CN113197548B - Intracranial implantable flexible multi-modal physiological and biochemical information monitoring equipment - Google Patents

Abstract

The invention provides intracranial implantation type flexible multi-mode physiological and biochemical information monitoring equipment which comprises a multi-mode sensor, a flexible flat cable and a signal processing and wireless transmission circuit, wherein the multi-mode sensor comprises a pressure sensor, a shearing force sensor, a temperature sensor, an oxygen partial pressure sensor, a potassium-sodium ion sensor and a flexible electrode; the multi-mode sensor and the signal processing and wireless transmission circuit are connected through the flexible flat cable. A flexible substrate and a multi-mode sensor are prepared based on a flexible material similar to a brain elasticity model, a sensor capable of measuring intracranial pressure, intracranial oxygen partial pressure, intracranial temperature, intracranial nerve electric signals, intracranial sodium ion concentration and intracranial potassium ion concentration in an integrated mode is prepared, and using stability and accuracy information of the sensor are detected in real time by preparing a shear force sensor and an electrode impedance sensor.

Description

Intracranial implantable flexible multi-modal physiological and biochemical information monitoring equipment

technical field

The invention relates to the technical field of physiological signal monitoring, in particular to an intracranial implantable flexible multi-mode physiological and biochemical information monitoring device.

Background technique

Traumatic brain injury is a global health problem with a high mortality rate. In China, about 600,000 people suffer from traumatic brain injury every year, of which about 100,000 people die, causing direct and indirect economic losses of more than 10 billion yuan. Preliminary statistics from the China Craniocerebral Trauma Database show that among more than 13,000 inpatients with acute craniocerebral trauma in 47 hospitals in China, the fatality rate of severe craniocerebral trauma is >20%, and the severe disability rate is >50%. Foreign data show that for patients with severe craniocerebral injury with Glasgow Coma Scale (Glasgow Coma Scale, GCS) ≤ 8 points, the mortality rate is as high as 35% to 45%.

In the prior art, there have been relevant studies on single or dual-mode brain physiological index monitoring technology, and the following introductions are made:

(1) Intracranial pressure: According to whether the sensor is implanted into the skull in the existing products, the intracranial pressure monitoring method can be divided into two types: non-implantable monitoring method and implantable monitoring method. The non-implantable intracranial pressure monitoring method does not need to implant sensors in the brain, can avoid the trauma caused by implanted monitoring to patients, and has no risk of infection during the monitoring process, and has the advantages of simple operation and low cost. Its principles include transcranial Doppler method, flash visual evoked potential method and infrared spectroscopy. The above non-implantable monitoring methods all convert other physical quantities into pressure values through indirect means, and there are many interference factors in the monitoring process, which makes the measurement error larger. The implantable intracranial pressure monitoring method refers to implanting a pressure sensor into the skull to directly measure the intracranial pressure value. The method is to insert one end of the pressure measuring catheter into the ventricle through brain surgery, and connect the other end to the hydraulic sensor interface to measure the intracranial pressure. The advantage of this method is that the detection accuracy is high, which can meet the requirements of clinical intracranial pressure monitoring indicators, and is regarded as the "gold standard" for intracranial pressure monitoring. However, this method and other existing implanted intracranial pressure monitoring devices are all wired, which has many inconveniences in the continuous monitoring process, is not suitable for long-term monitoring, and increases the chance of wound infection. There are foreign literature reports that the probability of infection is even as high as 11%.

(2) Intracranial EEG signal: The measurement of intracranial EEG signal can be divided into the measurement of the electrical signal in the cortical area and the measurement of the electrical signal inside the cortex. Cortical EEG is to place cortical electrodes on the intracranial cerebral cortex by surgery, and the electrode wires pass through the skull and scalp and are connected to an external EEG. By analyzing the power spectral density, beta wave (15-25Hz), alpha wave (8-13Hz), delta-theta wave (2-8Hz) of the cortical signal, the non-convulsive seizures caused by secondary brain injury can be analyzed. Detect the initial moment or status epilepticus, and give timely feedback and treatment. In 2016, De Marchies et al. adopted the flexible multi-channel implantable cortical brain electrode product of Ad-Tech Medical Company of the United States. The deep cerebral cortex implantable electrode uses silicon-based or stainless steel neural probes, which are implanted into the cerebral cortex to measure brain nerve signals. By analyzing field potentials and action potentials, it is possible to obtain the degree of brain damage to the motor or sensory cortex. .

(3) Intracranial temperature: Under different physiological and pathological conditions, the temperature of the brain will not only fluctuate, but also affect the changes of multiple physiological metabolisms of cortical cells. Researchers have tried to use infrared imaging to display the temperature of the cortex, but due to its lack of accuracy and the inability to monitor in real time, it has only stayed in a few research areas. At present, implantable temperature sensor probes are generally used to measure cortical temperature by implanting them into the cerebral cortex. The intracranial temperature sensor is usually integrated with the intracranial pressure detection or intracranial oxygen partial pressure detection probe, but the wound is large, the measurement time is short, and it is In a craniotomy, where the skull and scalp are not covered, the values may vary considerably from the physiological state.

(4) Intracranial cerebral oxygen partial pressure: In the treatment of brain injury, the method of evaluating the oxygenation status of the whole body is not reliable, so the application of the method of detecting brain tissue oxygenation is very important. Currently more mature methods include whole brain oxygen measurement and local oxygen measurement. The whole-brain oxygen measurement method uses a probe to measure the jugular vein. This method is more convenient, but it cannot accurately obtain the distribution of oxygenation in the brain tissue, and cannot obtain accurate positioning of the damaged brain tissue. Local oxygen measurement methods include near-infrared spectrometer non-invasive technology and tissue probe implantation method. The former obtains the partial pressure of oxygen by comparing the light intensity information entering the brain and returning, such as the near-infrared non-invasive brain tissue oxygen monitor from CAS Medical in the United States. However, because the light penetration path has greater uncertainty than the finger oxygenation measurement method, such as brain swelling after traumatic brain injury will change the optical path distance, thereby reducing the measurement accuracy and reliability. At present, the more reliable method is to use tissue probes to measure the partial pressure of oxygen in local tissues. The monitoring equipment companies for partial pressure of oxygen and temperature of the brain mainly include Licox implantable products from Integra in Germany and Neurovent-PTO implantable products provided by Raumedic in the United States. product.

(5) Intracranial electrolyte concentration

About 62% of the patients died of various complications after traumatic brain trauma, and among them, the endocrine dysfunction after traumatic brain trauma led to water and electrolyte disorders, and then the internal environment balance was destroyed, causing pathological stress reaction, leading to water and electrolyte disorders or damage or failure of multiple organ functions. Water and electrolyte disorders in patients after traumatic brain trauma have long plagued the majority of medical workers, including hyponatremia, hypernatremia, and hypokalemia. Therefore, it is very important to simultaneously detect the component concentrations of potassium and sodium ions in brain interstitial fluid. At present, there is no implantable electrolyte detection method for the brain.

