A method of downsampling neuronal membrane potential data

Abstract

The application discloses a kind of neuron membrane potential data downsampling methods, applied to pulse neuron simulation, comprising: the original acquisition membrane potential data is divided into bucket;Filtering local strict extreme point in bucket;If there is no local strict extreme point in bucket, using the first and last membrane potential data in bucket represent all the membrane potential data in bucket;Otherwise, using the first membrane potential data, the last membrane potential data and all or part of local strict extreme points represent all the membrane potential data in bucket;Collect representative data in all buckets to form the final downsampling data.The application realizes the premise of effectively reducing the total amount of membrane potential data to the greatest extent preserving the variation characteristics of membrane potential data.

Description

Technical Field

[0001] This invention belongs to the field of neuron simulation technology, specifically relating to a method for downsampling neuron membrane potential data. Background Technology

[0002] In neuron simulations, information is transmitted within and between neurons via electrical and chemical signals. Pulse signals are particularly important for the rapid, long-distance transmission of time-sensitive information. However, pulse signals are very brief and difficult to capture. Therefore, during simulations, the acquisition interval for neuronal membrane potential data is typically set very short to maximize the recording of neuronal membrane potential data and thus preserve its pulse characteristics.

[0003] However, due to the extremely short acquisition intervals of neuronal membrane potential data, the acquired data expands linearly as the acquisition interval decreases. For example, in typical research with an acquisition interval of 0.1 milliseconds, a single neuron generates 10,000 data points within a one-second simulation period. Of these 10,000 data points, less than one percent may exhibit pulse characteristics. This clearly poses a significant challenge to both data acquisition and analysis. Furthermore, in the field of whole-brain simulation, the number of neurons involved can reach hundreds of billions, requiring the recording of hundreds or even thousands of membrane potentials, with simulation times sometimes exceeding hours. This presents a significant challenge to the visualization of membrane potential data. Therefore, a downsampling method is urgently needed to effectively reduce the data volume while preserving the characteristics of membrane potential changes to the maximum extent.

[0004] Currently, the commonly used data downsampling methods include the following:

[0005] 1) Mode-Median-Bucket Downsampling Method: This method first determines the number of buckets for all data points. Except for the first and last buckets which contain the first and last data points respectively, each bucket contains approximately the same number of original data points. Then, it iterates through each bucket and checks if there is a unique mode in the data points within the bucket. If there is, this unique mode is used to represent all data points in the bucket; otherwise, the median is used to represent all data points in the bucket. Finally, the representative data from all buckets are collected to form the final downsampled data.

[0006] 2) Minimum Standard Deviation Bucket Downsampling Method (Min-Std-Error-Bucket): This method first determines the number of buckets for all data points. Except for the first and last buckets which contain the first and last data points respectively, each bucket contains approximately the same number of original data points. Then, it connects the points in each bucket to the points in the next bucket in pairs to form line segments. It calculates the distance from the line segment to the points whose horizontal projection is contained within the line segment. Using these distances, it applies the SEE formula to calculate the standard error as the weight of the line segment. Finally, it treats the line segment with weight information as a directed acyclic graph and applies the shortest path algorithm to obtain all the data points that constitute the shortest path as the final downsampled data.

[0007] 3) Largest Triangle-Three-Buckets Downsampling Method: This method first determines the number of buckets for all data points. Except for the first and last buckets which contain the first and last data points respectively, each of the remaining buckets contains approximately the same number of original data points. Then, starting from the second bucket, it iterates through each data point in the bucket and calculates the area of ​​the triangle formed by the points in the bucket, the points in the previous bucket, and all data points in the next bucket using an averaging algorithm. The triangle with the largest area is selected to represent all data points in the bucket. Finally, the representative data from all buckets are collected to form the final downsampled data.

[0008] While the commonly used data downsampling methods mentioned above can reduce the amount of data, they ignore the changes in membrane potential data during the processing, resulting in the final downsampled data not effectively retaining the changing characteristics of the membrane potential data. Summary of the Invention

[0009] One of the objectives of this invention is to provide a method for downsampling neuronal membrane potential data, which can preserve the variation characteristics of membrane potential data to the greatest extent while effectively reducing the total amount of membrane potential data.

