lab 111

Benjamin Gorko Sung Soo Kim Lab 107

Nseraf Flyers (citizen scientist) 107

Mareike Selcho Mareike Selcho Lab 102

Haein Kim Nilay Yapici Lab 72

Minsik Yun Young-Joon Kim Lab 71

Damian Demarest Michael Pankratz Lab 70

István Taisz Greg Jefferis Lab 66

Marissa Sorek Mala Murthy Lab, Sebastian Seung Lab 62

Andrea Sandoval Kristin Scott Lab 58

Diego A. Pacheco Mala Murthy Lab 55

Kyle Patrick Willie Mala Murthy Lab, Sebastian Seung Lab 53

Zhihao Zheng Sebastian Seung Lab 51

Benjamin Bargeron Salil Bidaye Lab 47

Burak Gür Marion Silies Lab 44

Sandeep Kumar Mala Murthy Lab 40

Tansy Yang Janelia Tracers 37

Amanda González-Segarra Kristin Scott Lab 36

Gianna Vitelli Salil Bidaye Lab 29

Joanna Eckhardt Mala Murthy Lab 26

Shuo Cao David Anderson Lab 24

Haley Croke Katie von Reyn Lab 22

Nino Mancini Salil Bidaye Lab 21

Jonas Chojetzki Marion Silies 18

Gabriella R. Sterne Gabriella Sterne Lab 16

Kate Maier Salil Bidaye Lab 16

Amy R Sterling Sebastian Seung Lab 15

68
bioRxiv preprint doi: https://doi.org/10.1101/2023.06.27.546656; this version posted June 30, 2023. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
made available under aCC-BY-ND 4.0 International license.

Yuta Mabuchi Nilay Yapici Lab 12

Lucas Encarnacion-Rivera Mala Murthy Lab 10

Alexander Del Toro Kristin Scott Lab 10

Zepeng Yao Kristin Scott Lab 10

Supplementary Table 2: Number of annotations by consortium members. Only members with
≥10 annotations are shown.

69