A method, device, electronic device and storage medium for processing brain electrical signal channels

Abstract

The present invention relates to an EEG signal channel processing method, device, electronic device and storage medium, comprising: obtaining a target EEG signal data set corresponding to a target object; the target EEG signal data set includes multiple EEG signal samples of the target object, each EEG signal sample includes multiple channels; calculating the target information gain of each channel in the target EEG signal data set; based on the target information gain of each channel, determining a channel combination consisting of multiple channels ranked top by the target information gain in a preset manner as a target channel combination. The present invention determines the target channel combination based on the information gain of each EEG signal channel, can quickly and accurately select effective channels in the EEG signal, and filter out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time performance of the brain-computer interface system, and improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities.

Description

Technical Field

[0001] The present invention relates to the field of brain-computer interface, and in particular to a method, device, electronic device and storage medium for processing an electroencephalogram (EEG) signal channel. Background Art

[0002] Brain-computer interface technology has made great progress in clinical and research fields. However, when recording EEG signals, due to differences in implantation technology and individual test animals, some channels often record abnormally. The existence of these abnormal channels may lead to a decrease in the quality of EEG signals and affect the accurate analysis of brain activity.

[0003] At the same time, EEG signals have the characteristics of large number of channels and high sampling rate. Such high-density and high-frequency signal acquisition consumes a lot of computing resources. In practical applications, this may lead to insufficient computing power of the brain-computer interface system, resulting in a decline in real-time performance, and limiting the effective application of brain-computer interface technology. Summary of the invention

[0004] In order to solve the above technical problems, the present invention discloses a method for processing EEG signal channels, which determines the target channel combination based on the information gain of each EEG signal channel, so that effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0005] In order to achieve the above object, the present invention provides a method for processing an EEG signal channel, characterized in that the method comprises:

[0006] Acquire a target EEG signal data set corresponding to a target object; the target EEG signal data set includes a plurality of EEG signal samples of the target object, each EEG signal sample includes a plurality of channels;

[0007] Calculating the target information gain of each channel in the target EEG signal data set;

[0008] Based on the target information gain of each channel, a channel combination consisting of a plurality of channels ranked high in the target information gain is determined as a target channel combination in a preset manner.

[0009] In an optional embodiment, based on the target information gain of each channel, determining a channel combination consisting of a plurality of channels ranked higher in the target information gain as a target channel combination in a preset manner includes:

[0010] According to the order of the target information gain from large to small, the channels corresponding to the target information gain are sorted to obtain an EEG signal channel sequence;

[0011] Determining a plurality of channel combinations based on the EEG signal channel sequence; the number of channels respectively included in the plurality of channel combinations is increasing, the plurality of channels included in each channel combination are continuous in the EEG signal channel sequence, and each channel combination includes the first channel in the EEG signal channel sequence;

[0012] Determine the indicator information corresponding to each of the plurality of channel combinations; the indicator information represents the prediction performance of each channel combination in the target prediction task;

[0013] Based on the indicator information, the target channel combination is determined from the multiple channel combinations.

[0014] In an optional embodiment, the calculating the target information gain of each channel in the target EEG signal data set includes:

[0015] Determine the data division points corresponding to each of the channels; each channel in the target EEG signal data set has a plurality of data division points;

[0016] Based on the data division points, the information gain of each channel is calculated respectively; each channel in the target EEG signal data set has multiple information gains, and the multiple information gains of each channel correspond to the multiple data division points of each channel one by one;

[0017] The maximum information gain among the multiple information gains of the channels is determined as the target information gain of the channels.

[0018] In an optional embodiment, determining the data division points corresponding to the channels respectively includes:

[0019] Sorting the EEG signal values ​​included in each channel respectively to obtain a plurality of EEG signal value sequences;

[0020] Calculate the average value of every two adjacent EEG signal values ​​in each EEG signal value sequence;

[0021] The average value is determined as the data division point corresponding to each channel.

[0022] In an optional embodiment, the calculating the information gain of each channel based on the data partition point includes:

[0023] Calculating a first information entropy of the target EEG signal data set;

[0024] Based on the data division points, the EEG signal values ​​included in each channel are divided to obtain a plurality of division intervals;

[0025] Based on the EEG signal samples respectively corresponding to the multiple divided intervals, calculating the information entropy respectively corresponding to the multiple divided intervals;

[0026] Data fusion processing is performed on the first information entropy and the information entropies corresponding to the multiple divided intervals to obtain the information gain of each channel.

[0027] In an optional embodiment, determining the indicator information corresponding to the plurality of channel combinations respectively includes:

[0028] Based on the EEG signal samples corresponding to each channel combination, determining a training set and a test set corresponding to each channel combination;

[0029] Based on the training sets corresponding to each channel combination, a machine learning model corresponding to each channel combination is established;

[0030] Calculating a first accuracy rate of the machine learning model in a training set respectively corresponding to the machine learning model;

[0031] Calculating second accuracy rates of the machine learning models in test sets respectively corresponding to the machine learning models;

[0032] Performing data fusion processing on the first accuracy rate and the second accuracy rate to obtain a target accuracy rate of the machine learning model;

[0033] The indicator information corresponding to each channel combination is determined based on the target accuracy and the calculation time of the target accuracy.

[0034] In an optional embodiment, determining the indicator information corresponding to the plurality of channel combinations respectively based on the target accuracy and the calculation time of the target accuracy includes:

[0035] The difference between the target accuracy and the calculation time of the target accuracy is determined as the indicator information respectively corresponding to the multiple channel combinations.

[0036] In an optional embodiment, the method further includes:

[0037] When the values ​​of the electroencephalogram signals transmitted in the plurality of channels change, acquiring updated electroencephalogram signal samples based on the channel identifiers;

[0038] The target EEG signal dataset is updated based on the updated EEG signal samples to obtain an updated EEG signal dataset;

[0039] The updated EEG signal dataset is determined as the target EEG signal dataset.

[0040] In an optional embodiment, the method further includes:

[0041] After the target channel combination is determined, the EEG signal of the target object is transmitted through each channel included in the target channel combination.

[0042] In an optional embodiment, the method further includes:

[0043] After determining the target channel combination, the target system decodes the EEG signal of the target object based on each channel included in the target channel combination.