[0010] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0011] A method for downsampling neuronal membrane potential data, applied to spiking neuron simulation, wherein the method for downsampling neuronal membrane potential data includes:

[0012] The raw membrane potential data is divided into buckets;

[0013] Local strict extrema within the sieving bucket;

[0014] If there are no local strict extrema in the bucket, the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket; otherwise, the first membrane potential data, the last membrane potential data, and all or some of the local strict extrema in the bucket are used to represent all the membrane potential data in the bucket.

[0015] Collect representative data from all buckets to form the final downsampled data.

[0016] Several alternative methods are provided below, but they are not intended as additional limitations on the overall solution above. They are merely further additions or optimizations. Provided there are no technical or logical contradictions, each alternative method can be combined individually with respect to the overall solution above, or multiple alternative methods can be combined with each other.

[0017] Preferably, the method further includes establishing a Cartesian coordinate system for visualizing membrane potential data, wherein the horizontal axis of the Cartesian coordinate system is the simulation time, and the vertical axis is the membrane potential corresponding to the simulation time.

[0018] Preferably, the process of binning the raw acquired membrane potential data includes:

[0019] The original acquired membrane potential data were sorted according to the simulation time sequence;

[0020] The first sorted membrane potential data is assigned to the first bucket, the last sorted membrane potential data is assigned to the last bucket, and the remaining membrane potential data is assigned to the remaining buckets.

[0021] Preferably, the step of allocating the remaining membrane potential data to the remaining buckets includes:

[0022] Let there be N buckets and M membrane potential data to be allocated;

[0023] Then the amount of data allocated to the 2nd to the (N-1)th buckets is (M-2) / (N-2);

[0024] Using (M-2) / (N-2) data points as the dividing unit, the remaining membrane potential data are divided sequentially according to the simulation time sequence, and the divided membrane potential data are then distributed sequentially according to the bucket sequence.

[0025] Preferably, the local strict extrema points in the screening bucket include:

[0026] Starting from the second bucket, by combining the last membrane potential data in the previous bucket with the first membrane potential data in the next bucket, it is determined whether each membrane potential data in the bucket is a local strict extremum point, thus completing the screening.

[0027] As a preferred method, the method for determining the local strict extremum point is as follows:

[0028] Take the membrane potential data to be judged as the current data point, take the membrane potential data before the current data point as the previous data point, and take the membrane potential data after the current data point as the next data point.

[0029] The slope of the line connecting the current data point and the previous data point is calculated as the first slope, and the slope of the line connecting the current data point and the next data point is calculated as the second slope.

[0030] If the product of the first slope and the second slope is less than zero, then the current data point is determined to be a local strict extremum; otherwise, the current data point is determined not to be a local strict extremum.

[0031] Preferably, the use of the first membrane potential data, the last membrane potential data, and all or part of the local strict extrema in the bucket to represent all membrane potential data in the bucket includes:

[0032] If the bucket contains one or two local strict extrema, then the first membrane potential data, the last membrane potential data, and all the local strict extrema in the bucket are used to represent all the membrane potential data in the bucket.

[0033] If the bucket contains more than two local strict extrema, then the maximum local maximum and minimum local minimum are selected from the more than two local strict extrema, and the first membrane potential data, the last membrane potential data, the maximum local maximum and the minimum local minimum in the bucket are used to represent all the membrane potential data in the bucket.

[0034] Preferably, if the bucket contains one or two local strict extrema, then the first membrane potential data, the last membrane potential data, and all the local strict extrema in the bucket are used to represent all the membrane potential data in the bucket, including:

[0035] When a bucket contains a local strict extremum, if the local strict extremum coincides with the first or last membrane potential data, then the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

[0036] When a bucket contains two local strict extrema, if one of the local strict extrema coincides with the first or last membrane potential data, then the first membrane potential data, the last membrane potential data, and the other local strict extrema in the bucket are used to represent all the membrane potential data in the bucket.

[0037] When a bucket contains two local strict extrema, if one of the local strict extrema coincides with the first membrane potential data and the other local strict extrema coincides with the last membrane potential data, then the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

[0038] Preferably, the step of selecting the maximum local maximum point and the minimum local minimum point from more than two local strict extrema points includes:

[0039] Take the local strict extreme point to be judged as the current extreme point, take the membrane potential data before the current extreme point as the previous extreme point, and take the membrane potential data after the current extreme point as the next extreme point.

