WO2021159571A1 - 一种有向动态脑功能网络多类情绪识别构建方法及其装置 - Google Patents

一种有向动态脑功能网络多类情绪识别构建方法及其装置 Download PDF

Info

Publication number
WO2021159571A1
WO2021159571A1 PCT/CN2020/078291 CN2020078291W WO2021159571A1 WO 2021159571 A1 WO2021159571 A1 WO 2021159571A1 CN 2020078291 W CN2020078291 W CN 2020078291W WO 2021159571 A1 WO2021159571 A1 WO 2021159571A1
Authority
WO
WIPO (PCT)
Prior art keywords
directed
brain function
network
function network
client
Prior art date
Application number
PCT/CN2020/078291
Other languages
English (en)
French (fr)
Inventor
王洪涛
刘旭程
唐聪
裴子安
许林峰
Original Assignee
五邑大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 五邑大学 filed Critical 五邑大学
Publication of WO2021159571A1 publication Critical patent/WO2021159571A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/369—Electroencephalography [EEG]
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
    • A61B5/165—Evaluating the state of mind, e.g. depression, anxiety
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7253—Details of waveform analysis characterised by using transforms
    • A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device

Definitions

  • the invention relates to the technical field of brain electricity data processing, in particular to a method and device for constructing a multi-type emotion recognition of a directed dynamic brain function network.
  • EEG processing methods people have begun to pay attention to emotion recognition based on EEG, and try to decode emotion-related neural activities.
  • the essence is the processing of EEG data, which can be used in fatigue driving detection and emotion.
  • Existing methods mainly use the construction of static networks to describe brain activity. This method regards networks under different time windows as independent components.
  • the regional interaction of the brain changes with time, especially the interaction between brain regions is highly dynamic and non-stationary, so the static network cannot reflect the entire brain activity on the time scale.
  • the undirected brain function matrix can only reflect the relationship between the channels, but cannot obtain the information flow between the channels, which will also cause the lack of information when constructing the brain network.
  • the purpose of the present invention is to provide a method and device for constructing a directional dynamic brain function network for multi-emotion recognition, which can simultaneously reflect brain activity from the flow of information and time scale, so as to explore different emotions.
  • a more accurate characterization of human brain function connections provides a data basis.
  • the present invention provides a multi-type emotion recognition construction method of a directed dynamic brain function network, which includes the following steps:
  • the client obtains EEG data, preprocesses the EEG data, and decomposes it into 4 standard sub-bands through wavelet packet transform;
  • the client reads the EEG signal of the standard sub-band, divides the EEG signal according to a preset time window and step size, and calculates a directed static network through GPDC;
  • the client computer calculates the directed brain function network matrix, selects important connections according to the set sparseness, and then binarizes them to obtain the brain function matrix of the important connections.
  • connection matrices are dynamically planned according to time series, Obtain a directed dynamic brain network;
  • the client computer calculates the global efficiency and local efficiency of information transmission in the dynamic brain network, divides these parameters into three categories according to different emotions, and calculates the statistical difference between them.
  • the preprocessing includes independent component analysis and removal of EEG signal baseline.
  • the calculation of the directed brain function network matrix by the GPDC specifically includes the following steps:
  • the client obtains the segmented EEG signal, and calculates the causal relationship between the two channels of the EEG signal;
  • the client computer calculates the GPDC value based on the causal relationship between nodes.
  • directed dynamic brain network is obtained by the following specific steps:
  • the client terminal reads the preset duration, and reads the brain signal corresponding to the preset duration;
  • the client obtains the binary directed brain function network through a sparse method
  • the client computer calculates a directed dynamic brain function network according to the threshold of the binarized directed brain function network.
  • the present invention provides a device for executing a method for constructing a multi-type emotion recognition of a directed dynamic brain function network, which includes a CPU unit configured to perform the following steps:
  • the client obtains EEG data, preprocesses the EEG data, and decomposes it into 4 standard sub-bands through wavelet packet transform;
  • the client reads the EEG signal of the standard sub-band, divides the EEG signal according to a preset time window and step size, and calculates a directed static network through GPDC;
  • the client computer calculates the directed brain function network matrix, selects important connections according to the set sparseness, and then binarizes them to obtain the brain function matrix of the important connections.
  • connection matrices are dynamically planned according to time series, Obtain a directed dynamic brain network;
  • the client computer calculates the global efficiency and local efficiency of information transmission in the dynamic brain network, divides these parameters into three categories according to different emotions, and calculates the statistical difference between them.
  • CPU unit is also used to perform the following steps:
  • the client obtains the segmented EEG signal, and calculates the causal relationship between the two channels of the EEG signal;
  • the client computer calculates the GPDC value based on the causal relationship between nodes.
  • CPU unit is also used to perform the following steps:
  • the client terminal reads the preset duration, and reads the EEG signal corresponding to the preset duration;
  • the client obtains the binary directed brain function network through a sparse method
  • the client computer calculates a directed dynamic brain function network according to the threshold of the binarized directed brain function network.