[0044] The present invention also provides an electroencephalogram signal channel processing device, characterized in that the device comprises:

[0045] A data set acquisition module, used to acquire a target EEG signal data set corresponding to a target object; the target EEG signal data set includes a plurality of EEG signal samples of the target object, and each EEG signal sample includes a plurality of channels;

[0046] An information gain calculation module, used to calculate the target information gain of each channel in the target EEG signal data set;

[0047] The target channel combination determination module is used to determine, based on the target information gain of each channel, a channel combination consisting of multiple channels with the highest target information gain ranking as a target channel combination in a preset manner.

[0048] The present invention also provides an electronic device, comprising a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the EEG signal channel processing method as described above.

[0049] The present invention also provides a computer-readable storage medium, in which at least one instruction or at least one program is stored. The at least one instruction or the at least one program is loaded and executed by a processor to implement the EEG signal channel processing method as described above.

[0050] The implementation of the embodiments of the present invention has the following beneficial effects:

[0051] The EEG signal channel processing method disclosed in the present invention determines the target channel combination based on the information gain of each EEG signal channel, so that effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in analyzing brain activities, and having wide applicability and portability. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the EEG signal channel processing method, device, electronic device and storage medium described in the present invention, the drawings required for the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0053] Figure 1 A schematic diagram of a flow chart of a method for processing an EEG signal channel provided by an embodiment of the present invention;

[0054] Figure 2 A schematic flow chart of a method for determining target information gain of each channel provided by an embodiment of the present invention;

[0055] Figure 3 A schematic flow chart of a method for determining data division points of each channel provided by an embodiment of the present invention;

[0056] Figure 4 A schematic flow chart of a method for calculating information gain of each channel provided by an embodiment of the present invention;

[0057] Figure 5 A schematic flow chart of a method for determining a target channel combination provided by an embodiment of the present invention;

[0058] Figure 6 A schematic flow chart of a method for determining index information of each channel combination provided by an embodiment of the present invention;

[0059] Figure 7 A schematic diagram of target information gain of multiple channels provided by an embodiment of the present invention;

[0060] Figure 8 A schematic diagram of the accuracy of a machine learning model corresponding to a plurality of channel combinations in a training set and a test set provided by an embodiment of the present invention;

[0061] Fig. 9 A schematic diagram of the calculation time of a target accuracy rate corresponding to a combination of multiple channels provided in an embodiment of the present invention;

[0062] Fig.10 A schematic diagram of indicator information corresponding to a combination of multiple channels provided by an embodiment of the present invention;

[0063] Fig.11 A schematic diagram of a flow chart of a method for updating a target EEG signal data set provided by an embodiment of the present invention;

[0064] Fig.12 A schematic diagram of the structure of an electroencephalogram signal channel processing device provided by an embodiment of the present invention;

[0065] Fig.13 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0067] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0068] Please refer to Figure 1 , which shows a flow chart of an EEG signal channel processing method provided by an embodiment of the present invention. This specification provides method operation steps as described in the embodiment or flow chart, but may include more or fewer operation steps based on conventional or non-creative labor. The order of steps listed in the embodiment is only one way of executing the steps among many, and does not represent the only execution order. When the actual system or server product is executed, it can be executed in the order of the method shown in the embodiment or the accompanying drawings. Specifically, Figure 1 As shown, the method includes:

[0069] S101, obtaining a target EEG signal dataset corresponding to a target object; the target EEG signal dataset includes a plurality of EEG signal samples of the target object, and each EEG signal sample includes a plurality of channels;

[0070] In the embodiments of this specification, the target object is the subject, that is, the individual whose EEG signal is recorded;

[0071] In the embodiment of this specification, each of the EEG signal samples includes EEG signal values ​​transmitted through multiple channels;

[0072] In the embodiment of the present specification, the EEG signal value included in each EEG signal sample may be the EEG signal value transmitted by the target object in each channel at the time of triggering the target task or the EEG signal value transmitted by the target object in each channel at the time related to the target task; the target task may include one or more behaviors of the target object;

[0073] In the embodiments of this specification, the target EEG signal dataset may be characterized by each channel and labeled by the target task;

[0074] For example, as shown in Table 1, Table 1 is an EEG signal dataset of monkey movement. Table 1 shows an EEG signal dataset consisting of EEG signal values ​​transmitted through 6 brain channels when the monkey moves to the left and to the right in a monkey movement experiment. The dataset includes 7 EEG signal samples. The values ​​in Table 1 represent the signals transmitted by each channel, and the label column represents the actual movement of the monkey.

[0075] Table 1 Monkey movement EEG signal dataset

[0076] <![CDATA[ Channel 1 ]]> <![CDATA[ Channel 2 ]]> <![CDATA[ Channel 3 ]]> <![CDATA[ Channel 4 ]]> <![CDATA[ Channel 5 ]]> <![CDATA[ Channel 6 ]]> <![CDATA[ Label ]]> 91 89 3 88 6 70 Move left 36 98 8 25 96 5 Move right 60 34 41 26 28 63 Move right 34 86 89 38 11 96 Move left 21 57 99 71 69 1 Move right 9 49 3 59 40 80 Move left 75 52 33 50 6 49 Move left

[0077] S103, calculating the target information gain of each channel in the target EEG signal data set;

[0078] Please refer to Figure 2 , which is a flow chart of a method for determining target information gain of each channel provided by an embodiment of the present invention. In the embodiment of this specification, the target information gain of each channel in the target EEG signal data set is calculated, including:

[0079] S201, determining the data division points corresponding to each channel respectively; each channel in the target EEG signal data set has a plurality of data division points;

[0080] In the embodiments of the present specification, the data division points corresponding to each channel can be determined based on the magnitude relationship of the electroencephalogram signal values ​​corresponding to each channel.

[0081] S203, based on the data division points, respectively calculating the information gain of each channel; each channel in the target EEG signal data set has a plurality of information gains, and the plurality of information gains of each channel correspond one-to-one to the plurality of data division points of each channel;

[0082] In the embodiment of the present specification, multiple partition intervals can be obtained based on the data partition points corresponding to each channel, and the information gain of each channel can be further obtained based on the information entropy of the target EEG signal data set and the information entropy corresponding to the multiple partition intervals;

[0083] In the embodiments of the present specification, for a channel, an information gain of the channel can be calculated based on each data division point corresponding to the channel. Since each channel has multiple data division points, each channel has multiple information gains.

[0084] S205: Determine the maximum information gain among the multiple information gains of the channels as the target information gain of the channels.

[0085] In the embodiment of the present specification, the target information gain of each channel is the maximum value of multiple information gains corresponding one-to-one to multiple data division points of each channel.