[0040] The slope of the line connecting the current extreme point and the previous extreme point is calculated as the third slope, and the slope of the line connecting the current extreme point and the next extreme point is calculated as the fourth slope.

[0041] If the third slope is greater than zero and the fourth slope is less than zero, then the local strict extremum point to be judged is a local maximum point; if the third slope is less than zero and the fourth slope is greater than zero, then the local strict extremum point to be judged is a local minimum point.

[0042] The point with the largest membrane potential among all local maxima is taken as the maximum local maximum, and the point with the smallest membrane potential among all local minima is taken as the minimum local minimum.

[0043] Preferably, the use of the first membrane potential data, the last membrane potential data, the maximum local maximum point, and the minimum local minimum point in the bucket to represent all membrane potential data in the bucket includes:

[0044] If one of the two points, the maximum local maximum point and the minimum local minimum point, coincides with the first or last membrane potential data point, then the first membrane potential data point, the last membrane potential data point, and the other point in the bucket are used to represent all the membrane potential data in the bucket.

[0045] If one of the two points, the maximum local maximum and the minimum local minimum, coincides with the first membrane potential data and the other coincides with the last membrane potential data, then the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

[0046] This invention provides a method for downsampling neuronal membrane potential data. It uses a bucketing approach to segment the original data, thereby effectively controlling the upper limit of the data. In addition, it finds local strict extrema in all membrane potential data in each bucket, and uses the first membrane potential data, the last membrane potential data in the bucket, plus all or some of the local strict extrema in the bucket to represent all membrane potential data in the bucket. This achieves the goal of preserving the data features of mutations to the maximum extent while effectively reducing the total amount of membrane potential data. Attached Figure Description

[0047] Figure 1 This is a flowchart of a method for downsampling neuronal membrane potential data according to the present invention;

[0048] Figure 2 This is a schematic diagram of the bucket division of the present invention;

[0049] Figure 3 This is a schematic diagram illustrating the visualization of membrane potential data using a Cartesian coordinate system as described in this invention.

[0050] Figure 4 This is a schematic diagram illustrating the determination of local strict extrema points according to the present invention;

[0051] Figure 5 This is a schematic diagram showing that there are no local strict extrema in the bucket of the present invention;

[0052] Figure 6 This is a schematic diagram showing the existence of local strict extrema points in the bucket of the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0055] Example 1

[0056] like Figure 1 As shown, in order to effectively downsample the membrane potential data of neurons in spiking neuron simulation, this embodiment provides a method for downsampling neuron membrane potential data applied to spiking neuron simulation, specifically including the following steps:

[0057] (1) The original membrane potential data is divided into buckets.

[0058] Since binning can effectively guarantee the upper limit of the data after final downsampling, this embodiment is based on the binning concept for downsampling. First, the number of bins for all data points (for ease of description, this embodiment describes one membrane potential data point as one data point) is determined (generally, the number of bins is much smaller than the number of data points, otherwise sampling is meaningless), so that all data points are distributed in each bin.

[0059] When dividing the data into buckets, the original collected membrane potential data are first sorted according to the simulation time sequence. Then, the first sorted membrane potential data is assigned to the first bucket, the last sorted membrane potential data is assigned to the last bucket, and the remaining membrane potential data is assigned to the remaining buckets.

[0060] When allocating the remaining membrane potential data to the remaining buckets, the data can be divided equally among the remaining buckets, or it can be allocated according to a preset rule. To ensure the downsampling effect, in this embodiment, the remaining membrane potential data is evenly distributed among the remaining buckets.

[0061] like Figure 2 As shown, if the number of buckets is N and the number of membrane potential data points to be allocated is M, then the amount of data allocated to the 2nd to the (N-1)th buckets is (M-2) / (N-2). During allocation, the remaining membrane potential data is divided into units of (M-2) / (N-2) data points, following the simulation time sequence, and then allocated according to the bucket order. For example, when 10 data points are divided into 4 buckets, the first data point is allocated to the first bucket, the second to fifth data points to the second bucket, the sixth to ninth data points to the third bucket, and the tenth data point to the fourth bucket.

[0062] In addition, to facilitate the visualization of membrane potential data during simulation, in another embodiment, a Cartesian coordinate system can be established for visualizing the membrane potential data, such as... Figure 3 As shown, the horizontal axis of the Cartesian coordinate system represents the simulation time (in milliseconds, composed of consecutive simulation moments), and the vertical axis represents the membrane potential corresponding to the simulation moment (in millivolts). Furthermore, to facilitate the demonstration of binning applications, this embodiment uses a Cartesian coordinate system for illustration in the binning and subsequent downsampling principle explanations.