To sum up, the current monitoring of intracranial physiological indicators mostly uses implantable sensor probes, which bring about problems such as large wound area, susceptibility to infection, and inconvenient wire connection, which significantly limit its clinical application. At present, there is no highly integrated instrument and method that can monitor multi-modal intracranial cerebral cortex signals in real time and synchronously at home and abroad. Therefore, studying the principle of multi-modal signal sensing and exploring accurate and reliable wireless miniature intracranial cerebral cortex signal monitoring equipment will help reveal the mechanism of secondary brain injury, analyze the cause of secondary brain injury, predict and treat secondary brain injury. Sub-brain injury is of great significance.

Contents of the invention

In view of this, the main purpose of the present invention is to provide an intracranial implantable flexible multi-modal physiological and biochemical information monitoring device, in order to partially solve at least one of the above technical problems.

In order to achieve the above purpose, as an aspect of the present invention, a flexible multimodal physiological and biochemical information monitoring device for intracranial implantation is provided, including multimodal sensors, flexible cables, and signal processing and wireless transmission circuits. The multimodal sensor includes a pressure sensor, a shear force sensor, a temperature sensor, an oxygen partial pressure sensor, a potassium and sodium ion sensor and a flexible electrode; the multimodal sensor and the signal processing and wireless transmission circuit pass through the flexible cable connect.

Wherein, the pressure sensor is used to detect the pressure value in the brain, and the pressure sensor adopts a resistive pressure sensor or a capacitive pressure sensor, and adopts a sandwich structure of upper and lower electrode plates and interlayer materials for pressure sensing.

Wherein, the shear force sensor is used to monitor the shear force on the flexible substrate, and uses the left and right parallel electrode plates and the structure of the middle sensitive material to sense the shear force.

Wherein, the temperature sensor is used for real-time monitoring of intracranial temperature, and metal Pt or Au is used as the temperature sensor.

Wherein, if the multimodal sensor is placed above the cerebral cortex, the oxygen partial pressure sensor measures the oxygen concentration of the cerebrospinal fluid; if the multimodal sensor is inserted into the cerebral cortex, the oxygen partial pressure sensor Oxygen concentration measurement in interstitial fluid inside the cortex.

Wherein, if the multimodal sensor is placed above the cerebral cortex, the potassium and sodium ion sensor measures the concentration of potassium and sodium ions in the cerebrospinal fluid; if the multimodal sensor is inserted into the cerebral cortex, the potassium and sodium ion The sensor measures the concentration of potassium and sodium ions in the interstitial fluid inside the cerebral cortex.

Wherein, the multimodal sensor integrates no less than 8 flexible electrodes, which are evenly distributed on the substrate, and the flexible electrode has the functions of electrical signal measurement and electric stimulator at the same time. Above the cerebral cortex, the flexible electrodes measure the neural signals of the cerebral cortex or electrically stimulate the cortical nerves. If the multimodal sensor is inserted into the cerebral cortex, the flexible electrodes measure the neural signals or stimulate the Internal nerves are electrically stimulated to complete neural function regulation.

Wherein, the multimodal sensor uses metal Pt or Ir as the flexible electrode material.

Wherein, the multi-modal sensor further includes a flexible base, and the elastic modulus of the flexible base material matches the brain tissue, so that the flexible base material has a commonality with the cerebral cortex.

Wherein, the multimodal sensor also includes a lead wire, which is used to connect the sensor and the connector at the end of the substrate; the material of the lead wire includes a solid metal material, a degradable metal material, a nanocomposite Conductive materials and liquid conductive metals.

Based on the above technical solutions, it can be seen that the intracranial implantable flexible multi-modal physiological and biochemical information monitoring device of the present invention has at least a part of the following beneficial effects compared with the prior art:

1. Prepare flexible substrates and multi-modal sensors based on flexible materials similar to the brain elastic model, and produce sensors that can measure intracranial pressure, intracranial oxygen partial pressure, intracranial temperature, intracranial nerve electrical signals, and intracranial sodium ions Concentration and intracranial potassium ion concentration integrated sensors, and real-time detection of sensor stability and accuracy information by making shear force sensors and electrode impedance sensors.

2. Flexible sensors including substrates, lead wires, multi-modal sensors and covering layers can be prepared in a variety of ways, flexible sensors can be bent into rolls, and the flexible sensor can be implanted with the intracranial cerebral cortex through the use of clamps above, or within the intracranial cortex. When placed above the cortex, the sensor mainly detects the relevant physiological and biochemical information of the intracranial cerebrospinal fluid, while for the sensor implanted inside the cerebral cortex, it mainly detects the relevant physiological and biochemical information in the cortical tissue and interstitial fluid.

3. In view of the characteristics of each block of the flexible sensor, within a certain deformation range, the functional use of the sensor will not be affected.

4. The present invention proposes an integrated method for multimodal sensing information acquisition, which is mainly divided into power management, multimodal sensing front end, analog-to-digital conversion, micro-processing control unit, digital processing unit and digital signal output unit. And because it has the characteristics of miniaturization and high integration, it is fixed above the skull, and the flexible wire is passed through the skull to connect with the intracranial sensor, and the sensing signal can be transmitted to the terminal equipment outside the hospital in a wireless manner.

5. The present invention proposes a brain disease judgment method based on neural network algorithm and random forest machine learning algorithm, through 6 groups of intracranial information (intracranial pressure, intracranial partial pressure of oxygen, intracranial temperature, intracranial nerve electrical signal, Intracranial sodium ion concentration and intracranial potassium ion concentration) input, multi-level data processing, automatic judgment output of various intracranial injury types, and designate intervention treatment combinations according to the combination of output layer types.

6. For the intracranial electrical signals collected by 8-channel electrodes, the present invention adopts the epilepsy detection algorithm of multi-channel mixed feature matrix fusion and the multi-channel functional electrical stimulation epilepsy treatment algorithm to automatically determine the type and orientation of epilepsy, and release it by controlling the microprocessor. Different types of electrical stimulation protocols.

Description of drawings

Fig. 1 is the overall schematic diagram provided by the embodiment of the present invention;

Fig. 2 is a mode in which the system provided by the embodiment of the present invention is placed inside the brain (outside the flexible sensor cortex);

Fig. 3 is another way in which the system provided by the embodiment of the present invention is placed inside the brain (inside the flexible sensor cerebral cortex);

Fig. 4 is a structural diagram of the flexible multimodal sensor provided by the embodiment of the present invention;

Fig. 5 is a cross-section of a flexible multimodal sensor provided by an embodiment of the present invention;

Fig. 6 is a cross-sectional structure diagram of a flexible pressure sensor provided by an embodiment of the present invention;

Fig. 7 is a cross-sectional structure diagram of a flexible shear force sensor provided by an embodiment of the present invention;

Fig. 8 is a structural diagram of a flexible temperature sensor provided by an embodiment of the present invention;

Fig. 9 is a cross-sectional view of an oxygen partial pressure sensor based on an electrochemical method provided by an embodiment of the present invention;

10 is a cross-sectional view of an ion sensor based on an electrochemical method provided by an embodiment of the present invention;

Figure 11 is a cross-sectional view of a flexible electrode provided by an embodiment of the present invention;

FIG. 12 is a structural block diagram of a circuit system provided by an embodiment of the present invention;

Fig. 13 is the disease prediction model based on the artificial neural network provided by the embodiment of the present invention;

Fig. 14 is the disease prediction model based on random forest provided by the embodiment of the present invention;