[0086] According to the above-mentioned technical means, the embodiment of this specification obtains multiple information gains based on the data division points corresponding to each channel, and determines the maximum value of the multiple information gains as the target information gain of each channel, and then further determines the target channel combination based on the target information gain of each channel, so that effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0087] Please refer to Figure 3 , which is a flow chart of a method for determining data partition points of each channel provided by an embodiment of the present invention. In the embodiment of this specification, determining the data partition points corresponding to each channel includes:

[0088] S301, sorting the EEG signal values ​​included in each channel respectively to obtain a plurality of EEG signal value sequences;

[0089] In the embodiments of this specification, the EEG signal values ​​included in each channel can be sorted in order from large to small or from small to large to obtain multiple EEG signal value sequences and labels corresponding to each EEG signal value;

[0090] For example, taking channel 1 in the EEG signal data set of monkey movement shown in Table 1 as an example, the EEG signal values ​​corresponding to channel 1 are sorted in order from large to small, and the EEG signal value sequence of channel 1 and the labels corresponding to each EEG signal value are obtained as shown in Table 2.

[0091] Table 2 EEG signal value sequence of channel 1 and the labels corresponding to each EEG signal value

[0092] <![CDATA[ Channel 1 ]]> <![CDATA[ Label ]]> 91 Move left 75 Move left 60 Move right 36 Move right 34 Move left 21 Move right 9 Move left

[0093] S303, calculating the average value of every two adjacent EEG signal values ​​in each EEG signal value sequence;

[0094] For example, taking the EEG signal value sequence of channel 1 shown in Table 2 as an example, the average values ​​of every two adjacent EEG signal values ​​in the EEG signal value sequence are calculated, which are: (91+75) / 2=83, (75+60) / 2=67.5, (60+36) / 2=48, (36+34) / 2=35, (34+21) / 2=27.5, (21+9) / 2=15.

[0095] S305: Determine the average value as the data division point corresponding to each channel.

[0096] For example, taking the EEG signal value sequence of channel 1 shown in Table 2 as an example, the average values ​​corresponding to the EEG signal value sequence are 83, 67.5, 48, 35, 27.5, and 15, that is, the data division points corresponding to channel 1 include 83, 67.5, 48, 35, 27.5, and 15. In the same way, the data division points corresponding to channels 2 to 6 in the EEG signal data set of the monkey movement shown in Table 1 can be obtained.

[0097] According to the above-mentioned technical means, the embodiment of this specification determines the data division points corresponding to each channel by sorting the EEG signal values ​​corresponding to each channel and averaging them, and further determines the target information gain of each channel based on the data division points corresponding to each channel, and determines the target channel combination based on the target information gain of each channel, so that effective channels can be quickly and accurately identified and selected in the EEG signals, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0098] Please refer to Figure 4 , which is a flow chart of a method for calculating the information gain of each channel provided by an embodiment of the present invention. In the embodiment of this specification, the information gain of each channel is calculated based on the data division point, including:

[0099] S401, calculating the first information entropy of the target EEG signal data set;

[0100] In the embodiment of the present specification, the first information entropy of the target EEG signal data set can be calculated according to the information entropy formula; specifically, the calculation formula of the first information entropy is:

[0101]

[0102] Where S represents the target EEG signal dataset, T represents the label, and p i represents the probability of label i appearing in the target EEG signal dataset S;

[0103] For example, taking the EEG signal dataset of monkey movement shown in Table 1 as an example, the probability of the label "left movement" in the EEG signal dataset is 4 / 7, and the probability of the label "right movement" in the EEG signal dataset is 3 / 7. Therefore, the first information entropy of the EEG signal dataset can be calculated as:

[0104]

[0105] S403, dividing the EEG signal values ​​included in each channel based on the data division points to obtain a plurality of division intervals;

[0106] In the embodiments of this specification, for a channel, based on multiple data division points corresponding to the channel, the EEG signal value corresponding to the channel can be divided into multiple division intervals; specifically, for a channel, based on each data division point corresponding to the channel, the EEG signal value corresponding to the channel can be divided into two division intervals, wherein the EEG signal values ​​in one division interval are all greater than the data division point, and the EEG signal values ​​in the other division interval are all less than the data division point;

[0107] For example, taking channel 1 in the EEG signal data set of monkey movement shown in Table 1 as an example, the data division points corresponding to channel 1 include 83, 67.5, 48, 35, 27.5, and 15. When 15 is used as the data division point, the EEG signal value corresponding to channel 1 can be divided into two division intervals, as shown in Table 3, which is the division interval obtained by channel 1 based on the data division point 15, where 91, 75, 60, 36, 34, and 21 are division interval 1, and the EEG signal values ​​in division interval 1 are all greater than the data division point 15, and 9 is division interval 2, and the EEG signal values ​​in division interval 2 are all less than the data division point 15; when 27.5 is used as the data division point, the EEG signal value corresponding to channel 1 can be divided into two division intervals, as shown in Table 4, which is the division interval obtained by channel 1 based on the data division point 27.5, where 91, 75, 60, 36, 34, and 21 are division intervals 1, and the EEG signal values ​​in division interval 1 are all greater than the data division point 15, and 9 is division interval 2, and the EEG signal values ​​in division interval 2 are all less than the data division point 15; 5, 60, 36, 34 are partition interval 1, and the EEG signal values ​​in partition interval 1 are all greater than the data partition point 27.5, 21, 9 are partition interval 2, and the EEG signal values ​​in partition interval 2 are all less than the data partition point 27.5; when 35 is the data partition point, the EEG signal value corresponding to channel 1 can be divided into two partition intervals, as shown in Table 5, which is the partition interval obtained by channel 1 based on the data partition point 35, among which 91, 75, 60, 36 are partition interval 1, and the EEG signal values ​​in partition interval 1 are all greater than the data partition point 35, 34, 21, 9 are partition interval 2, and the EEG signal values ​​in partition interval 2 are all less than the data partition point 35; When 48 is used as the data division point, the EEG signal value corresponding to channel 1 can be divided into two division intervals, as shown in Table 6. Table 6 is the division intervals obtained for channel 1 based on the data division point 48, where 91, 75, and 60 are division interval 1, and the EEG signal values ​​in division interval 1 are all greater than the data division point 48, and 36, 34, 21, and 9 are division interval 2, and the EEG signal values ​​in division interval 2 are all less than the data division point 48; when 67.5 is used as the data division point, the EEG signal value corresponding to channel 1 can be divided into two division intervals, as shown in Table 7. Table 7 is the division intervals obtained for channel 1 based on the data division point 67.5, where 91, 75, and 60 are division intervals 1, and the EEG signal values ​​in division interval 1 are all greater than the data division point 48, and 36, 34, 21, and 9 are division intervals 2, and the EEG signal values ​​in division interval 2 are all less than the data division point 48. 5 is partition interval 1, and the EEG signal values ​​in partition interval 1 are all greater than the data partition point 67.5; 60, 36, 34, 21, and 9 are partition interval 2, and the EEG signal values ​​in partition interval 2 are all less than the data partition point 67.5; when 83 is the data partition point, the EEG signal values ​​corresponding to channel 1 can be divided into two partition intervals, as shown in Table 8, which is the partition interval obtained for channel 1 based on the data partition point 83, among which 91 is partition interval 1, and the EEG signal values ​​in partition interval 1 are all greater than the data partition point 83; 75, 60, 36, 34, 21, and 9 are partition interval 2, and the EEG signal values ​​in partition interval 2 are all less than the data partition point 83.