[0063] (2) Local strict extrema in the screening bucket.

[0064] To effectively preserve data variation characteristics, this embodiment filters for local strict extrema among all membrane potential data in each bucket. Since the first and last buckets contain only one data point, this embodiment starts from the second bucket, combining the last membrane potential data from the previous bucket with the first membrane potential data from the next bucket to determine whether each membrane potential data point in the bucket is a local strict extrema, thus completing the filtering.

[0065] The method for determining local strict extrema of membrane potential data in a bucket is as follows: the product of the slope of a data point relative to its preceding data point and the slope of its following data point is less than zero. The specific determination is described below:

[0066] Take the membrane potential data to be judged as the current data point, take the membrane potential data before the current data point as the previous data point, and take the membrane potential data after the current data point as the next data point.

[0067] The slope of the line connecting the current data point and the previous data point is calculated as the first slope, and the slope of the line connecting the current data point and the next data point is calculated as the second slope.

[0068] If the product of the first slope and the second slope is less than zero, then the current data point is determined to be a local strict extremum; otherwise, the current data point is determined not to be a local strict extremum.

[0069] like Figure 4 As shown, in bucket 2 (the second bucket), the membrane potential data to be judged is data point C, with coordinates (x2, y2). The preceding data point is data point B, with coordinates (x1, y1). The following data point is data point D, with coordinates (x3, y3). Therefore, the slope product calculated based on data point C according to the judgment rule is as follows:

[0070]

[0071] It is obvious that the slope of line segment BC is positive, and the slope of line segment CD is also positive. Therefore, their product is greater than zero, which does not meet the rules for determining local strict extrema. Thus, data point C is not a local strict extrema.

[0072] (3) If there are no local strict extrema in the bucket, the first and last membrane potential data in the bucket are used as representative data to represent all membrane potential data in the bucket; otherwise, the first membrane potential data, the last membrane potential data, and all or some local strict extrema in the bucket are used as representative data to represent all membrane potential data in the bucket.

[0073] Local strict extrema can effectively characterize the data change features. Therefore, depending on whether there are local strict extrema within the bucket, this embodiment provides different data adoption logics to effectively preserve the change trend of the data within the bucket.

[0074] (3.1) If no local strict extremum is found within the bucket, the first and last data points in the bucket are used to represent all data points in the bucket, such as... Figure 5 As shown, there are no local strict extrema within bucket 2, so data point B and data point F are used to represent all data points within bucket 2.

[0075] (3.2) If there are local strict extrema within the bucket, in order to effectively control the final amount of downsampled data, when there are too many local strict extrema, it is also necessary to filter the local Yang Ge extrema.

[0076] (3.2.1) If the bucket contains one or two local strict extrema, then the first membrane potential data, the last membrane potential data, and all the local strict extrema in the bucket are used to represent all the membrane potential data in the bucket.

[0077] In this embodiment, when a bucket contains only one or two local strict extrema, all local strict extrema are retained to preserve data variation characteristics to the greatest extent. Furthermore, since the selection of local strict extrema is based on each data point within the bucket, the possibility of overlap between local strict extrema and the first or last data point in the bucket cannot be ruled out. Therefore, considering overlap, the selection rules for representative data are as follows:

[0078] A. When a bucket contains a local strict extremum, if the local strict extremum coincides with the first or last membrane potential data, then the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

[0079] B. When the bucket contains two local strict extrema, if one of the local strict extrema coincides with the first membrane potential data or the last membrane potential data, then the first membrane potential data, the last membrane potential data, and the other local strict extrema in the bucket are used to represent all the membrane potential data in the bucket.

[0080] C. When the bucket contains two local strict extrema, if one of the local strict extrema coincides with the first membrane potential data and the other local strict extrema coincides with the last membrane potential data, then the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

[0081] (3.2.2) If the bucket contains more than two local strict extrema, then the maximum local maximum and minimum local minimum are selected from the more than two local strict extrema, and the first membrane potential data, the last membrane potential data, the maximum local maximum and the minimum local minimum in the bucket are used to represent all the membrane potential data in the bucket.