Fig. 15 is a flow chart of intelligent identification and electrical feedback treatment of epilepsy under brain injury provided by the embodiment of the present invention;

Fig. 16 is a block diagram of an algorithm for intelligent identification and electrical feedback treatment of epilepsy under brain injury provided by an embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

In view of the current lack of integrated minimally invasive subtraumatic brain multiple physiological parameters detection methods and equipment, as well as the unquantifiable changes in intracranial pressure, brain electrical activity, intracranial temperature, intracranial oxygen partial pressure, and electrolyte concentration indicators after secondary traumatic trauma The present invention proposes a multi-sensing fusion detection technology based on flexible sensors integrating intracranial pressure, intracranial electrical signals, intracranial temperature, intracranial oxygen partial pressure, and intracranial electrolyte concentration flexible sensors, and cooperates with related detection of small It realizes synchronous, real-time, and wireless implantable intracranial physiological information monitoring, and realizes automatic diagnosis of brain disease types through machine learning algorithms. At the same time, for epilepsy and other electrical signal disorder diseases caused by brain injury, the flexible electrode integrated in the system has the function of electrical stimulation. After the system predicts the disease, it can electrically stimulate the local cortex, and then relieve or treat the electrical signal disorder through electrical intervention.

The overall schematic diagram of the invention is shown in Fig. 1 .

The system includes a flexible sensor and a small signal processing and wireless transmission circuit, and the two parts are connected through a flexible cable. The flexible sensor integrates a flexible substrate and multiple flexible sensing units, which can be placed in the cerebrospinal fluid above the cortex (Fig. 2) or implanted longitudinally in the cortex (Fig. 3); small circuits are usually fixed on the outside of the skull and passed through Flexible FPC cables, nylon filaments or other biocompatible flexible wires are used as an insulating layer to electrically connect with flexible sensors through the skull cavity or turning hole at the brain injury to monitor multiple physiological parameters. The flexible data cable can be fixed on the sensor substrate by pasting at the end of the flexible integrated sensor.

The implantation method in the body of the proposed system of the present invention is described as follows:

method one:

(1) The flexible sensor is bent into a roll, clamped on the front end of a related fixture such as tweezers, and keeps the sensor in a roll;

(2) Find the skull injury on the skull, and the opening diameter is less than 5mm; or use the skull electroporation to create an aperture of similar size;

(3) The clamp passes through the opening of the skull, goes deep into the skull, and releases the roll-shaped sensor;

(4) The flexible sensor is laid directly above the cerebral cortex to detect intracranial pressure, oxygen partial pressure, temperature, potassium and sodium ions and electrical signals, and the electrodes can be used for electrical stimulation of the cortex;

(5) Use screws or glue to fix the small circuit board under the scalp on the outside of the skull;

(6) The flexible wiring of the circuit board and the sensor passes through the opening of the skull and is fixed.

Method 2:

(1) Bend the flexible sensor along the long side, wrap and paste it on the outside of the fixture with a biosoluble glue (such as polysaccharide), and keep the sensor in a roll shape;

(2) Find the skull injury on the skull, and the opening diameter is less than 5mm; or use the skull electroporation to create an aperture of similar size;

(3) The jig passes through the opening of the skull, goes deep into the skull, and inserts it into the cerebral cortex longitudinally;

(4) After a certain period of time, the biosoluble glue decomposes, and the sensor is separated from the jig. At this time, the jig is removed, so that the sensor remains in a roll shape inside the cortex, and the sensor units are distributed along the outside of the roll. At this time, intracranial pressure, oxygen partial pressure, temperature, potassium and sodium ions and electrical signals can be detected, and the electrodes can be used for electrical stimulation of the cortex;

(5) Screws or glue are often used to fix small circuit boards under the scalp on the outside of the skull;

(6) The flexible wiring of the circuit board and the sensor passes through the opening of the skull and is fixed.

Below, the present invention will be described in detail in three aspects: sensor system composition, small circuit system, disease type analysis and electric stimulation curative effect.

1. Flexible multimodal sensor

The composition diagram of the flexible multimodal sensor is shown in Fig. 4.

The flexible multimodal sensor proposed by the present invention is composed of flexible base materials and various types of flexible sensors. Manufacture no less than 2 flexible pressure sensors and 2 flexible shear force sensors, no less than 2 flexible temperature sensors, no less than 2 flexible oxygen partial pressure sensors, and no less than 2 flexible oxygen partial pressure sensors through different flexible sensor manufacturing processes. There are no less than 2 flexible potassium and sodium ion sensors, and no less than 9 electrodes (the electrodes can be used as nerve signal acquisition or nerve electrical stimulation, or the same electrode can be used for both electrical signal acquisition and nerve electrical stimulation through time division multiplexing). electrical stimulation function). A cross-section of the flexible multimodal sensor shows the structure of its longitudinal layers, as shown in Fig. 5.

Among them, the lowermost layer is a flexible substrate, and the upper part is a metal or polymer lead wire. A flexible sensor is made or placed at the node of the lead wire, and finally a flexible polymer layer is covered on the top for insulation, waterproof and protection. etc., the covering layer can choose to open or not open above the sensor according to the sensing type, for example, for the electrodes, the opening of the flexible covering layer must be selected. The present invention is specifically stated below to different structural layers:

1. Flexible substrate

Flexible substrate materials choose organic materials with high biocompatibility, such as PMMA, polyimide, PDMS, parylene, SEBS, Ecoflex, SU-8, PET, PEE, PVC, Cumene-PSMA, PSE, PVP and other rubber or resin materials, etc. Degradable materials with high biocompatibility, such as PLGA, PVA, PGS, silk protein, Zyvox, collagen, chitosan, POMaC, PLLA, PCL, etc., by choosing the ratio of different components, the elastic modulus of the material can be achieved. The amount is matched with the brain tissue, so that it has a commonality with the cerebral cortex, and the damage to the brain tissue of the implanted device is minimized. Among them, the outer dimensions of the flexible substrate and the multimodal sensor are less than 6cm 6cm, and the thickness is usually less than 20μm. The specific size can be selected according to the actual situation of the patient's wound and disease type. Flexible multimodal sensors of different sizes and shapes.

There are two options for the structure of the substrate: one is composed of a non-patterned homogeneous substrate material, and all parts of the substrate have the same material density; the second is made of a hollow patterned substrate material. The base material under the wire is made into the same or similar pattern as the lead wire. For example, the base under the lead wire is made into a serpentine, arc or arbitrary curved shape, which is more conducive to the ductility of the base and can be tightly attached. Hold the cortex. The substrate material beneath the sensor unit remains solid and unpatterned.

For non-patterned homogeneous substrate materials, semiconductor processes, such as PECVD, evaporation, sputtering, ALD and other thin film deposition processes, or liquid synthesis and solidification can be used for production. For patterned substrate materials, laser cutting, shadow mask plasma etching, mask chemical wet etching, or mold casting can be used.