[0108] Table 3. The partition intervals of channel 1 based on data partition point 15

[0109]

[0110] Table 4 The division intervals of channel 1 based on the data division point 27.5

[0111]

[0112] Table 5 The division intervals of channel 1 based on data division point 35

[0113]

[0114] Table 6 The division intervals obtained by channel 1 based on data division point 48

[0115]

[0116] Table 7 The division interval of channel 1 based on the data division point 67.5

[0117]

[0118] Table 8 The division intervals obtained by channel 1 based on data division point 83

[0119]

[0120] S405, calculating information entropies corresponding to the plurality of divided intervals respectively based on the EEG signal samples corresponding to the plurality of divided intervals respectively;

[0121] In the embodiment of this specification, according to the information entropy formula, the information entropy corresponding to the multiple partition intervals obtained by each channel based on different data partition points can be calculated;

[0122] For example, taking the partition interval obtained by channel 1 based on data partition point 15 shown in Table 3 as an example, for partition interval 1, it includes 6 EEG signal samples, of which there are 3 samples with the label "left movement" and 3 samples with the label "right movement", so the information entropy corresponding to partition interval 1 can be calculated as:

[0123]

[0124] For partition interval 2, it includes 1 EEG signal sample, of which there is 1 label "left movement" and 0 label "right movement", so the information entropy corresponding to partition interval 2 can be calculated as:

[0125]

[0126] S407, performing data fusion processing on the first information entropy and the information entropies corresponding to the multiple divided intervals respectively, to obtain the information gain of each channel.

[0127] In the embodiment of this specification, according to the information gain formula, the first information entropy of the target EEG signal data set and the information entropy corresponding to the multiple divided intervals of each channel are substituted to calculate the information gain of each channel; specifically, the calculation formula of the information gain of each channel is:

[0128]

[0129] Where Ent(S) represents the first information entropy of the target EEG signal dataset S, V represents that channel c is divided into V intervals on the target EEG signal dataset S, and p i represents the probability that the ith interval of channel c appears in the target EEG signal dataset S, Ent(c i ) represents the information entropy corresponding to the i-th interval of channel c;

[0130] For example, taking the partition interval obtained by channel 1 based on data partition point 15 shown in Table 3 as an example, in Table 3, the probability of partition interval 1 appearing in the target EEG signal data set is 6 / 7, and the information entropy corresponding to partition interval 1 is 1. The probability of partition interval 2 appearing in the target EEG signal data set is 1 / 7, and the information entropy corresponding to partition interval 2 is 0. The first information entropy of the target EEG signal data set is 0.9852. Therefore, the information gain of channel 1 based on data partition point 15 is:

[0131]

[0132] Taking the partition interval obtained by channel 1 based on the data partition point 27.5 shown in Table 4 as an example, in Table 4, the probability of partition interval 1 appearing in the target EEG signal data set is 5 / 7, and the information entropy corresponding to partition interval 1 is 0.97095. The probability of partition interval 2 appearing in the target EEG signal data set is 2 / 7, and the information entropy corresponding to partition interval 2 is 1. The first information entropy of the target EEG signal data set is 0.9852. Therefore, the information gain of channel 1 based on the data partition point 27.5 is:

[0133]

[0134] Taking the partition interval obtained by channel 1 based on data partition point 35 shown in Table 5 as an example, in Table 5, the probability of partition interval 1 appearing in the target EEG signal data set is 4 / 7, and the information entropy corresponding to partition interval 1 is 1. The probability of partition interval 2 appearing in the target EEG signal data set is 3 / 7, and the information entropy corresponding to partition interval 2 is 0.9183. The first information entropy of the target EEG signal data set is 0.9852. Therefore, the information gain of channel 1 based on data partition point 35 is:

[0135]

[0136] Taking the partition intervals obtained by channel 1 based on data partition point 48 shown in Table 6 as an example, in Table 6, the probability of partition interval 1 appearing in the target EEG signal data set is 3 / 7, and the information entropy corresponding to partition interval 1 is 0.918296. The probability of partition interval 2 appearing in the target EEG signal data set is 4 / 7, and the information entropy corresponding to partition interval 2 is 1. The first information entropy of the target EEG signal data set is 0.9852. Therefore, the information gain of channel 1 based on data partition point 48 is:

[0137]

[0138] Taking the partition interval obtained by channel 1 based on the data partition point 67.5 shown in Table 7 as an example, in Table 7, the probability of partition interval 1 appearing in the target EEG signal data set is 2 / 7, and the information entropy corresponding to partition interval 1 is 0. The probability of partition interval 2 appearing in the target EEG signal data set is 5 / 7, and the information entropy corresponding to partition interval 2 is 0.971. The first information entropy of the target EEG signal data set is 0.9852. Therefore, the information gain of channel 1 based on the data partition point 67.5 is:

[0139]

[0140] Taking the partition interval obtained by channel 1 based on data partition point 83 shown in Table 8 as an example, in Table 8, the probability of partition interval 1 appearing in the target EEG signal data set is 1 / 7, and the information entropy corresponding to partition interval 1 is 0. The probability of partition interval 2 appearing in the target EEG signal data set is 6 / 7, and the information entropy corresponding to partition interval 2 is 1. The first information entropy of the target EEG signal data set is 0.9852. Therefore, the information gain of channel 1 based on data partition point 83 is:

[0141]

[0142] In summary, the six information gains corresponding to the six data division points of channel 1 are 0.12806, 0.00597, 0.0202143, 0.02021, 0.2916, and 0.1281, respectively. Since the target information gain of each channel is the maximum information gain among the multiple information gains of each channel, the target information gain of channel 1 is 0.2916. In the same way, the target information gain of channels 2 to 6 in the EEG signal data set of monkey movement shown in Table 1 can be obtained.