[0082] In this embodiment, if the bucket contains more than two local strict extrema, then the local strict extrema need to be filtered to select one maximum local maximum and one minimum local minimum. The filtering process is as follows:

[0083] Take the local strict extreme point to be judged as the current extreme point, take the membrane potential data before the current extreme point as the previous extreme point, and take the membrane potential data after the current extreme point as the next extreme point.

[0084] The slope of the line connecting the current extreme point and the previous extreme point is calculated as the third slope, and the slope of the line connecting the current extreme point and the next extreme point is calculated as the fourth slope.

[0085] If the third slope is greater than zero and the fourth slope is less than zero, then the local strict extremum point to be judged is a local maximum point. If the third slope is less than zero and the fourth slope is greater than zero, then the local strict extremum point to be judged is a local minimum point.

[0086] The point with the largest membrane potential among all local maxima is taken as the maximum local maximum, and the point with the smallest membrane potential among all local minima is taken as the minimum local minimum.

[0087] like Figure 6 As shown, there are three local strict extrema in bucket 2, namely data point C, point D and point E. Since data point C and data point E are both local maxima and the ordinate value of data point C is less than that of data point E, data point E is selected as the maximum local maximum. Data point D is the only local minimum and is selected as the minimum local minimum. Finally, data points B, D, E and F are used to represent all data points in bucket 2, and data point C is excluded.

[0088] It should be noted that, in order to clearly describe the principle, this embodiment introduces the first slope and the second slope in the determination of local strict extrema, and the third slope and the fourth slope in the screening of local strict extrema. However, in actual calculation, for a membrane potential data, the third slope and the first slope are the same, and the fourth slope and the second slope are the same, which can be directly reused without repeated calculation.

[0089] Similarly, considering that the maximum and minimum local maxima and minimum local minima may overlap with the first and last data points in the bucket, this embodiment proposes the following selection rule for representative data after considering the overlap:

[0090] A. If one of the two points, the maximum local maximum and the minimum local minimum, coincides with the first or the last membrane potential data, then the first membrane potential data, the last membrane potential data, and the other point in the bucket are used to represent all the membrane potential data in the bucket.

[0091] B. If one of the two points, the maximum local maximum and the minimum local minimum, coincides with the first membrane potential data and the other coincides with the last membrane potential data, then the first and last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

[0092] (4) Collect representative data from all buckets to form the final downsampled data.

[0093] The method in this embodiment is fast, using only subtraction, multiplication, and division operations; it also has low complexity, is easy to understand and implement on a computer; and it can preserve the data characteristics of abrupt changes to the maximum extent while effectively reducing the total amount of membrane potential data.

[0094] Example 2

[0095] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method for downsampling neuronal membrane potential data of the present invention.

[0096] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0097] Any references to memory, storage, database, or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0098] Example 3

[0099] A computer device includes a processor and a memory storing a plurality of computer instructions, wherein the computer instructions, when executed by the processor, implement the steps of the method for downsampling neuronal membrane potential data of the present invention.

[0100] The memory and processor are electrically connected directly or indirectly to enable data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines. The memory stores a computer program that can run on the processor, which implements the method of the present invention by running the computer program stored in the memory.

[0101] The memory may be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory stores the program, and the processor executes the program upon receiving an execution instruction.

[0102] The processor may be an integrated circuit chip with data processing capabilities. The aforementioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0103] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0104] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A method of down-sampling of neuronal membrane potential data, applied to spiking neuron simulation, characterized in that, The downsampling method of the neuron membrane potential data comprises: The original collected membrane potential data is divided into buckets, comprising: sorting the original collected membrane potential data according to the simulation time sequence; assigning the first membrane potential data after sorting to the first bucket, assigning the last membrane potential data after sorting to the last bucket, and assigning the remaining membrane potential data to the remaining buckets; wherein the assigning of the remaining membrane potential data to the remaining buckets comprises: letting the number of buckets be N and the number of membrane potential data to be assigned be M; then the data amount assigned to the 2nd to N-1th buckets is (M-2) / (N-2); the remaining membrane potential data is divided according to the simulation time sequence in (M-2) / (N-2) data units, and the divided membrane potential data is assigned according to the sequence of the buckets; Local strict extreme points in the buckets are screened; the judgment method of the local strict extreme points is as follows: taking the membrane potential data to be judged as a current data point, taking the previous membrane potential data of the current data point as a previous data point, and taking the next membrane potential data of the current data point as a next data point; calculating the slope of the line connecting the current data point and the previous data point as a first slope, and calculating the slope of the line connecting the current data point and the next data point as a second slope; if the product of the first slope and the second slope is less than zero, it is judged that the current data point is a local strict extreme point; otherwise, it is judged that the current data point is not a local strict extreme point; If there is no local strict extreme point in the bucket, the first and last membrane potential data in the bucket represent all the membrane potential data in the bucket; otherwise, the first membrane potential data, the last membrane potential data, and all or part of the local strict extreme points in the bucket represent all the membrane potential data in the bucket. The representative data in all the buckets are collected to form the final downsampling data.