2. Lead wire

After the flexible substrate is manufactured, a curved lead wire is made on the substrate to connect the sensor and the connector at the end of the substrate. To maintain the electrical stability of the lead wire in the stretched or compressed state, the lead wire material is usually made of materials with good flexibility and high biocompatibility, such as Ag, Au, Mg, Mo, carbon nanotube, polyaniline, PEDOT:PSS , Carbon black and other solid metal materials, degradable metal materials, nanocomposite conductive materials, liquid conductive metals. The conductive liquid material can be mixed with polymers in a liquid state such as PDMS, PMMA, SEBS, etc., to form lead wires after curing. The lead wires can be made in the following ways:

(1) Manufacture 1 by semiconductor technology: This method can produce patterned solid metal wires, such as Ag, Au, Mg, and Mo. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. Firstly, a thin solid metal film is deposited by sputtering or evaporating equipment, the thickness of which is less than 3 μm, and then the pattern of the metal line is patterned on the photoresist through the steps of spinning glue, exposure, development, post-baking, etc., and then wet etching solution is used The metal is etched by etching or dry gas etching, and finally the residual photoresist is removed to complete the preparation of the metal lead wire.

(2) Manufacture 2 by semiconductor technology: This method can produce patterned solid metal wires, such as Ag, Au, Mg, and Mo. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. First, on the flexible substrate, pattern the pattern of the metal line on the photoresist through steps such as glue rejection, exposure, development, and post-baking, and then use a sputtering or evaporation device to deposit a thin layer of solid metal film with a thickness of less than 3 μm, and then use The metal film lift-off process removes the photoresist and the metal above the photolithography to complete the preparation of the metal lead wire.

(3) Manufacture by metal deposition process under shadow mask: This method can produce patterned solid metal wires, such as Ag, Au, Mg, Mo. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. First, use mechanical processing (such as CNC, laser cutting, etc.), 3D printing, or semiconductor substrate material etching to make a hollowed-out lead line pattern on the substrate material, and then place the shadow mask material on the flexible sensor substrate. contact, and then use a sputtering or evaporation device to deposit a thin layer of solid metal film, the thickness of which is less than 3 μm, and finally remove the shadow mask material to complete the preparation of the metal lead wire.

(4) Manufactured by the process of deposition of conductive liquid under shadow mask: This method can produce patterned conductive liquid, such as liquid Au particles, liquid Ag particles, carbon nanotube, polyaniline, PEDOT:PSS, carbon black, etc. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. First, use mechanical processing (such as CNC, laser cutting, etc.), 3D printing, or semiconductor substrate material etching to make a hollowed-out lead line pattern on the substrate material, and then place the shadow mask material on the flexible sensor substrate. Contact, and then cover the conductive liquid with a spin coater or glue sprayer or dispenser, the thickness of which is usually not less than 10 μm, and then heat and cure the conductive liquid to complete the preparation of the lead wire.

(5) Process production by screen printing: This method can produce patterned conductive liquids, such as liquid Au particles, liquid Ag particles, carbon nanotube, polyaniline, PEDOT:PSS, carbon black, etc. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. First make a screen printing mask according to the pattern of the metal lead wire, then place the flexible sensor base on the screen printing machine, pour conductive liquid into the liquid storage tank, and carry out the printing pattern operation, and then heat and solidify the conductive liquid to complete the conduction In-line preparation.

(6) Manufacture by pipetting: This method can produce patterned conductive liquids, such as liquid Au particles, liquid Ag particles, carbon nanotube, polyaniline, PEDOT:PSS, carbon black, etc. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. First program the electronic pipette, import the graphic file of the lead wire, then pour the conductive liquid into the liquid reservoir of the electronic pipette, then perform the printing operation, eject the liquid material, print the lead wire pattern, and finally The conductive liquid is heated and solidified to complete the lead wire preparation.

(7) Fabrication by means of pattern transfer: This method can produce all types of liquids suitable for the lead wires of the present invention. The manufacturing process can generally be summarized as follows, and the process details can be adjusted according to the process conditions. First deposit a thin film on a hard substrate such as silicon or glass, such as silicon oxide, silicon nitride, or polymer, and then pattern the lead wires on top, and then use the pattern transfer process to place a high-viscosity film. The flexible sensor substrate is placed above the lead wire, and a certain pressure is applied, and then slowly peeled off from the edge, finally completing the transfer of the lead wire pattern.

3. Sensor fabrication

(1) Flexible pressure sensor

The overall elastic modulus of the pressure sensor should be close to that of the flexible substrate to ensure the flexible characteristics of the multimodal sensor. Therefore, flexible materials are used to make pressure sensors. A piezoresistive or capacitive type can be used to make a flexible pressure sensor. According to the requirements of the present invention, the flexible sensor is customized for design and processing, including pressure sensitivity range (5-50mmHg), resolution (better than 1mmHg), high biocompatibility of materials, and water resistance.

The pressure sensor of the present invention is used to detect the pressure value in the brain, so it is designed to detect the pressure value perpendicular to the flexible substrate. Therefore, for pressure sensors with the three principles of piezoresistive and capacitive, it is necessary to use a sandwich structure of upper and lower electrode plates with interlayer materials for pressure sensing. The cross-sectional structural diagram of the pressure sensor is shown in Figure 6.

The process of making the pressure sensor is: firstly, the lower plate is made on the flexible substrate. The electrodes of the plate are made of metal materials with better biocompatibility, such as Au, Ag, Mg, Mo, etc., which can be sputtered and evaporated by semiconductor equipment. , photolithography, etching, liquid metal solution printing, screen printing and other processes to prepare the patterned lower plate metal, the lower plate can be optionally covered with a thin insulating layer, which is also a part of the lower plate Then make the pressure sensor on the bottom plate, also can adopt the mode of pressure sensing sensitive film transfer, the sensitive layer that will make is transferred to the top of the bottom plate, or adopt to directly generate or deposit the pressure sensitive film on the top of the bottom plate, and The pressure-sensitive film is fabricated by patterning the pressure-sensitive film by etching or corrosion technology; then, the upper electrode plate (usually including electrodes and insulating covering layers) is pasted on the pressure-sensitive film, and the size of the upper plate is longer than that of the pressure-sensitive film. film, so that the metal of the upper plate is connected to the lead wire on the flexible substrate; finally, a thin layer of flexible insulating layer is covered to complete the production of the flexible pressure sensor. In addition, the pressure sensor made in the present invention can be transferred on other substrates, such as silicon, silicide, glass and other substrate materials, and then transferred, and the two electrodes and the sensitive film are transferred and fixed on the flexible substrate as a whole. .

For the piezoresistive pressure sensor, the resistance value of the sensitive film is changed by the change of pressure. Nano-silver and gold particles, CNT, Polyaniline, PEDOT:PSS and other materials can be selectively mixed into PDMS rubber materials in proportion to synthesize Pressure sensitive film. At this time, the metal layers of the upper and lower electrode plates should be attached to the pressure sensitive film to achieve the effect of conduction. In addition, the pressure-sensitive film can also be made into different surface forms and shapes, such as rhombus, inverted pyramid, high-roughness surface, etc., to adjust the pressure-sensitive range and resolution required by the present invention.

For capacitive sensors, the capacitance value of the sensitive film is changed by the change of pressure, and insulating high-elastic dielectric materials, such as PDMS, SEBS, etc., can be used as pressure-sensitive materials. At this time, the metal of the upper and lower electrode plates does not need to be attached to the pressure-sensitive film, and the change of the capacitance value with respect to the pressure can be measured. In addition, the pressure-sensitive film can also be made into different surface forms and shapes, such as rhombus, inverted pyramid, high-roughness surface, etc., to adjust the pressure-sensitive range and resolution required by the present invention.