[0143] According to the above-mentioned technical means, the embodiment of this specification obtains multiple information gains of each channel based on the information entropy of the target EEG signal data set and the information entropy corresponding to the multiple divided intervals of each channel, and further determines the target information gain of each channel based on the multiple information gains of each channel, and determines the target channel combination based on the target information gain of each channel, so that effective channels can be quickly and accurately identified and selected in the EEG signals, and channels that may bring noise or invalid information can be filtered out, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0144] S105 , based on the target information gain of each channel, determine a channel combination consisting of a plurality of channels with higher rankings of the target information gain as a target channel combination in a preset manner.

[0145] In the embodiment of the present specification, the preset method may be: a plurality of channels ranked high in target information gain form a plurality of channel combinations, and based on the indicator information respectively corresponding to the plurality of channel combinations, a target channel combination is determined from the plurality of channel combinations.

[0146] Please refer to Figure 5 , which is a flow chart of a method for determining a target channel combination provided by an embodiment of the present invention. In the embodiment of the present specification, based on the target information gain of each channel, a channel combination consisting of multiple channels ranked at the top of the target information gain is determined as the target channel combination in a preset manner, including:

[0147] S501, sorting the channels corresponding to the target information gain in descending order of the target information gain to obtain an EEG signal channel sequence;

[0148] In the embodiment of the present specification, the target information gain of each channel included in the EEG signal channel sequence is decreasing; the larger the target information gain of a channel is, the higher the weight of the channel is.

[0149] S503, determining a plurality of channel combinations based on the EEG signal channel sequence; the number of channels respectively included in the plurality of channel combinations is increasing, the plurality of channels included in each channel combination are continuous in the EEG signal channel sequence, and each channel combination includes the first channel in the EEG signal channel sequence;

[0150] For example, assuming that the EEG signal channel sequence includes N channels (N is a positive integer greater than 2), the target information gain of the N channels is decreasing, and the number of channels included in the multiple channel combinations is 2, 3, ..., N, that is, the top 2 channels in the EEG signal channel sequence are combined, the top 3 channels in the EEG signal channel sequence are combined..., and so on, until the top N channels in the EEG signal channel sequence (i.e. all channels) are combined, and N-1 channel combinations can be obtained.

[0151] S505, determining the index information corresponding to each of the plurality of channel combinations; the index information represents the prediction performance of each channel combination in the target prediction task;

[0152] In the embodiments of the present specification, the prediction performance may include the prediction accuracy, prediction time and / or other prediction-related indicators of each channel combination in the target prediction task.

[0153] S507: Determine the target channel combination from the multiple channel combinations based on the indicator information.

[0154] In the embodiment of the present specification, the channel combination with the largest indicator information value may be determined as the target channel combination.

[0155] According to the above-mentioned technical means, the embodiment of this specification determines multiple channel combinations based on the information gain of each channel of the EEG signal, and then determines the channel combination composed of multiple channels with higher information gain ranking as the target channel combination according to the indicator information corresponding to the multiple channel combinations. In this way, effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0156] Please refer to Figure 6 , which is a flow chart of a method for determining index information of each channel combination provided by an embodiment of the present invention. In the embodiment of this specification, determining the index information corresponding to the plurality of channel combinations respectively includes:

[0157] S601, determining a training set and a test set corresponding to each channel combination based on the EEG signal samples corresponding to each channel combination;

[0158] In the embodiments of the present specification, a cross-validation method can be used to divide the EEG signal samples corresponding to each channel combination to determine the training set and the test set corresponding to each channel combination; for example, in order to ensure the generalization of the machine learning model, a K-fold cross-validation method can be used to determine the training set and the test set corresponding to each channel combination. Specifically, the EEG signal samples corresponding to each channel combination are divided into K sub-samples, one of which is used as the test set, and the other K-1 samples are used as the training set. A different sub-sample is used as the test set each time, and the cross-validation is repeated K times.

[0159] S603, establishing a machine learning model corresponding to each channel combination one by one based on the training sets corresponding to each channel combination;

[0160] In the embodiment of the present specification, considering the time complexity, the machine learning model may be a linear model; for example, the machine learning model may be a logistic regression model.

[0161] S605, calculating a first accuracy rate of the machine learning model in a training set respectively corresponding to the machine learning model;

[0162] In the embodiment of the present specification, the first accuracy rate represents the proportion of the number of correct predictions made by the machine learning model to the total number in the training sets corresponding to the machine learning models respectively.

[0163] S607, calculating a second accuracy rate of the machine learning model in a test set respectively corresponding to the machine learning model;

[0164] In the embodiment of the present specification, the second accuracy rate represents the proportion of the number of correct predictions made by the machine learning model to the total number in the test sets corresponding to the machine learning models respectively.

[0165] S609, performing data fusion processing on the first accuracy rate and the second accuracy rate to obtain a target accuracy rate of the machine learning model;

[0166] In the embodiment of this specification, the average accuracy rate obtained by K-fold cross validation is used as the target accuracy rate; specifically, the target accuracy rate of the machine learning model is:

[0167]

[0168] Among them, C 1～n represents each channel combination, n represents the nth channel in the EEG signal channel sequence; Model(C 1～n ) represents the machine learning model based on the combination of each channel; T represents the label; K represents the K value in K-fold cross validation; A CC Indicates the accuracy, Represents the first accuracy of the machine learning model in the training set, Represents the second accuracy of the machine learning model in the test set; α is a parameter ranging from 0 to 1, which is used to configure the weight ratio of the training set (train) and the test set (test), and the default is 0.7.

[0169] S611, determining the indicator information corresponding to each channel combination based on the target accuracy and the calculation time of the target accuracy.

[0170] In the embodiment of this specification, the calculation time of the target accuracy rate includes the calculation time of the first accuracy rate, the second accuracy rate, and the time of performing data fusion processing on the first accuracy rate and the second accuracy rate;

[0171] Figure 6 The method for determining the indicator information of each channel combination shown is to divide the EEG signal samples corresponding to each channel combination into K sub-samples, use one of the sub-samples as a test set, and the other K-1 samples as a training set, use the training set to train the machine learning model, and calculate the accuracy of the machine learning model on the training set and the test set, respectively, use different sub-samples as the test set each time, repeat cross-validation K times, and take the average accuracy of K cross-validations as the target accuracy, and determine the indicator information corresponding to each channel combination based on the target accuracy and the calculation time of the target accuracy.