  • the present invention provides a device for executing a method for constructing a multi-type emotion recognition of a directed dynamic brain function network, which includes at least one control processor and a memory for communicating with the at least one control processor; memory storage There are instructions that can be executed by at least one control processor, and the instructions are executed by at least one control processor, so that the at least one control processor can execute the method for constructing multiple types of emotion recognition in a directed dynamic brain function network as described above.
  • the present invention provides a computer-readable storage medium, the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to make the computer execute the above-mentioned directed dynamic brain function network multiple types of emotions Identify the construction method.
  • the present invention also provides a computer program product, the computer program product includes a computer program stored on a computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by a computer , Make the computer execute the above-mentioned directional dynamic brain function network multi-type emotion recognition construction method.
  • the present invention preprocesses the acquired EEG data and decomposes it into 4 standard sub-bands, and then reads from each standard sub-band.
  • the directed static network is calculated through GPDC, and then the directed static network is dynamically planned according to the time series, and the dynamic network is obtained on the time scale to realize the dynamic, dynamic and functional network of the brain under different emotions. Fine observation.
  • FIG. 1 is a flowchart of a method for constructing multi-type emotion recognition in a directed dynamic brain function network according to the first embodiment of the present invention
  • FIG. 2 is a flow chart of calculating a directed brain function network matrix by GPDC in a method for constructing a directed dynamic brain function network multi-emotion recognition according to the first embodiment of the present invention
  • FIG. 3 is a flow chart of dynamic brain network construction of a multi-type emotion recognition construction method for a directed dynamic brain function network according to the first embodiment of the present invention
  • Fig. 4 is a schematic diagram of an apparatus for executing a method for constructing a multi-type emotion recognition of a directed dynamic brain function network according to a second embodiment of the present invention.
  • the first embodiment of the present invention provides a method for constructing a multi-type emotion recognition of a directional dynamic brain function network, which includes the following steps:
  • Step S100 the client obtains EEG data, preprocesses the EEG data, and decomposes the EEG data into 4 standard sub-bands through wavelet packet transform;
  • Step S200 The client reads the EEG signal of the standard sub-band, divides the EEG signal according to a preset time window and step size, and calculates a directed static network through GPDC;
  • step S300 the client computer calculates the directed brain function network matrix, selects important connections according to the set sparsity, and binarizes them to obtain the brain function matrix of the important connections, and divides these connection matrices into time series Dynamic programming to obtain a directed dynamic brain network;
  • step S400 the client computer calculates the global efficiency and the local efficiency of information transmission in the dynamic brain network, divides these parameters into three categories according to different emotions, and calculates the statistical difference between them.
  • the EEG data can be collected by any prior art method.
  • the preferred collection method in this embodiment is to use the ESI neural scanning system (Advanced Medical Equipment Ltd), and the electrodes are placed in accordance with the international 10-20 system, totaling 62 Two electrodes, the sampling rate is 1000 Hz.
  • the wavelet packet transform is preferably adopted as the decomposition method in this embodiment. It is understandable that using GPDC (Generalized Partial Directed Coherence) to calculate a multi-band directed static network is the preferred embodiment of this embodiment, and other calculation methods that can achieve the same function can also be used, and will not be repeated here.
  • GPDC Generalized Partial Directed Coherence
  • this embodiment preferably pre-sets the set corresponding to the connection feature before classification, such as positive emotion set, calm emotion set, negative emotion set, etc. Others can also be used.
  • the classification method can be used to describe emotions. It can be understood that the use of the statistical difference method to calculate the difference of the parameters is the preferred embodiment of this embodiment.
  • the preprocessing includes independent component analysis and removal of EEG signal baselines.
  • this embodiment preferably down-samples the EEG data to 200 Hz to reduce the computational complexity, as the calculation of causality will be affected by various human factors, especially the falseness caused by blinking. Therefore, this embodiment preferably adopts Independent Component Analysis (ICA) method to remove artifacts mainly caused by electrooculogram signals, and then extracts and removes the baseline of each EEG signal to complete the removal of physiological artifacts.
  • ICA Independent Component Analysis
  • the GPDC calculating the directed brain function network matrix specifically includes the following steps:
  • Step S210 the client obtains the segmented EEG signal, and calculates the causal relationship between the two channels of the EEG signal;
  • step S220 the client computer calculates the GPDC value based on the causal relationship between nodes.
  • the unidirectional index of each channel is expressed by the following formula: Where A(f) is the corresponding Fourier transform of the time series x i (t), which satisfies the following formula: Where ⁇ ij describes the influence of signal j on i.
  • the causal relationship can be expressed as: When D is greater than 0, it means that node j has a greater impact on node i, and vice versa.
  • the directed dynamic brain network is obtained by the following specific steps:
  • Step S410 the client terminal reads a preset duration, and reads an EEG signal corresponding to the preset duration;