(2) Flexible shear sensor

In the present invention, the flexible multi-modal sensor integrates a shear force sensor, which can effectively monitor the shear force on the flexible substrate, and then analyze the working conditions of the sensor in the implant according to the multi-modal sensing data, and analyze multiple Stability and reliability of modal data, etc.

The shearing force of the present invention measures the magnitude of the shearing force of the flexible substrate, so the left and right parallel electrode plates and the structure of the middle sensitive material are used to sense the shearing force. The cross-sectional structure diagram of the shear force sensor is shown in Figure 7.

The process of making the shear force sensor is: firstly, the left and right plates are made on the flexible substrate. The electrodes of the plates are made of metal materials with good biocompatibility, such as Au, Ag, Mg, Mo, etc., which can be sputtered by semiconductor equipment. , evaporation, photolithography, etching, liquid metal solution printing, screen printing and other processes to prepare patterned left and right plate metals, and then make a shear force sensing sensitive film between the two plates above the flexible substrate, or use shearing In the method of transferring the shear sensor sensitive film, the prepared sensitive layer is transferred to the middle of the two polar plates above the flexible substrate and fixed; finally, a thin flexible insulating layer is covered to complete the production of the flexible shear force sensor. In addition, the shear sensor made in the present invention can be transferred on other substrates, such as silicon, silicide, glass and other substrate materials, and then transferred, and the two electrodes and the sensitive film are integrally transferred and fixed on the flexible substrate. . Sensitive films can be made of nanoparticles such as silver and gold, CNT, Polyaniline, PEDOT:PSS and other materials that are selectively mixed into PDMS rubber materials in proportion to synthesize pressure-sensitive films. At this time, the metal of the left and right electrode plates should be attached to the shear force sensing sensitive film to achieve the effect of conduction. In addition, the pressure-sensitive film can also be made into different surface forms and shapes, such as rhombus, inverted pyramid, high roughness surface, etc. to adjust the shear force sensitivity range and resolution required by the present invention.

(3) Flexible temperature sensor

In the present invention, the flexible multi-modal sensor integrates a temperature sensor for real-time monitoring of intracranial temperature. The invention utilizes the resistance-temperature relationship of metals as a sensing principle to detect intracranial temperature. Metal Pt or Au is commonly used as a temperature sensor because of its good linearity and high temperature coefficient of resistance. The initial resistance value of the temperature sensor should be larger, so that a larger resistance value change can be obtained under a unit temperature change. Since the lead wire is usually made of metal, it also has a resistance temperature effect. In order to counteract the resistance temperature effect caused by the lead wire, the width of the lead wire is at least 5 times larger than the metal width of the temperature sensor, and the temperature sensor has a longer size to increase the initial resistance.

The present invention selects metal Pt or Au as the temperature sensor, and provides design drawings of the two temperature sensors, as shown in FIG. 8 .

The temperature sensor in Figure 8(a) is wound around a square, and the horizontal and vertical wiring have the same spacing, so that while increasing the resistance length, most of the resistance caused by the tangential stress of the sensor substrate can be offset. value changes. The temperature sensor graphic scheme in Figure 8(b) can partially offset the resistance change caused by the tangential stress of the temperature sensor substrate, and at the same time increase the sensing area of the sensor temperature, and play the role of averaging the temperature values of multiple points in space. Avoid sudden changes in temperature sensor output caused by local temperature changes.

If Au is used as the material of the temperature sensor, its fabrication can be completed simultaneously with the fabrication of the lead wire. If Pt is used as the temperature sensor material, patterned Pt can be prepared by sputtering, evaporation, photolithography, etching and other processes of semiconductor equipment; or after the preparation on other substrates, the Pt temperature sensor can be completed by pattern transfer. preparation. For specific steps, please refer to the lead wire preparation process. Finally, a thin layer of flexible covering layer is covered to complete the preparation of the temperature sensor.

(4) Flexible oxygen partial pressure sensor

In the present invention, the flexible multimodal sensor integrates the oxygen partial pressure sensor. If the multimodal sensor is placed above the cerebral cortex (Figure 2), the oxygen partial pressure sensor can measure the oxygen concentration of the cerebrospinal fluid. If the multimodal sensor is inserted into the cerebral cortex (FIG. 3), the oxygen partial pressure sensor measures the oxygen concentration in the interstitial fluid inside the cerebral cortex. It can be seen that the implantation methods of the two sensors can directly reflect the concentration of oxygen components in the brain. The present invention proposes a method for measuring oxygen partial pressure based on an electrochemical method, and FIG. 9 schematically shows a cross-sectional view of an oxygen partial pressure sensor.

The manufacturing method is that above the flexible substrate of the multi-modal sensor, metal Pt three electrodes are fabricated by using the metal film lift off process, which are the working electrode (WE), the counter electrode (CE), and the reference electrode (RE); Among them, Ag and AgCl can be selectively deposited on the two electrodes of CE and RE to achieve the purpose of standard measurement potential. The area above the electrode is covered with a thin layer of proton exchange membrane (nafion) and cured by spray printing or dispensing, and then the liquid gas filter membrane (such as PDMS, PTFE) is sprayed and printed above the nafion and cured, and finally the The flexible insulating layer covers the oxygen partial pressure sensor. At this time, the region of the flexible insulating layer directly above the oxygen partial pressure sensor needs to be etched by photolithography (or shadow mask) to ensure that oxygen molecules can enter. The oxygen partial pressure sensing unit can also be prepared on hard materials such as silicon, silicide, or glass, and transferred to the sensor and the flexible substrate by pattern transfer, and cover the uppermost flexible layer. The area of the counter electrode is larger than the working electrode and the reference electrode, so the working electrode and the reference electrode can be placed in the middle area in the design, and the counter electrode surrounds the other two electrodes to form an efficient current path.

The role of the gas filter membrane is to only allow oxygen molecules to penetrate and reach the proton exchange membrane. The purpose of the proton exchange membrane is to provide flowable protons (H+) for the redox reaction generated by the three electrodes to form a continuous redox reaction. Under the flow of protons, oxygen gets hydrogen ions on the working electrode to generate water and electrons, and the water and electrons are electrolyzed into oxygen and hydrogen ions on the counter electrode. The support of proton circulation is required in the whole redox reaction process, but oxygen or protons are not consumed in the whole process. If the oxygen concentration increases, a greater current will be generated at the working electrode. In this way, the relationship between oxygen partial pressure and current intensity is established.

(5) Flexible potassium and sodium ion sensor

In the present invention, the flexible multimodal sensor integrates potassium and sodium ion sensors. If the multimodal sensor is placed above the cerebral cortex (Fig. 2), the potassium and sodium ion sensor will measure the concentration of potassium and sodium ions in the cerebrospinal fluid. If the multimodal sensor is inserted Inside the cerebral cortex (FIG. 3), the potassium and sodium ion sensors measure the concentration of potassium and sodium ions in the interstitial fluid inside the cerebral cortex. The present invention proposes a method for measuring potassium and sodium ions based on an electrochemical method, and FIG. 10 schematically shows a cross-sectional view of a potassium and sodium ion sensor.