[0172] According to the above-mentioned technical means, the embodiment of this specification determines the indicator information corresponding to each channel combination based on the machine learning model established for each channel combination, and further determines the target channel combination from multiple channel combinations based on the indicator information, so that effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0173] In the embodiment of this specification, the determining the indicator information corresponding to the plurality of channel combinations respectively based on the target accuracy and the calculation time of the target accuracy includes:

[0174] The difference between the target accuracy and the calculation time of the target accuracy is determined as the indicator information respectively corresponding to the multiple channel combinations.

[0175] In the embodiment of this specification, the calculation formula of the indicator information can be:

[0176] f(C 1～n ,duration)=L(Model(C 1～n ),T)-duration

[0177] Among them, L(Model(C 1～n ), T) represents the target accuracy, duration represents the calculation time of the target accuracy; C 1～n represents each channel combination, n represents the nth channel in the EEG signal channel sequence; Model(C 1～n ) represents a machine learning model based on the combination of each channel; T represents a label;

[0178] In another embodiment of the present specification, a weight factor α can be added between the target accuracy and the calculation time of the target accuracy to balance the conversion relationship between the target accuracy and the calculation time of the target accuracy, so as to better ensure that the accuracy is improved while controlling the delay and improve the overall performance of the brain-computer interface system. Accordingly, the calculation formula can be:

[0179] f(C 1～n ,duration)=L(Model(C 1～n ),T)-α*duration

[0180] Here, α is a constant determined according to demand or user experience.

[0181] According to the above-mentioned technical means, the embodiment of this specification determines the indicator information corresponding to multiple channel combinations based on the target accuracy and the calculation time of the target accuracy, and then determines the target channel combination from the multiple channel combinations according to the indicator information, so that effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, thereby reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and having wide applicability and portability.

[0182] Please refer to Figure 7 , which is a schematic diagram of target information gain of multiple channels provided by an embodiment of the present invention, Figure 7There are 250 channels in total, the x-axis represents the channel, and the y-axis represents the target information gain corresponding to the channel. Among them, channels 73-79, 199, 217, and 247 provide high information gains;

[0183] Please refer to Figure 8 , which is a schematic diagram showing the accuracy of a machine learning model corresponding to a plurality of channel combinations provided by an embodiment of the present invention in a training set and a test set, Figure 8 There are 250 channels in total. The x-axis represents the top 5 channels (top5) to the top 250 channels (top250) in terms of target information gain, and the y-axis represents the accuracy. By observing the accuracy of the test set, it can be found that the accuracy obtained by using the top 30 channels is equivalent to the accuracy obtained by using all channels, which also shows that there is noise in the collected EEG signal data.

[0184] Please refer to Fig. 9 , which is a schematic diagram of the calculation time of the target accuracy corresponding to a combination of multiple channels provided by an embodiment of the present invention, Fig. 9 There are 250 channels in total. The x-axis represents the top 5 channels (top5) to the top 250 channels (top250) in terms of target information gain, and the y-axis represents the calculation time (in seconds).

[0185] Please refer to Fig.10 , which is a schematic diagram of indicator information corresponding to a plurality of channel combinations provided by an embodiment of the present invention, Fig.10 There are 250 channels in total. The x-axis represents the top 5 channels (top5) to the top 250 channels (top250) in terms of target information gain, and the y-axis represents the value of the indicator information. It can be found that the indicator information value corresponding to the top76 channel combination is the largest, that is, the target channel combination is the channel combination composed of the top 76 channels in terms of target information gain.

[0186] Please refer to Fig.11 , which is a flow chart of a method for updating a target EEG signal data set provided by an embodiment of the present invention. In the embodiment of this specification, the method further includes:

[0187] S1101, when the values ​​of the electroencephalogram signals transmitted in the multiple channels change, obtaining updated electroencephalogram signal samples based on the channel identifiers;

[0188] In the embodiment of this specification, the channel identifier may be a channel number or other identifier used to distinguish each channel.

[0189] S1103, updating the target EEG signal dataset based on the updated EEG signal samples to obtain an updated EEG signal dataset;

[0190] In the embodiment of the present specification, samples in the target EEG signal dataset are supplemented or replaced based on the updated EEG signal samples to obtain an updated EEG signal dataset.

[0191] S1105: Determine the updated EEG signal dataset as the target EEG signal dataset.

[0192] In the embodiment of this specification, after the updated EEG signal dataset is determined as the target EEG signal dataset, Figure 1 The EEG signal channel processing method shown recalculates the target information gain of each channel and redetermines the target channel combination.

[0193] According to the above technical means, whenever the channel situation changes, the target EEG signal data set can be updated, the information gain of each channel can be recalculated, and the target channel combination can be re-determined, thereby improving the real-time performance of the brain-computer interface system and the accuracy of analyzing brain activity.

[0194] In the embodiment of this specification, the method further includes:

[0195] After the target channel combination is determined, the EEG signal of the target object is transmitted through each channel included in the target channel combination.

[0196] In the embodiment of the present specification, after determining the target channel combination, the front end of the brain-computer interface system may only use the channels included in the target channel combination to transmit the EEG signal of the target object.

[0197] According to the above-mentioned technical means, after determining the target channel combination, the front end of the brain-computer interface system can only use the channels included in the target channel combination to transmit EEG signals, thereby reducing the transmission amount of EEG signals and improving the transmission efficiency of EEG signals.

[0198] In another embodiment of this specification, the method further includes:

[0199] After determining the target channel combination, the target system decodes the EEG signal of the target object based on each channel included in the target channel combination.

[0200] In the embodiments of this specification, the target system may be a brain-computer interface system;

[0201] In an embodiment of the present specification, after determining the target channel combination, the back end of the brain-computer interface system can decode the EEG signal of the target object only based on the signals in each channel included in the target channel combination after acquiring the EEG signal of the target object; the back end of the brain-computer interface system can be a decoding end.

[0202] According to the above-mentioned technical means, after determining the target channel combination, the back-end of the brain-computer interface system can decode the EEG signals only based on the signals in the channels included in the target channel combination, thereby reducing the resources consumed by decoding and improving decoding efficiency.