  • Step S420 the client obtains a binary directed brain function network through a sparse method
  • Step S430 the client computer calculates a directed dynamic brain function network according to the threshold of the binarized directed brain function network.
  • the preset duration of this embodiment is the unit duration of EEG data collection, which can be adjusted according to actual needs, and will not be repeated here.
  • both the sliding window and the step size are set to 2 seconds.
  • the sparse method is to set the threshold of the connection strength (1% to 16%), and keep the connections above the threshold. Therefore, calculate the dynamic brain network Where T is the number of corresponding static networks in the cycle.
  • the second embodiment of the present invention also provides a device for executing the emotion recognition method based on dynamic brain network.
  • the device is a smart device, such as a smart phone, a computer, and a tablet. Take the computer as an example.
  • the computer 4000 for executing the emotion recognition method based on the dynamic brain network includes a CPU unit 4100, and the CPU unit 4100 is configured to perform the following steps:
  • the client obtains EEG data, preprocesses the EEG data, and decomposes it into 4 standard sub-bands through wavelet packet transform;
  • the client reads the EEG signal of the standard sub-band, divides the EEG signal according to a preset time window and step size, and calculates a directed static network through GPDC;
  • the client computer calculates the directed brain function network matrix, selects important connections according to the set sparseness, and then binarizes them to obtain the brain function matrix of the important connections.
  • connection matrices are dynamically planned according to time series, Obtain a directed dynamic brain network;
  • the client computer calculates the global efficiency and local efficiency of information transmission to the dynamic brain network, divides these parameters into three categories according to different emotions, and calculates the statistical difference between them.
  • the CPU unit 4100 is further configured to perform the following steps:
  • the client obtains the segmented EEG signal, and calculates the mutual causality of each channel of the EEG signal;
  • the CPU unit 4100 is further configured to perform the following steps:
  • CPU unit 4100 is further configured to perform the following steps:
  • the client obtains the segmented EEG signal, and calculates the causal relationship between the two channels of the EEG signal;
  • the client computer calculates the GPDC value based on the causal relationship between nodes.
  • CPU unit 4100 is further configured to perform the following steps:
  • the client terminal reads the preset duration, and reads the EEG signal corresponding to the preset duration;
  • the client obtains the binary directed brain function network through a sparse method
  • the client computer calculates a directed dynamic brain function network according to the threshold of the binarized directed brain function network.
  • the computer 4000 and the CPU unit 4100 can be connected by a bus or other means.
  • the computer 4000 also includes a memory.
  • the memory can be used to store non-transitory software programs and non-transitory Computer-executable programs and modules, such as program instructions/modules corresponding to the device for executing the method for emotion recognition based on dynamic brain network in the embodiment of the present invention.
  • the computer 4000 runs the non-transient software programs, instructions and modules stored in the memory to control the CPU unit 4100 to execute various functional applications and data processing for performing the emotion recognition method based on the dynamic brain network, that is, to realize the implementation of the above method Example of emotion recognition method based on dynamic brain network.
  • the memory may include a storage program area and a storage data area, where the storage program area may store an operating system and an application program required by at least one function; the storage data area may store data created according to the use of the CPU unit 4100 and the like.
  • the memory may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices.
  • the memory may optionally include a memory remotely provided with respect to the CPU unit 4100, and these remote memories may be connected to the computer 4000 via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.
  • the one or more modules are stored in the memory, and when executed by the CPU unit 4100, the method for emotion recognition based on dynamic brain network in the foregoing method embodiment is executed.
  • the embodiment of the present invention also provides a computer-readable storage medium that stores computer-executable instructions that are executed by the CPU unit 4100 to realize the dynamic brain network described above for emotions. recognition methods.
  • the device embodiments described above are only illustrative, and the devices described as separate components may or may not be physically separated, that is, they may be located in one place, or they may be distributed on multiple network devices. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • each implementation manner can be implemented by means of software plus a general hardware platform.
  • All or part of the processes in the methods of the foregoing embodiments can be implemented by computer programs instructing relevant hardware.
  • the programs can be stored in a computer-readable storage medium.
  • the storage medium can be a magnetic disk, an optical disc, a read-only memory (Read Only Memory, ROM), or a random access memory (Random Access Memory, RAM), etc.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Public Health (AREA)
  • General Health & Medical Sciences (AREA)
  • Psychiatry (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Animal Behavior & Ethology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Artificial Intelligence (AREA)
  • Signal Processing (AREA)
  • Physiology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mathematical Physics (AREA)
  • Psychology (AREA)
  • Child & Adolescent Psychology (AREA)
  • Developmental Disabilities (AREA)
  • Educational Technology (AREA)
  • Hospice & Palliative Care (AREA)
  • Social Psychology (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

