Neural Correlates of Visuospatial Working Memory: An fMRI Study Mind & Brain Lab, Department of Psychology of the Faculty of Arts of the University of Ljubljana Nina Demšar, Aleš Oblak, Anka Slana, Grega Repovš Introduction The following research falls into the long line of inquiries into the functional localization of visuospatial working memory (VSWM). Using functional magnetic resonance imaging (fMRI), we have attempted to reaffirm some of the hypotheses, previously established when it comes to the way the prefrontal cortex (PFC) codifies different types of information about objects, when they are being manipulated by the visual working memory. Depending on the demands of the task, the PFC can encode information either in a domain-general or domain-specific way [1]. We looked into how processing of integrated stimuli differs from segregated stimuli. EEG studies have shown that the contralateral delay activation (CDA) - a wave of activation seen in such cognitive functions as attention - rises with each new stimulus held in our working memory, reaching an asymptote at 3 – 4 stimuli, the maximum number of objects we can maintain [2]. The question here is whether an integrated stimulus is perceived as a single item or its constituent features represent unique objects in and of themselves, and what are the neural correlates of this processing. Method While in the fMRI machine, the participants had to perform a VSWM task. They were presented with four stimuli. In the visual task, the stimulus was one of 25 colors. In the spatial task, the stimuli were blacked out squares. In the integration task, these two possibilities were “bound.” Following the initial training sequence, there was a pause, during which the participants had to memorize the presented stimuli. Finally, they were presented with the test stimulus. Their task was to decide whether this is one of the four stimuli presented initially or if a change occurred. There were two possible presentations of stimuli. The “2x2 load” task presented the participants with two distinct stimuli, which were both repeated once. In the “4 load” task the participants were given four unique stimuli, which were never repeated. Further, the participants had to conduct a control task, during which they had to detect the occurrence of a smaller square. The functional magnetic resonance imaging was conducted with the use of a Philips Achieva 3.0T TX machine. T1 T2 N = 34; mean age was 20.3 years, STRUCT. Results and discussion The statistical analysis of fMRI data has not yet been completed and so we are unable at this time to make any claims about the functional localization of VSWM within the PFC and the global connectivity involved in processing of integrated stimuli. However, we have made some surprising findings regarding the behavioral data. Earlier iterations of a similar research design have shown that the participants tend to perform better on the visual task (graph 2). This; however, is not consistent with our findings. In fact, the participants performed better in the spatial task, while the integration task showed the worst results by far (graph 1). TASK RS. 28 of them were female 2 + 2 load 4 load K 3 2 1 T In the original research, the presentation of stimuli was slightly different. There was no grid. Instead, participants were presented with a blank space, where stimuli appeared at seemingly random locations. What we can assume from this discrepancy is that the participants perform significantly better once they have a visual frame of reference into which they can place the presented stimuli. So small differences in the design of the task can translate into large differences in behavioral data. Integrated stimuli are not codified as single unique object as some have suggested. The poor performance in the integrated task shows that the disparate features are perceived as unique objects, rather than constituent components of a larger stimulus. What this means for the hierarchical dynamic model hypothesis proposed by Sala and Courtney [1], remains to be ascertained by way of fMRI data. C P I Graph 1 Simultaneous Sequential K 3 2 1 T C P I Graph 2 References: [1] J. B. Sala and S. M. Courtney, “Binding of what and where during working memory maintenance,” Cortex, vol. 43, no. 1, pp. 5-21, Jan 2007. [2] E. K. Vogel and M. G. Machizawa, “Neural activity predicts individual differences in visual working memory capacity,” Nature, vol. 428, no. 6984, pp. 748-51, Apr 2004.