[0203] It can be seen from the above-mentioned embodiments of the EEG signal channel processing method provided by the present invention that the embodiments of the present invention obtain a target EEG signal data set corresponding to a target object; the target EEG signal data set includes multiple EEG signal samples of the target object, and each EEG signal sample includes multiple channels; the target information gain of each channel in the target EEG signal data set is calculated; based on the target information gain of each channel, a channel combination consisting of multiple channels ranked top in the target information gain is determined as a target channel combination in a preset manner. The technical solution provided in the embodiments of this specification determines the target channel combination based on the information gain of each EEG signal channel, so that effective channels can be quickly and accurately identified and selected in the EEG signal, while filtering out channels that may bring noise or invalid information, reducing the computational burden of the brain-computer interface system, improving the real-time and stability of the brain-computer interface system, and improving the efficiency and accuracy of the brain-computer interface system in parsing brain activities, and has wide applicability and portability.

[0204] The embodiment of the present invention also provides a device for processing an electroencephalogram signal channel. Fig.12 As shown, it is a structural schematic diagram of an electroencephalogram signal channel processing device provided by an embodiment of the present invention; specifically, the device includes:

[0205] The data set acquisition module 1210 is used to acquire a target EEG signal data set corresponding to a target object; the target EEG signal data set includes a plurality of EEG signal samples of the target object, and each EEG signal sample includes a plurality of channels;

[0206] An information gain calculation module 1220 is used to calculate the target information gain of each channel in the target EEG signal data set;

[0207] The target channel combination determination module 1230 is used to determine, based on the target information gains of the channels, a channel combination consisting of a plurality of channels with higher rankings of the target information gains as a target channel combination in a preset manner.

[0208] In the embodiment of this specification, the target channel combination determination module 1230 includes:

[0209] A sorting unit, used to sort the channels corresponding to the target information gain in descending order of the target information gain, to obtain an EEG signal channel sequence;

[0210] A channel combination unit, used for determining a plurality of channel combinations based on the EEG signal channel sequence; the number of channels respectively included in the plurality of channel combinations is increasing, the plurality of channels included in each channel combination are continuous in the EEG signal channel sequence, and each channel combination includes the first channel in the EEG signal channel sequence;

[0211] An indicator information determination unit, used to determine the indicator information corresponding to each of the plurality of channel combinations; the indicator information represents the prediction performance of each channel combination in the target prediction task;

[0212] The target channel combination determining unit is used to determine a target channel combination from the multiple channel combinations based on the indicator information.

[0213] In the embodiment of this specification, the information gain calculation module 1220 includes:

[0214] A data division point determination unit, used to determine the data division points corresponding to each of the channels; each channel in the target EEG signal data set has a plurality of data division points;

[0215] An information gain calculation unit, used for calculating the information gain of each channel based on the data division point; each channel in the target EEG signal data set has multiple information gains, and the multiple information gains of each channel correspond to the multiple data division points of each channel one by one;

[0216] The target information gain determination unit is used to determine the maximum information gain among the multiple information gains of each channel as the target information gain of each channel.

[0217] In the embodiment of this specification, the data division point determination unit includes:

[0218] A sorting subunit, used for sorting the EEG signal values ​​included in each channel to obtain a plurality of EEG signal value sequences;

[0219] an average value calculation subunit, used for calculating the average value of every two adjacent EEG signal values ​​in each EEG signal value sequence;

[0220] The data division point determination subunit is used to determine the average value as the data division point corresponding to each channel.

[0221] In this embodiment of the specification, the information gain calculation unit includes:

[0222] A first calculation subunit, used to calculate a first information entropy of the target EEG signal data set;

[0223] An interval division subunit, used for dividing the EEG signal values ​​included in each channel based on the data division points to obtain a plurality of division intervals;

[0224] A second calculation subunit is used to calculate the information entropy corresponding to the plurality of divided intervals respectively based on the EEG signal samples corresponding to the plurality of divided intervals respectively;

[0225] The third calculation subunit is used to perform data fusion processing on the first information entropy and the information entropies corresponding to the multiple divided intervals to obtain the information gain of each channel.

[0226] In the embodiment of this specification, the indicator information determination unit includes:

[0227] The data set division subunit is used to determine the training set and the test set corresponding to each channel combination based on the EEG signal samples corresponding to each channel combination;

[0228] A machine learning model establishment subunit, used to establish a machine learning model corresponding to each channel combination based on the training set corresponding to each channel combination;

[0229] A first accuracy calculation subunit, used to calculate a first accuracy of the machine learning model in a training set respectively corresponding to the machine learning model;

[0230] A second accuracy calculation subunit, used to calculate a second accuracy of the machine learning model in a test set respectively corresponding to the machine learning model;

[0231] a target accuracy calculation subunit, configured to perform data fusion processing on the first accuracy and the second accuracy to obtain a target accuracy of the machine learning model;

[0232] The indicator information determination subunit is used to determine the indicator information corresponding to each channel combination based on the target accuracy and the calculation time of the target accuracy.

[0233] In the embodiment of this specification, the device also includes:

[0234] A sample updating module, configured to obtain updated EEG signal samples based on channel identifiers when the EEG signal values ​​transmitted in the multiple channels change;

[0235] A data set updating module, used for updating the target EEG signal data set based on the updated EEG signal samples to obtain an updated EEG signal data set;

[0236] The data set determination module is used to determine the updated EEG signal data set as the target EEG signal data set.

[0237] An embodiment of the present invention also provides an electronic device, comprising a processor and a memory; the memory stores one or more instructions, and the one or more instructions are suitable for the processor to load and execute to implement the EEG signal channel processing method as described in the above method embodiment.

[0238] The memory can be used to store software programs and modules. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include a high-speed random access memory and can also include a non-volatile memory, such as at least one disk storage device, a flash memory device or other volatile solid-state storage device. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0239] Fig.13 1301, memory 1302, and communication interface 1303. The processor 1301, memory 1302, and communication interface 1303 in the electronic device can be connected through a bus or other means. Fig.13 The example of connecting through bus is taken in the following.

[0240] Among them, the processor 1301 (or CPU (Central Processing Unit)) is the computing core and control core of the electronic device. The communication interface 1303 is used for communication between the memory 1302 and the processor 1301. The memory 1302 (Memory) is used to store programs and data. It can be understood that the memory here can be a high-speed RAM storage device or a non-volatile storage device (non-volatile memory), such as at least one disk storage device; optionally, it can also be at least one storage device located away from the aforementioned processor. The memory provides a storage space, which stores the operating system of the electronic device, which may include but is not limited to: Windows system (an operating system), Linux system (an operating system), etc., and the present invention does not limit this; and, a computer program (including program code) suitable for being loaded and executed by the processor is also stored in the storage space. In the embodiment of this specification, the processor 1301 loads and executes the computer program stored in the memory 1302 to implement the EEG signal channel processing method provided in the above method embodiment.