本发明公开了一种有向动态脑功能网络多类情绪识别构建方法及其装置，将获取的脑电数据进行预处理后分解成4个标准子频带，然后对每个标准子频带的脑电信号进行时间窗分割，通过GPDC计算出有向的脑功能网络矩阵，再根据时间序列对有向脑功能网络矩阵进行动态规划，从而获得在一段时间尺度上的动态网络，对得到的动态网络可以计算其信息交流的性能参数，性能参数包括全局效率和局部效率，对不同情绪下参数进行统计学分析。

Description

一种有向动态脑功能网络多类情绪识别构建方法及其装置 技术领域
本发明涉及脑电数据处理技术领域，特别是一种有向动态脑功能网络多类情绪识别构建方法及其装置。
背景技术
目前，随着脑电处理方法的发展，人们开始关注基于脑电的情绪识别，并尝试对情绪相关的神经活动进行解码，其本质为对脑电数据的处理，能够应用在疲劳驾驶检测、情绪状态检测等领域。现有方法主要采用构建静态网络对大脑活动进行描述，这种方法把处于不同时间窗口下的网络看作是独立组成部分。然而大脑的区域交互是随时间变化的，特别是大脑区域间的相互作用是高度动态和非平稳的，因此静态网络无法在时间尺度上反映整个大脑活动。此外，无向脑功能矩阵仅能反应通道间的相互关系，却无法获得通道间的信息流向，亦会造成构建脑网络时信息缺失。
发明内容
为了克服现有技术的不足，本发明的目的在于提供一种有向动态脑功能网络多类情绪识别构建方法及其装置，能够同时从信息流向和时间尺度上反映大脑活动，为挖掘不同情绪下人脑功能连接的更精确刻画提供数据基础。
本发明解决其问题所采用的技术方案是：第一方面，本发明提供了一种有向动态脑功能网络多类情绪识别构建方法，包括以下步骤：
客户端获取脑电数据，对所述脑电数据进行预处理，并通过小波包变换分解为4个标准子频带；
所述客户端读取所述标准子频带的脑电信号，根据预先设定的时间窗口和步长对所述脑电信号进行分割，并通过GPDC计算出有向静态网络；
所述客户端计算有向的脑功能网络矩阵，根据设定的稀疏度，筛选出重要连接，将其二值化后可获得重要连接的脑功能矩阵，将这些连接矩阵按照时间序列动态规划，得出有向动态脑网络；
所述客户端计算有向动态脑网络中信息传递的全局效率、局部效率，按照情绪的不同将这些参数分为三类，计算它们之间的统计学差异。
进一步，所述预处理包括独立分量分析、移除脑电信号基线。
进一步，所述GPDC计算出有向的脑功能网络矩阵具体包括以下步骤：
所述客户端获取分割后的脑电信号，计算出所述脑电信号两两通道相互的因果关系；
所述客户端计依据节点间的因果关系，计算出GPDC值。
进一步，所述有向动态脑网络由以下具体步骤得出：
所述客户端读取预设时长，读取与所述预设时长所对应的脑电信 号；
所述客户端通过稀疏方法获取二值化的有向脑功能网络；
所述客户端根据所述二值化的有向脑功能网络的阈值，计算出有向动态脑功能网络。
第二方面，本发明提供了一种用于执行有向动态脑功能网络多类情绪识别构建方法的装置，包括CPU单元，所述CPU单元用于执行以下步骤：
客户端获取脑电数据，对所述脑电数据进行预处理，并通过小波包变换分解为4个标准子频带；
所述客户端读取所述标准子频带的脑电信号，根据预先设定的时间窗口和步长对所述脑电信号进行分割，并通过GPDC计算出有向静态网络；
所述客户端计算有向的脑功能网络矩阵，根据设定的稀疏度，筛选出重要连接，将其二值化后可获得重要连接的脑功能矩阵，将这些连接矩阵按照时间序列动态规划，得出有向动态脑网络；
所述客户端计算有向动态脑网络中信息传递的全局效率、局部效率，按照情绪的不同将这些参数分为三类，计算它们之间的统计学差异。
进一步，所述CPU单元还用于执行以下步骤：
所述客户端获取分割后的脑电信号，计算出所述脑电信号两两通道相互的因果关系；
所述客户端计依据节点间的因果关系，计算出GPDC值。
进一步，所述CPU单元还用于执行以下步骤：