The fabrication method is as follows: above the flexible substrate of the multimodal sensor, metal Pt three electrodes are fabricated by using the metal film lift off process, which are the potassium ion working electrode (WE1), the reference electrode (RE) and the sodium ion working electrode. (WE2); where Ag and AgCl can be selectively deposited on the two electrodes of RE to achieve the purpose of standard measurement potential, and a layer of PEDOT:PSS can be selectively deposited on WE1 and WE2 to reduce the potential drift. The area above the CE1 and CE2 electrodes is covered with a potassium ion selective film and a sodium ion selective film by liquid jet printing or dispensing, respectively. The main chemicals are K ionophore X, K-TFPB and Na ionophore X, Na-TFPB, and Mixed binders such as PVC and DOS are synthesized and cured to form solid potassium and sodium ion selective films. Finally, the flexible covering layer is covered on the potassium and sodium ion sensor. At this time, the area directly above the potassium and sodium ion sensor needs to be etched by photolithography (or shadow mask) to ensure that ions can enter the sensor. The potassium and sodium ion sensing unit can also be prepared on hard materials such as silicon, silicide, or glass, and transferred to the sensor and the flexible substrate by pattern transfer, and cover the uppermost flexible layer. Due to the dual-electrode mode adopted, the reference electrode (RE) is also the counter electrode, and its area is larger than that of the working electrode. Therefore, the working electrodes (WE1 and WE2) are placed in the middle area in the design, and the reference electrode (RE) surrounds the other electrodes. Two electrodes to form an efficient current path.

The function of the potassium and sodium ion selective membrane is to selectively allow potassium ions and sodium ions to pass through the selective membrane. After the ions reach the working electrode, the potential of the working electrode increases relative to the reference electrode RE. The higher the ion concentration, the higher the WE and RE. The potential difference is also higher, establishing a relationship between ion concentration and voltage.

(6) Flexible electrodes and electrical stimulators

In the present invention, the flexible multi-modal sensor integrates no less than 8 flexible electrodes, which are evenly distributed above the substrate (Figure 1). The 8 flexible electrodes can simultaneously have the functions of electrical signal measurement and electrical stimulator. If the multimodal sensor is placed above the cortex (Fig. 2), the electrodes can measure the cortical nerve signal (Ecog) or perform electrical stimulation on the cortical nerves. If the multimodal sensor is inserted inside the cortex (Fig. 3), The electrodes can measure nerve signals or electrically stimulate the inner nerves of the cerebral cortex to complete neural function regulation. In addition, 8 flexible electrodes are connected to the impedance detector of the circuit at the same time, and the working status of the flexible electrodes is analyzed by monitoring the change of impedance. Fig. 11 schematically shows a cross-sectional view of a flexible electrode and an electrical stimulator.

The choice of flexible electrode material is crucial. For the nerve signal acquisition function, the electrode should have a lower electrochemical impedance to obtain better signal amplification effect, that is, to obtain a higher signal-to-noise ratio of the nerve signal; for the electrical stimulation function, the electrode should have a higher The function of charge storage and release is to improve the efficiency of electrical stimulation, and the stability of the electrode should be further improved, that is, the dissolution of the electrode will not occur under the application of current. For the multimodal sensor of the present invention, metal Pt or Ir is used as the flexible electrode material.

The manufacturing method is that on the flexible substrate of the multi-modal sensor, metal Pt or Ir electrodes are fabricated by using the metal film lift off (Lift off) process, and the metal surface can be etched in a short time by using the physical plasma etching technology selectively ( less than 30sec) to increase the surface roughness, increase the surface specific area, reduce the electrochemical impedance or improve the storage and release capacity of the charge; further, a layer of PEDOT:PSS can be selectively deposited on the electrode to further reduce the electrochemical resistance. Impedance, the purpose of improving the signal-to-noise ratio, but this solution is not suitable for use as an electrical stimulator.

2. Small signal processing and wireless transmission circuit

The flexible multi-modal integrated sensor is connected to a row of contact points on one side of the substrate through the lead wires on the flexible substrate, and the flexible cables (made of polyimide, nylon silk and other flexible materials with high biocompatibility as the electrical insulation layer) pass through Through the skull space, one end is glued to the contact point of the lead wire of the intracranial sensor, and the other end is connected to the FPC/FFC interface on the side of the small circuit board placed outside the skull to realize the data and power transmission between the sensor and the circuit board .

The small circuit board includes multi-mode and multi-channel data acquisition analog terminals, power management, microprocessor, digital processing compression, radio frequency communication and other modules. Firstly, the conditioning function of multi-modal and multi-channel sensing signals (pressure, shear force, oxygen partial pressure, temperature, ion concentration, electrical signal) is realized, and its conditioning function mainly includes signals (current, voltage, resistance, capacitance, electrical signal) Chemical impedance) signal amplification and filtering, and then digitize the multi-channel analog signal through the digital-to-analog converter (ADC), and transmit it to the microprocessor for processing, encrypt and compress the data in the digital processing module, and finally transmit it through the radio frequency circuit In addition, the electrical stimulation module can output voltage/current dual-mode pulse width and amplitude-adjustable signals, and the electrical stimulation signals are applied to the electrodes through time-division multiplexing with the signal acquisition module. The overall block diagram of the system is shown in Figure 12.

The working principle of each module is introduced as follows:

(1) Power management module

The power detection module provides accurate and low-noise constant voltage sources for each module of the circuit system. The input of the module is a 3-6V lithium battery. Get ±3V, ±4.5V voltage output.

(2) Resistance detection module

This module detects the variable resistance signals of flexible pressure sensor, flexible shear force sensor and flexible temperature sensor. A voltage divider circuit or a Wheatstone bridge is used to convert the resistance signal into a voltage signal, and then the resistance signal is converted and output through a negative feedback voltage amplifier.

(3) Current detection module

This module detects the current output of the flexible oxygen partial pressure sensor. Under the action of a constant voltage source, the current intensity changes with the concentration of oxygen partial pressure. First, the micro-current signal passes through the transimpedance amplifier circuit to convert the micro-current signal into an amplified voltage. After the signal passes through the voltage follower, a voltage signal with low output impedance and high current driving capability is obtained, and then passes through a low-pass filter to improve the signal-to-noise ratio of the signal.

(4) Potential detection module

This module detects the concentration of potassium and sodium ions. The concentration of potassium and sodium ions is measured by detecting the potential difference between the working electrode (WE) and the reference electrode (RE). The input impedance between the electrodes is relatively large (usually hundreds of megabytes), so a circuit with high input impedance and low common-mode rejection ratio must be used to amplify the electrode signal to ensure that the signal is not distorted and reduce the measurement error of the system. First, the two-electrode system obtains a lower output impedance through a voltage follower, and then the two-electrode signals pass through a differential amplifier to obtain the potential difference between the electrodes with a higher common-mode rejection ratio, which improves the common-mode rejection ratio, and then passes through a low-pass filter device, which improves the signal-to-noise ratio of the signal.