2. The method of downsampling neuronal membrane potential data of claim 1, wherein, It also includes establishing a rectangular coordinate system for visualizing the membrane potential data, wherein the abscissa of the rectangular coordinate system is the simulation time, and the ordinate is the membrane potential corresponding to the simulation time.

3. The method of downsampling neuronal membrane potential data of claim 1, wherein, The screening of the local strict extreme points in the buckets comprises: Starting from the second bucket, the last membrane potential data in the previous bucket and the first membrane potential data in the next bucket are combined to judge whether each membrane potential data in the bucket is a local strict extreme point, and the screening is completed.

4. The method of downsampling neuronal membrane potential data of claim 1, wherein, The use of the first membrane potential data, the last membrane potential data, and all or part of the local strict extreme points in the bucket to represent all the membrane potential data in the bucket comprises: If the bucket contains one or two local strict extreme points, the first membrane potential data, the last membrane potential data, and all local strict extreme points in the bucket are used to represent all the membrane potential data in the bucket; If the bucket contains more than two local strict extreme points, the maximum local maximum point and the minimum local minimum point are selected from the more than two local strict extreme points, and the first membrane potential data, the last membrane potential data, the maximum local maximum point, and the minimum local minimum point in the bucket are used to represent all the membrane potential data in the bucket.

5. The method of downsampling neuronal membrane potential data of claim 4, wherein, If the bucket contains one or two local strict extreme points, the first membrane potential data and the last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket, including: If the bucket contains one local strict extreme point, and the local strict extreme point coincides with the first membrane potential data or the last membrane potential data, the first membrane potential data and the last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket; If the bucket contains two local strict extreme points, and one of the local strict extreme points coincides with the first membrane potential data or the last membrane potential data, the first membrane potential data, the last membrane potential data and the other local strict extreme point are used to represent all the membrane potential data in the bucket; If the bucket contains two local strict extreme points, and one of the local strict extreme points coincides with the first membrane potential data, and the other local strict extreme point coincides with the last membrane potential data, the first membrane potential data and the last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

6. The method of downsampling neuronal membrane potential data of claim 4, wherein, The filtering of the maximum local maximum point and the minimum local minimum point from the local strict extreme points greater than two includes: Taking the local strict extreme point to be judged as a current extreme point, taking the membrane potential data before the current extreme point as a previous extreme judgment point, and taking the membrane potential data after the current extreme point as a next extreme judgment point; The slope of the line connecting the current extreme point and the previous extreme judgment point is calculated as a third slope, and the slope of the line connecting the current extreme point and the next extreme judgment point is calculated as a fourth slope; If the third slope is greater than zero and the fourth slope is less than zero, the local strict extreme point to be judged is a local maximum point, and if the third slope is less than zero and the fourth slope is greater than zero, the local strict extreme point to be judged is a local minimum point; The maximum local maximum point is taken as the maximum membrane potential value in all local maximum points, and the minimum local minimum point is taken as the minimum membrane potential value in all local minimum points.

7. The method of downsampling neuronal membrane potential data of claim 4, wherein, The first membrane potential data, the last membrane potential data, the maximum local maximum point and the minimum local minimum point in the bucket are used to represent all the membrane potential data in the bucket, including: If one of the maximum local maximum point and the minimum local minimum point coincides with the first membrane potential data or the last membrane potential data, the first membrane potential data, the last membrane potential data and the other point are used to represent all the membrane potential data in the bucket; If one of the maximum local maximum point and the minimum local minimum point coincides with the first membrane potential data, and the other point coincides with the last membrane potential data, the first membrane potential data and the last membrane potential data in the bucket are used to represent all the membrane potential data in the bucket.

Images