所述客户端读取预设时长，读取与所述预设时长所对应的脑电信号；
所述客户端通过稀疏方法获取二值化的有向脑功能网络；
所述客户端根据所述二值化的有向脑功能网络的阈值，计算出有向动态脑功能网络。
第三方面，本发明提供了一种用于执行有向动态脑功能网络多类情绪识别构建方法的设备，包括至少一个控制处理器和用于与至少一个控制处理器通信连接的存储器；存储器存储有可被至少一个控制处理器执行的指令，指令被至少一个控制处理器执行，以使至少一个控制处理器能够执行如上所述的有向动态脑功能网络多类情绪识别构建方法。
第四方面，本发明提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机可执行指令，计算机可执行指令用于使计算机执行如上所述的有向动态脑功能网络多类情绪识别构建方法。
第五方面，本发明还提供了一种计算机程序产品，所述计算机程序产品包括存储在计算机可读存储介质上的计算机程序，所述计算机程序包括程序指令，当所述程序指令被计算机执行时，使计算机执行如上所述的有向动态脑功能网络多类情绪识别构建方法。
本发明实施例中提供的一个或多个技术方案，至少具有如下有益 效果：本发明将获取的脑电数据进行预处理后分解成4个标准子频带，再对从每个标准子频带读取脑电信号后进行分割，通过GPDC计算出有向静态网络，再根据时间序列对有向静态网络进行动态规划，在时间尺度上得出动态网络，以实现不同情绪下对脑功能网络的动态、细微观测。
附图说明
下面结合附图和实例对本发明作进一步说明。
图1是本发明第一实施例提供的一种有向动态脑功能网络多类情绪识别构建方法的流程图；
图2是本发明第一实施例提供的一种有向动态脑功能网络多类情绪识别构建方法的GPDC计算出有向的脑功能网络矩阵的流程图；
图3是本发明第一实施例提供的一种有向动态脑功能网络多类情绪识别构建方法的动态脑网络构建的流程图；
图4是本发明第二实施例提供的一种用于执行有向动态脑功能网络多类情绪识别构建方法的装置示意图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本发明，并不用于限定本发明。
需要说明的是，如果不冲突，本发明实施例中的各个特征可以相互结合，均在本发明的保护范围之内。另外，虽然在装置示意图中进 行了功能模块划分，在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于装置中的模块划分，或流程图中的顺序执行所示出或描述的步骤。
参考图1，本发明的第一实施例提供了一种有向动态脑功能网络多类情绪识别构建方法，包括以下步骤：
步骤S100，客户端获取脑电数据，对所述脑电数据进行预处理，并通过小波包变换分解为4个标准子频带；
步骤S200，所述客户端读取所述标准子频带的脑电信号，根据预先设定的时间窗口和步长对所述脑电信号进行分割，并通过GPDC计算出有向静态网络；
步骤S300，所述客户端计算有向的脑功能网络矩阵，根据设定的稀疏度，筛选出重要连接，将其二值化后可获得重要连接的脑功能矩阵，将这些连接矩阵按照时间序列动态规划，得出有向动态脑网络；
步骤S400，所述客户端计算有向动态脑网络中信息传递的全局效率、局部效率，按照情绪的不同将这些参数分为三类，计算它们之间的统计学差异。
其中，需要说明的是，脑电数据可以通过任意现有技术的方式采集，本实施例优选采集方式为采用ESI神经扫描系统(Advanced Medical Equipment Ltd)，电极按照国际10-20系统放置，共62个电极，采样率为1000Hz。
其中，需要说明的是，由于db4的Daubechies小波和6级分解 小波适用于脑电频带提取，因此本实施例优选采用小波包变换为分解方式。可以理解的是，采用GPDC(Generalized partial directed coherence，广义一致性算法)计算多频带有向静态网络为本实施例的优选，也可以采用其他能够实现相同功能的计算方式，在此不再赘述。