[0241] An embodiment of the present invention also provides a computer-readable storage medium, which can be set in an electronic device to store at least one instruction, at least one program, code set or instruction set related to an EEG signal channel processing method in an implementation method embodiment. The at least one instruction, the at least one program, the code set or instruction set can be loaded and executed by a processor of the electronic device to implement the EEG signal channel processing method provided in the above-mentioned method embodiment.

[0242] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0243] It should be noted that the sequence of the embodiments of the present invention described above is for description only and does not represent the advantages and disadvantages of the embodiments. The above describes specific embodiments of this specification, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0244] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0245] Those skilled in the art will appreciate that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk, or an optical disk, etc.

[0246] The above disclosure is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for processing an electroencephalogram signal channel, characterized in that: include: Acquire a target EEG signal data set corresponding to the target object; The target EEG signal data set includes a plurality of EEG signal samples of the target object, each of which includes a plurality of channels; Sort the EEG signal values ​​included in each channel respectively to obtain multiple EEG signal value sequences; Calculate the average value of every two adjacent EEG signal values ​​in each EEG signal value sequence; Determine the average value as the data division point corresponding to each channel respectively; Each channel in the target EEG signal data set has a plurality of data division points; Based on the data division points, respectively calculating the information gain of each channel; Each channel in the target EEG signal data set has multiple information gains, and the multiple information gains of each channel correspond one-to-one to the multiple data division points of each channel; Determine the maximum information gain among the multiple information gains of each channel as the target information gain of each channel; According to the order of the target information gain from large to small, the channels corresponding to the target information gain are sorted to obtain an EEG signal channel sequence; Determine a plurality of channel combinations based on the EEG signal channel sequence; the number of channels respectively included in the plurality of channel combinations is increasing, the plurality of channels included in each channel combination are continuous in the EEG signal channel sequence, and each channel combination includes the first channel in the EEG signal channel sequence; one of the plurality of channel combinations includes all the channels, and the number of the plurality of channel combinations is the number of the plurality of channels minus one; Based on the EEG signal samples corresponding to each channel combination, determining a training set and a test set corresponding to each channel combination; Based on the training sets corresponding to each channel combination, a machine learning model corresponding to each channel combination is established; Calculating a first accuracy rate of the machine learning model in a training set respectively corresponding to the machine learning model; Calculating second accuracy rates of the machine learning models in test sets respectively corresponding to the machine learning models; Performing data fusion processing on the first accuracy rate and the second accuracy rate to obtain a target accuracy rate of the machine learning model; The difference between the target accuracy and the calculation time of the target accuracy is determined as the indicator information corresponding to the multiple channel combinations respectively; the indicator information represents the prediction performance of each channel combination in the target prediction task; Based on the indicator information, a target channel combination is determined from the multiple channel combinations.

2. The method for processing EEG signal channels according to claim 1, characterized in that: The calculating the information gain of each channel based on the data partition point includes: Calculating a first information entropy of the target EEG signal data set; Based on the data division points, the EEG signal values ​​included in each channel are divided to obtain a plurality of division intervals; Based on the EEG signal samples respectively corresponding to the multiple divided intervals, calculating the information entropy respectively corresponding to the multiple divided intervals; Data fusion processing is performed on the first information entropy and the information entropies corresponding to the multiple divided intervals to obtain the information gain of each channel.

3. The method for processing EEG signal channels according to claim 1, characterized in that: The method further comprises: When the values ​​of the electroencephalogram signals transmitted in the plurality of channels change, acquiring updated electroencephalogram signal samples based on the channel identifiers; The target EEG signal dataset is updated based on the updated EEG signal samples to obtain an updated EEG signal dataset; The updated EEG signal dataset is determined as the target EEG signal dataset.

4. The method for processing EEG signal channels according to claim 1, characterized in that: The method further comprises: After the target channel combination is determined, the EEG signal of the target object is transmitted through each channel included in the target channel combination.

5. The method for processing EEG signal channels according to claim 1, characterized in that: The method further comprises: After determining the target channel combination, the target system decodes the EEG signal of the target object based on each channel included in the target channel combination.

6. An electroencephalogram signal channel processing device, characterized in that: include: A data set acquisition module is used to acquire a target EEG signal data set corresponding to a target object; The target EEG signal data set includes a plurality of EEG signal samples of the target object, each of which includes a plurality of channels; An information gain calculation module is used to sort the EEG signal values ​​included in each channel to obtain multiple EEG signal value sequences; Calculating the average value of every two adjacent EEG signal values ​​in each EEG signal value sequence; determining the average value as the data division point corresponding to each channel; Each channel in the target EEG signal data set has a plurality of data division points; based on the data division points, the information gain of each channel is calculated respectively; Each channel in the target EEG signal data set has multiple information gains, and the multiple information gains of each channel correspond to the multiple data division points of each channel one by one; the maximum information gain among the multiple information gains of each channel is determined as the target information gain of each channel; A target channel combination determination module is used to sort the channels corresponding to the target information gain in descending order of the target information gain to obtain an EEG signal channel sequence; Determine multiple channel combinations based on the EEG signal channel sequence; the number of channels respectively included in the multiple channel combinations is increasing, the multiple channels included in each channel combination are continuous in the EEG signal channel sequence, and each channel combination includes the first channel in the EEG signal channel sequence; one of the multiple channel combinations includes all channels, and the number of the multiple channel combinations is the number of the multiple channels minus one; based on the EEG signal samples corresponding to each channel combination, determine the training set and the test set corresponding to each channel combination; based on the training set corresponding to each channel combination, establish a machine learning model corresponding to each channel combination one by one; Calculating a first accuracy rate of the machine learning model in a training set respectively corresponding to the machine learning model; Calculate the second accuracy of the machine learning model in the test sets respectively corresponding to the machine learning models; perform data fusion processing on the first accuracy and the second accuracy to obtain the target accuracy of the machine learning model; determine the difference between the target accuracy and the calculation time of the target accuracy as the indicator information respectively corresponding to the multiple channel combinations; the indicator information characterizes the prediction performance of each channel combination in the target prediction task; based on the indicator information, determine the target channel combination from the multiple channel combinations.

7. An electronic device, characterized in that: The device includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the EEG signal channel processing method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the EEG signal channel processing method as described in any one of claims 1-5.

Images