Adaptive training method for P300 brain-computer interface classifier

Abstract

The present invention relates to an adaptive training method for a P300 brain-computer interface classifier. Whenever a P300 brain-computer interface trial occurs, the brain-computer interface platform acquires EEG data. e t , and then determine whether this trial contains P300, and use EEG data e t and discriminative feedback y t Update the weight coefficients of the linear combination and the discriminant matrix of the current user's linear classifier. This method is conducive to dynamically updating and optimizing the P300 brain-computer interface classifier during operation.

Description

Technical Field

[0001] The present invention belongs to the field of biometric identification technology, and in particular relates to an adaptive training method for a P300 brain-computer interface classifier. Background Art

[0002] A brain-computer interface (BCI) creates a direct connection between the human or animal brain and an external device, enabling information exchange between the two. The P300 potential, an EEG component frequently used in BCIs, is an event-related potential generated during the brain's cognitive process, primarily related to psychological factors such as expectation, intention, arousal, and attention. A classifier that identifies whether an EEG signal segment contains the P300 component is a key technology in P300 BCIs. However, existing P300 BCI classifiers are largely static and cannot be dynamically updated based on feedback. Summary of the Invention

[0003] The purpose of the present invention is to provide an adaptive training method for a P300 brain-computer interface classifier, which is conducive to dynamically updating and optimizing the P300 brain-computer interface classifier during operation.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is: an adaptive training method for a P300 brain-computer interface classifier, each time a P300 brain-computer interface trial occurs, the brain-computer interface platform obtains EEG data e t Then use the formula Determine whether this trial contains P300, and use EEG data t and discriminant feedback y t Update the weight coefficients of the linear combination and the discriminant matrix w of the current user linear classifier n+1 .

[0005] Furthermore, the method comprises the following steps:

[0006] Step 1: Each P300 flicker stimulus is called a trial; use the d-dimensional vector e t represents the EEG feature vector of the tth trial; y t ∈{-1,+1} represents the feedback of the tth trial; represents the discriminant matrix of the linear classifier trained on n other users' existing data sets; b is a vector whose initial value is a d-dimensional 0 vector; Φ is a matrix whose initial value is a d-dimensional unit matrix; w n+1 Represents the discriminant matrix of the current user linear classifier, with an initial value of Φ -1 b;κ represents the learning parameter, the initial value is 0; Represents the weights of n+1 classifiers, and the initial values ​​are

[0007] Step 2: Whenever the EEG signal and feedback of the t-th trial arrive, execute steps 2.1-2.6 in sequence:

[0008] Step 2.1:

[0009] Step 2.2: Put As the classifier's output of whether the current EEG signal contains P300, +1 means it contains P300, and -1 means it does not contain P300;

[0010] Step 2.2:

[0011] Step 2.3: Update

[0012] Step 2.4:

[0013] Step 2.5: b = b + y t e t ;

[0014] Step 2.6: w n+1 =Φ -1 b.

[0015] Furthermore, in step 2.3, by calling the function Update in the form The first two parameters of the function FBWU() are passed by address when calling, and the last four parameters are passed by value when calling. The variables corresponding to the last four parameters inside the function are τ, y, e, and k respectively. The implementation steps of FBWU() are as follows:

[0016] Step 3.1:

[0017] Step 3.2: α t =θ n+1 ,

[0018] Step 3.3:

[0019] Step 3.4: r t =I(yw n+1 e≤0);

[0020] Step 3.5:

[0021] Step 3.6: α = α s +α t ,

[0022] Step 3.7:

[0023] Step 3.8: i =I(yw i e≤0), for i∈{1,…,n};

[0024] Step 3.9:

[0025] Step 3.10: for i∈{1,…,n};

[0026] Among them, I(·) maps the logical value true or false to 1 or 0.

[0027] Compared with existing technologies, the present invention has the following advantages: Based on the principles of transfer learning and reinforcement learning, the present invention proposes an adaptive training method for the P300 brain-computer interface classifier. This method allows the transfer of knowledge from other users to construct a classifier without the user's EEG data, activates the operation of the P300 brain-computer interface, and dynamically updates the classifier based on platform feedback during operation, making the classifier increasingly consistent with the characteristics of the current user. The present invention enables the P300 brain-computer interface system to start directly without calibration and continuously self-optimize during operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a diagram showing the principle of implementing the method of an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0030] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present application belongs.

[0031] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0032] This embodiment provides an adaptive training method for a P300 brain-computer interface classifier. Whenever a P300 brain-computer interface trial occurs, the brain-computer interface platform obtains EEG data e tThen use the formula Determine whether this trial contains P300, and use EEG data t and discriminant feedback y t Update the weight coefficients of the linear combination and the discriminant matrix w of the current user linear classifier n+1 The processing mechanism for a P300 brain-computer interface trial in this embodiment is as follows: Figure 1 shown.