其中，需要说明的是，为了对情绪进行描述，本实施例优选在分类之前预先设定好与连接特征所对应的集合，例如积极情绪集合、平静情绪集合和消极情绪集合等，也可以采用其他分类方式，能够用于描述情绪即可。可以理解的是，采用统计学差异方法计算参数的差异为本实施例的优选。
进一步，在本发明的另一个实施例中，所述预处理包括独立分量分析、移除脑电信号基线。
其中，需要说明的是，在预处理阶段，本实施例优选将脑电数据降采样到200Hz以降低计算复杂度，由于因果关系的计算会受到各种人为因素的影响，尤其是眨眼导致的虚假的连接，因此本实施例优选采用独立分量分析(Independent Component Analysis,ICA)方法去除主要由眼电信号引起的伪迹，再提取并移除每个脑电信号的基线，完成去除生理伪迹。
参考图2，进一步，在本发明的另一个实施例中，所述GPDC计算出有向的脑功能网络矩阵具体包括以下步骤：
步骤S210，所述客户端获取分割后的脑电信号，计算出所述脑电信号两两通道相互的因果关系；
步骤S220，所述客户端计依据节点间的因果关系，计算出GPDC值。
在本实施例中，以下以一个具体实施例对GPDC的计算进行说明：
对于每个标准子频带，依据时间窗口和步长分割脑电信号：窗长/步长＝4/2s。在步骤S210中，每个通道的单方向指数由以下公式表示：
Figure PCTCN2020078291-appb-000001
其中A(f)为时间序列x i(t)的对应傅里叶变换，满足以下公式：
Figure PCTCN2020078291-appb-000002
其中π ij描述了信号j对i的影响。在综合考虑两个方向的因果关系可以表示为：
Figure PCTCN2020078291-appb-000003
其中当D大于0表示节点j对节点i影响更大，反之亦然。
参考图3，进一步，在本发明的另一个实施例中，所述有向动态脑网络由以下具体步骤得出：
步骤S410，所述客户端读取预设时长，读取与所述预设时长所对应的脑电信号；
步骤S420，所述客户端通过稀疏方法获取二值化的有向脑功能网络；
步骤S430，所述客户端根据所述二值化的有向脑功能网络的阈值，计算出有向动态脑功能网络。
其中，需要说明的是，本实施例的预设时长为进行脑电数据采集 时的单位时长，可以根据实际需求调整，在此不再赘述。
其中，需要说明的是，在本实施例中，滑动窗口和步长都设置为2秒。通过稀疏方法获取二值化的有向静态网络
Figure PCTCN2020078291-appb-000004
稀疏方法是设置连接强度的阈值(1％到16％)，保留阈值上的连接。因此，计算出动态脑网络
Figure PCTCN2020078291-appb-000005
其中T是周期中相应静态网络的数目。
参照图4，本发明的第二实施例还提供了一种用于执行基于动态脑网络进行情绪识别方法的装置，该装置为智能设备，例如智能手机、计算机和平板电脑等，本实施例以计算机为例加以说明。
在该用于执行基于动态脑网络进行情绪识别方法的计算机4000中，包括CPU单元4100，所述CPU单元4100用于执行以下步骤：
客户端获取脑电数据，对所述脑电数据进行预处理，并通过小波包变换分解为4个标准子频带；
所述客户端读取所述标准子频带的脑电信号，根据预先设定的时间窗口和步长对所述脑电信号进行分割，并通过GPDC计算出有向静态网络；
所述客户端计算有向的脑功能网络矩阵，根据设定的稀疏度，筛选出重要连接，将其二值化后可获得重要连接的脑功能矩阵，将这些连接矩阵按照时间序列动态规划，得出有向动态脑网络；
所述客户端计算有向动态脑网络中信息传递的全局效率、局部效率，按照情绪的不同将这些参数分为三类，计算它们之间的统计学差 异。