(5) Neuroelectric detection module

Since the nerve signal is usually less than 10 μV, the present invention uses an instrumentation amplifier as the first stage of the amplifier to amplify the potential difference signal between the nerve electrode and the reference segment electrode in a differential form. Instrumentation amplifiers have extremely low DC offset, low noise, very high open-loop gain, very large common-mode rejection ratio, and high input impedance, making them ideal for the measurement of neural signals with noisy signals and high impedance. After the signal is amplified in the instrumentation amplifier, it is amplified by the second stage voltage amplifier, and the signal reaches the mV-V level. Finally, the notch filter removes power frequency interference, and the bandpass filter removes high and low frequency interference. There are many types of neural signals, the amplitude is from 5μV to 50mV, and the frequency range is from 0.5Hz to 3000Hz. Therefore, the present invention uses a voltage amplifier with adjustable magnification, and a band filter with independently adjustable upper and lower cut-off bands to suit various types of neural signals. Acquisition of neural signals.

(6) Impedance detection module

After long-term implantation in the body, the cells will produce a rejection reaction to the electrode, so it is very important to monitor the change of the electrode impedance value in real time. As the impedance increases, the signal-to-noise ratio of the detected nerve electrical signal will become lower and lower. As for the final failure. Therefore, by monitoring the impedance value of the electrode, the working stability of the device in vivo can be understood. The impedance detection proposed by the present invention utilizes the excitation signal to generate the current excited by the alternating current, uses a microprocessor to control the multi-channel selector switch, applies the current to multiple signal acquisition electrodes in time-sharing, and detects the impedance signal (voltage) generated by it Demodulate to obtain the magnitude and phase information of the impedance.

(7) Digital processing module and communication module

The multi-channel analog signal is time-divisionally input to the analog-to-digital converter (ADC) through the multiplexer, and the generated digital signal then enters the low-power micro-processing unit for data processing, because the nerve electrical signal collected is a continuous wave signal , the amount of data is large, you can optionally perform DPCM encoding and compression on the original signal after analog-to-digital conversion to remove the time correlation between the data, and then perform huffman encoding to remove the statistical correlation between the data to achieve the purpose of lossless data compression. Compression encoding for data transmission has the advantages of encryption and anti-environmental interference. The compression processing hardware unit can be realized by FPGA or DSP chip, and the compressed data is transmitted wirelessly by Bluetooth module, VUF or UHF module, and the sensing signals of each channel are time-shared. On the upper computer side, the radio frequency module is used to receive multi-channel sensing signals.

(8) Electrical stimulation module

The circuit in the present invention can output voltage and current electrical stimulation modes, wherein the voltage amplitude is 0-20V, the current amplitude is 0-2mA, the frequency is 0-1KHz, and the pulse width is adjustable. The frequency and pulse width of the electrical stimulation output are realized through the programming of the microprocessor, and then the signal is converted into an analog signal through a digital-to-analog conversion chip. For the voltage electrical stimulation mode, the signal enters the programmable boost module to adjust the output amplitude; for the current electrical stimulation mode, the signal passes through the voltage-current conversion module to convert the analog voltage signal into a current for output.

3. Disease type analysis and electrical stimulation prognosis

Traumatic brain injury has two stages, primary brain injury and secondary brain injury. Primary injury is the primary head injury caused by external physical injury to the head, such as impact, falling, extrusion, violent shaking, etc. After the onset of primary brain injury, complications caused by intracranial hemorrhage and tissue infection, such as intracranial edema, intracranial hypertension, epilepsy, local cerebral blood flow disturbance, metabolic and ion disturbance, etc. The degree of damage will eventually lead to complications such as loss of brain autoregulation function, brain tissue damage, endocrine dysfunction, systemic hypotension, hypoxemia, acid-base balance, and glucose metabolism disorders, which will endanger life.

According to the American Traumatic Brain Injury Foundation's recommendations for the treatment of severe traumatic brain trauma, severe brain injury with a Glasgow coma score of 8 or less requires real-time monitoring of intracranial parameters, at least monitoring of intracranial pressure. Often according to the hardware and software conditions of the hospital, the measurement of intracranial partial pressure of oxygen and intracranial temperature will be selectively increased on the basis of intracranial pressure monitoring. By monitoring brain tissue blood oxygen, temperature, intracranial pressure, and EEG, it can provide more information than simple intracranial pressure monitoring, so as to detect and prevent the occurrence of secondary brain injury early. The intracranial cortical electrical signals can provide highly authentic intracranial information, and help to discover cortical spreading inhibitory waves that are not obvious on traditional EEG, that is, depolarization slow waves.

The types of intracranial parameter detection proposed by the present invention include pressure, oxygen partial pressure, temperature, nerve electrical signal, potassium and sodium ion concentration, and the like. Multi-modal and multi-channel physiological signal detection can be used as the "eyes" of doctors to timely and accurately observe and judge the degree of brain damage of patients in real time, as well as the type of secondary brain damage that may occur, and propose relevant treatment plans, such as craniotomy and decompression treatment, hypothermia treatment, hypertonic dehydration treatment, nutrition treatment and anti-infection treatment and so on.

The detection range of each parameter in cerebrospinal fluid is as follows: Pressure: 5-50mmHg (normal value 6-13mmHg), oxygen partial pressure: 10-50mmHg (normal value 40-44mmHg), temperature: 20-50°C (normal value 36-37.5°C ), nerve electrical signal: 20μV-50mV, sodium ion concentration 100-200mmol/L (normal value 136-150mmol/L), potassium ion concentration 1.0-4.0mmol/L (normal value 2.5-3.2mmol/L). Predicted diseases include: intracranial hematoma, intracranial hypertension, partial epilepsy, ion metabolism disorder or mixed disease, and timely intervention and treatment can be given to improve the cure rate and reduce the mortality rate.

There are two methods of disease prediction proposed in the present invention.

1. Disease prediction model 1: artificial neural network

A neural network is an operational model consisting of a large number of nodes (or neurons) connected to each other. Each node represents a specific output function called an activation function. Each connection between two nodes represents a weighted value for the signal passing through the connection, called weight, which is equivalent to the memory of the artificial neural network. The output of the network is different according to the way the network is connected, the weight value and the activation function. The network itself is usually an approximation to a certain algorithm or function in nature, or it may be an expression of a logical strategy.

In the present invention, the model is divided into three layers, namely an input layer, a hidden layer and an output layer. The input layer is six intracranial sensor parameters, namely pressure, oxygen partial pressure, temperature, nerve electrical signal, sodium ion concentration and potassium ion concentration, and the output is a combination of various injury types. According to the combination of injury types in the output layer of the model, the intervention treatment combination is formulated, such as craniotomy decompression treatment, hypothermia treatment, hyperosmotic dehydration treatment, nutrition treatment and anti-infection treatment, etc. The algorithm block diagram of the disease prediction model 1 is shown in Figure 13.

2. Disease Prediction Model II: Random Forest

Random forest is an algorithm that integrates multiple trees through the idea of ensemble learning. Its basic unit is a decision tree, which is a classifier containing multiple decision trees, and its output category is the category output by individual trees. Depends on the number. Its essence belongs to a large branch of machine learning - integrated learning method.