[0033] This embodiment is based on the background that the classifier has been trained on 20 other users' EEG datasets, that is, 20 is used to replace the parameter n to further illustrate this method.

[0034] In this embodiment, the specific implementation of the method includes the following steps:

[0035] Step 1: Each P300 flicker stimulus is called a trial; use a d (d represents a constant) dimension vector e t represents the EEG feature vector of the tth trial; y t ∈{-1,+1} represents the feedback of the tth trial; represents the discriminant matrix of a linear classifier trained on n (n represents a constant) other users' existing data sets; b is a vector whose initial value is a d-dimensional 0 vector; Φ is a matrix whose initial value is a d-dimensional unit matrix; w n+1 Represents the discriminant matrix of the current user linear classifier, with an initial value of Φ -1 b;κ represents the learning parameter, the initial value is 0; Represents the weights of n+1 classifiers, and the initial values ​​are In this embodiment, n is 20, then That is w n+1 That is w 21 , That is

[0036] Step 2: Whenever the EEG signal and feedback of the tth trial (t = 1, 2, ..., T) are received, execute steps 2.1-2.6 in sequence:

[0037] Step 2.1:

[0038] Step 2.2: Put As the classifier's output of whether the current EEG signal contains P300, +1 means it contains P300, and -1 means it does not contain P300;

[0039] Step 2.2:

[0040] Step 2.3: Perform step 3 update

[0041] Step 2.4:

[0042] Step 2.5: b = b + y t e t ;

[0043] Step 2.6: w 21 =Φ -1 b.

[0044] Step 3: Call the function Update in the form The first two parameters of the function FBWU() are passed by address when calling, and the last four parameters are passed by value when calling. The variables corresponding to the last four parameters inside the function are τ, y, e, and k respectively. The implementation steps of FBWU() are as follows:

[0045] Step 3.1:

[0046] Step 3.2: α t =θ 21 ,

[0047] Step 3.3: for i∈{1,…,20};

[0048] Step 3.4: r t =I(yw 21 e≤0);

[0049] Step 3.5:

[0050] Step 3.6: α = α s +α t ,

[0051] Step 3.7:

[0052] Step 3.8: i =I(yw i e≤0), for i∈{1,…,20};

[0053] Step 3.9:

[0054] Step 3.10: for i∈{1,…,20};

[0055] Among them, I(·) maps the logical value true or false to 1 or 0.

[0056] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0057] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0058] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0059] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0060] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other manner. Any person skilled in the art may utilize the above-disclosed technical content to modify or modify the present invention into equivalent embodiments. However, any simple modifications, equivalent variations, and modifications to the above embodiments that do not depart from the technical content of the present invention and are based on the technical essence of the present invention remain within the scope of protection of the present invention.

Claims

1. An adaptive training method for a P300 brain-computer interface classifier, characterized in that: Whenever a P300 BCI trial occurs, the BCI platform acquires EEG data t Then use the formula Determine whether this trial contains P300, and use EEG data t and discriminant feedback t t Update the weight coefficients of the linear combination and the discriminant matrix w of the current user linear classifier n+1 ; The method comprises the following steps: Step 1: Each P300 flicker stimulus is called a trial; use the d-dimensional vector e t represents the EEG feature vector of the t-th trial; t t ∈{-1,+1} represents the feedback of the tth trial; represents the discriminant matrix of the linear classifier trained on n other users' existing data sets; b is a vector whose initial value is a d-dimensional 0 vector; Φ is a matrix whose initial value is a d-dimensional unit matrix; w n+1 Represents the discriminant matrix of the current user linear classifier, with an initial value of Φ -1 b;κ represents the learning parameter, the initial value is 0; Represents the weights of n+1 classifiers, and the initial values ​​are Step 2: Whenever the EEG signal and feedback of the t-th trial arrive, execute steps 2.1-2.6 in sequence: Step 2.1: Step 2.2: Put As the classifier's output of whether the current EEG signal contains P300, +1 means it contains P300, and -1 means it does not contain P300; Step 2.2: Step 2.3: Update Step 2.4: Step 2.5: b = b + y t e t ; Step 2.6: w n+1 =Φ -1 b; In step 2.3, call the function Update in the form The first two parameters of the function FBWU() are passed by address when calling, and the last four parameters are passed by value when calling. The variables corresponding to the last four parameters inside the function are τ, y, e, and k respectively. The implementation steps of FBWU() are as follows: Step 3.1: Step 3.2: Alpha t =θ n+1 , Step 3.3: Step 3.4: Step 3.5: Step 3.6: a=a s +a t , Step 3.7: Step 3.8: Step 3.9: Step 3.10: Among them, I(·) maps the logical value true or false to 1 or 0.

Images