进一步，本发明的另一个实施例中，所述CPU单元4100还用于执行以下步骤：
所述客户端获取分割后的脑电信号，计算出所述脑电信号每个通道的相互的因果关系；
所述客户端节点相互的因果关系，计算出GPDC值；
进一步，本发明的另一个实施例中，所述CPU单元4100还用于执行以下步骤：
进一步，所述CPU单元4100还用于执行以下步骤：
所述客户端获取分割后的脑电信号，计算出所述脑电信号两两通道相互的因果关系；
所述客户端计依据节点间的因果关系，计算出GPDC值。
进一步，所述CPU单元4100还用于执行以下步骤：
所述客户端读取预设时长，读取与所述预设时长所对应的脑电信号；
所述客户端通过稀疏方法获取二值化的有向脑功能网络；
所述客户端根据所述二值化的有向脑功能网络的阈值，计算出有向动态脑功能网络。
计算机4000和CPU单元4100之间可以通过总线或者其他方式连接，计算机4000中还包括存储器，所述存储器作为一种非暂态计算机可读存储介质，可用于存储非暂态软件程序、非暂态性计算机可 执行程序以及模块，如本发明实施例中的用于执行基于动态脑网络进行情绪识别方法的设备对应的程序指令/模块。计算机4000通过运行存储在存储器中的非暂态软件程序、指令以及模块，从而控制CPU单元4100执行用于执行基于动态脑网络进行情绪识别方法的各种功能应用以及数据处理，即实现上述方法实施例的基于动态脑网络进行情绪识别方法。
存储器可以包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需要的应用程序；存储数据区可存储根据CPU单元4100的使用所创建的数据等。此外，存储器可以包括高速随机存取存储器，还可以包括非暂态存储器，例如至少一个磁盘存储器件、闪存器件、或其他非暂态固态存储器件。在一些实施方式中，存储器可选包括相对于CPU单元4100远程设置的存储器，这些远程存储器可以通过网络连接至该计算机4000。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
所述一个或者多个模块存储在所述存储器中，当被所述CPU单元4100执行时，执行上述方法实施例中的基于动态脑网络进行情绪识别方法。
本发明实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可执行指令，该计算机可执行指令被CPU单元4100执行，实现上述所述的动态脑网络进行情绪识别方法。
以上所描述的装置实施例仅是示意性的，其中所述作为分离部件 说明的装置可以是或者也可以不是物理上分开的，即可以位于一个地方，或者也可以分布到多个网络装置上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
需要说明的是，由于本实施例中的用于执行基于动态脑网络进行情绪识别方法的装置与上述的有向动态脑功能网络多类情绪识别构建方法基于相同的发明构思，因此，方法实施例中的相应内容同样适用于本装置实施例，此处不再详述。
通过以上的实施方式的描述，本领域技术人员可以清楚地了解到各实施方式可借助软件加通用硬件平台的方式来实现。本领域技术人员可以理解实现上述实施例方法中的全部或部分流程是可以通过计算机程序来指令相关的硬件来完成，所述的程序可存储于计算机可读取存储介质中，该程序在执行时，可包括如上述方法的实施例的流程。其中，所述的存储介质可为磁碟、光盘、只读存储记忆体(ReadOnly Memory,ROM)或随机存储记忆体(Random Access Memory,RAM)等。
以上是对本发明的较佳实施进行了具体说明，但本发明并不局限于上述实施方式，熟悉本领域的技术人员在不违背本发明精神的前提下还可作出种种的等同变形或替换，这些等同的变形或替换均包含在本申请权利要求所限定的范围内。