In the present invention, multiple sub-decision numbers are constructed by using the random forest model, and each sub-decision number includes several sensor parameter binary classification nodes, and the arrangement and combination of nodes of different decision numbers are different. When six sensor parameter values are input to the random forest model, each decision number will determine a certain damage result, and the model will combine these damage results to give corresponding intervention methods. Disease prediction model 2 The block diagram of the algorithm is shown in Figure 14.

Based on the framework of the intelligent brain injury type recognition algorithm proposed by the invention, the multi-channel flexible electrodes and electrical stimulation function proposed by the invention can effectively use eight flexible electrodes placed in different spatial positions in the case of non-manual intervention treatment, Effectively locate the site of epilepsy with high precision, obtain the physical information of the epilepsy signal, judge the type of epilepsy, and automatically give different types of electrical stimulation feedback (different cycle, amplitude, voltage/current, duty cycle) to the epileptic focus, So as to effectively relieve or rule out epileptic complications, such as cerebral hemorrhage, vasospasm and so on. The closed-loop process of epilepsy detection and feedback treatment is shown in Figure 15.

The closed-loop process of epilepsy detection and feedback treatment is shown in Figure 16, which consists of a multi-channel mixed feature matrix fusion epilepsy detection algorithm and a multi-channel functional electrical stimulation epilepsy treatment algorithm. The epilepsy detection algorithm mainly includes three parts: data preprocessing, multi-channel mixed feature matrix construction, convolutional neural network feature compression and classification detection. The electrical stimulation feedback algorithm mainly includes three parts: the location of the lesion, the extraction of multiple features of the epileptic signal, and the encoding of the electrical stimulator.

Epilepsy detection algorithm:

(1) Data preprocessing: The data preprocessing part of the algorithm mainly includes two parts: abnormal frequency value elimination and noise removal. The cascade filter preprocessing method is used to preprocess intracranial multi-channel electrical signals.

(2) Multi-channel mixed feature matrix construction: This part considers the nonlinear characteristics of EEG signals and the spatial information of multi-channel intracranial electrical signals. First, select the appropriate wavelet function to decompose the intracranial electrical signals in the full frequency range, and reconstruct the EEG signals in different frequency bands. Then, the classical features of the signal at different frequencies are calculated, and the channel space information is fused to construct the feature matrix of the signal.

(3) Feature compression and epilepsy classification: This part considers that the multi-domain features of EEG signals are not absolutely independent, and considers the spatial information of the channel, so a convolutional neural network is introduced to perform multi-domain feature matrix The fusion and learning of features can dig deep into the hidden information contained in the features, and at the same time, the final classification and detection of the fused features can be performed.

Electrical stimulation feedback algorithm:

(1) Positioning. The multi-channel signal is judged by the convolutional neural network. If it is judged to be an epileptic seizure, the area corresponds to the potential electrical stimulation feedback area.

(2) calculation. Calculate multiple features of epileptic signals, judge different types of epileptic seizures, and formulate corresponding electrical stimulation feedback therapy.

(3) Coding. Spatial encoding of multi-channel electrical stimulators, as well as encoding of frequency, amplitude, duty cycle, etc.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention, and are not intended to limit the present invention. Within the spirit and principles of the present invention, any modifications, equivalent replacements, improvements, etc., shall be included in the protection scope of the present invention.

Claims (10)

1. An intracranial implantable flexible multimodal physiological and biochemical information monitoring device is characterized in that it includes a multimodal sensor, a flexible cable and a signal processing and wireless transmission circuit, and the multimodal sensor includes a pressure sensor, A shear force sensor, a temperature sensor, an oxygen partial pressure sensor, a potassium and sodium ion sensor and a flexible electrode; the multimodal sensor and the signal processing and wireless transmission circuit are connected through the flexible cable; Wherein, the multimodal sensor also includes a curved lead wire, the lead wire is used to connect the sensor and the connector at the end of the flexible substrate, the pressure sensor, shear force sensor, temperature sensor, oxygen partial pressure sensor , a potassium and sodium ion sensor, flexible electrodes and the lead wires are arranged on the flexible base, and the flexible base has a pattern matching the shape of the lead wires, so that the lead wires are embedded with the on a flexible substrate; the flexible substrate has a material modulus of elasticity matching that of brain tissue; The temperature sensor is wound around a square, and the horizontal and vertical wires have the same spacing. 2. The intracranial implantable flexible multimodal physiological and biochemical information monitoring device according to claim 1, wherein the pressure sensor is used to detect the pressure value in the brain, and the pressure sensor adopts resistive pressure The sensor or capacitive pressure sensor uses a sandwich structure of upper and lower electrode plates and interlayer materials for pressure sensing. 3. The intracranial implantable flexible multi-modal physiological and biochemical information monitoring device according to claim 1, wherein the shear force sensor is used to monitor the shear force on the flexible substrate, using left and right parallel electrode plates Cooperate with the structure of the intermediate sensitive material for shear force sensing. 4. The intracranial implantable flexible multi-modal physiological and biochemical information monitoring device according to claim 1, wherein the temperature sensor is used for real-time monitoring of the intracranial temperature, and metal Pt or Au is used as the temperature sensor . 5. The intracranial implantable flexible multi-modal physiological and biochemical information monitoring device according to claim 1, wherein if the multi-modal sensor is placed above the cerebral cortex, the oxygen partial pressure sensor conducts cerebrospinal fluid For the measurement of oxygen concentration, if the multimodal sensor is inserted into the cerebral cortex, the oxygen partial pressure sensor can measure the oxygen concentration in the interstitial fluid inside the cerebral cortex. 6. The intracranial implantable flexible multi-modal physiological and biochemical information monitoring device according to claim 1, wherein if the multi-modal sensor is placed above the cerebral cortex, the potassium and sodium ion sensor conducts cerebrospinal fluid For the measurement of the concentration of potassium and sodium ions, if the multimodal sensor is inserted into the cerebral cortex, the potassium and sodium ion sensor will measure the concentration of potassium and sodium ions in the interstitial fluid inside the cerebral cortex. 7. The intracranial implantable flexible multimodal physiological and biochemical information monitoring device according to claim 1, wherein the multimodal sensor integrates no less than 8 flexible electrodes, which are evenly distributed above the base, The flexible electrode has the functions of electrical signal measurement and electrical stimulator at the same time. If the multimodal sensor is placed above the cerebral cortex, the flexible electrode will measure the cerebral cortex nerve signal or perform electrical stimulation on the cerebral cortex nerve. If the multimodal sensor is inserted into the cerebral cortex, the flexible electrode can measure nerve signals or perform electrical stimulation on the nerves inside the cerebral cortex to complete neural function regulation. 8. The intracranial implantable flexible multimodal physiological and biochemical information monitoring device according to claim 1, wherein the multimodal sensor uses metal Pt or Ir as the flexible electrode material. 9. The intracranial implantable flexible multimodal physiological and biochemical information monitoring device according to claim 1, wherein the multimodal sensor also includes a flexible substrate, and the elastic modulus of the flexible substrate material is the same as that of the brain. Tissue matching such that the flexible base material is commensurate with the cortex. 10. The intracranial implantable flexible multi-modal physiological and biochemical information monitoring device according to claim 1, wherein the material of the lead wire includes solid metal material, degradable metal material, nanocomposite conductive material and liquid conductive metals.