Claims (8)

  1. 一种有向动态脑功能网络多类情绪识别构建方法，其特征在于，包括以下步骤：
    客户端获取脑电数据，对所述脑电数据进行预处理，并通过小波包变换分解为4个标准子频带；
    所述客户端读取所述标准子频带的脑电信号，根据预先设定的时间窗口和步长对所述脑电信号进行分割，并通过GPDC计算出有向静态网络；
    所述客户端计算有向的脑功能网络矩阵，根据设定的稀疏度，筛选出重要连接，将其二值化后可获得重要连接的脑功能矩阵，将这些连接矩阵按照时间序列动态规划，得出有向动态脑网络；
    所述客户端计算有向动态脑网络中信息传递的全局效率、局部效率，按照情绪的不同将这些参数分为三类，计算它们之间的统计学差异。
  2. 根据权利要求1所述的一种有向动态脑功能网络多类情绪识别构建方法，其特征在于：所述预处理包括独立分量分析、移除脑电信号基线。
  3. 根据权利要求1所述的一种有向动态脑功能网络多类情绪识别构建方法，其特征在于，所述GPDC计算出有向的脑功能网络矩阵具体包括以下步骤：
    所述客户端获取分割后的脑电信号，计算出所述脑电信号两两通道相互的因果关系；
    所述客户端计依据节点间的因果关系，计算出GPDC值。
  4. 根据权利要求1所述的一种有向动态脑功能网络多类情绪识别构建方法，其特征在于，所述有向动态脑网络由以下具体步骤得出：所述客户端读取预设时长，读取与所述预设时长所对应的脑电信号；所述客户端通过稀疏方法获取二值化的有向脑功能网络；
    所述客户端根据所述二值化的有向脑功能网络的阈值，计算出有向动态脑功能网络。
  5. 一种用于执行有向动态脑功能网络多类情绪识别构建方法的装置，其特征在于，包括CPU单元，所述CPU单元用于执行以下步骤：
    客户端获取脑电数据，对所述脑电数据进行预处理，并通过小波包变换分解为4个标准子频带；
    所述客户端读取所述标准子频带的脑电信号，根据预先设定的时间窗口和步长对所述脑电信号进行分割，并通过GPDC计算出有向静态网络；
    所述客户端计算有向的脑功能网络矩阵，根据设定的稀疏度，筛选出重要连接，将其二值化后可获得重要连接的脑功能矩阵，将这些连接矩阵按照时间序列动态规划，得出有向动态脑网络；
    所述客户端计算有向动态脑网络中信息传递的全局效率、局部效率，按照情绪的不同将这些参数分为三类，计算它们之间的统计学差异。
  6. 根据权利要求5所述的一种用于执行有向动态脑功能网络多类情绪识别构建方法的装置，其特征在于，所述CPU单元还用于执行以下 步骤：
    所述客户端获取分割后的脑电信号，计算出所述脑电信号两两通道相互的因果关系；
    所述客户端计依据节点间的因果关系，计算出GPDC值。
  7. 根据权利要求5所述的一种用于执行有向动态脑功能网络多类情绪识别构建方法的装置，其特征在于，所述CPU单元还用于执行以下步骤：
    所述客户端读取预设时长，读取与所述预设时长所对应的脑电信号；所述客户端通过稀疏方法获取二值化的有向脑功能网络；
    所述客户端根据所述二值化的有向脑功能网络的阈值，计算出有向动态脑功能网络。
  8. 一种计算机可读存储介质，其特征在于：所述计算机可读存储介质存储有计算机可执行指令，所述计算机可执行指令用于使计算机执行如权利要求1-4任一项所述的一种基于有向动态脑功能网络多类情绪识别构建方法。
PCT/CN2020/078291 2020-02-12 2020-03-06 一种有向动态脑功能网络多类情绪识别构建方法及其装置 WO2021159571A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010088922.8A CN111150392A (zh) 2020-02-12 2020-02-12 一种有向动态脑功能网络多类情绪识别构建方法及其装置
CN202010088922.8 2020-02-12

Publications (1)

Publication Number Publication Date
WO2021159571A1 true WO2021159571A1 (zh) 2021-08-19

Family

ID=70565538

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/078291 WO2021159571A1 (zh) 2020-02-12 2020-03-06 一种有向动态脑功能网络多类情绪识别构建方法及其装置

Country Status (2)

Country Link
CN (1) CN111150392A (zh)
WO (1) WO2021159571A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114818786A (zh) * 2022-04-06 2022-07-29 五邑大学 通道筛选方法、情绪识别方法、系统及存储介质

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111544017A (zh) * 2020-05-25 2020-08-18 五邑大学 基于gpdc图卷积神经网络的疲劳检测方法、装置及存储介质
CN112057089A (zh) * 2020-08-31 2020-12-11 五邑大学 情绪识别方法、装置及存储介质
CN113509148B (zh) * 2021-04-28 2022-04-22 东北大学 一种基于混合高阶脑网络的精神分裂症检测系统
CN113558641A (zh) * 2021-06-23 2021-10-29 瑞尔明康（杭州）医疗科技有限公司 脑功能网络构建方法、装置、电子设备及存储介质

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104473636A (zh) * 2014-12-30 2015-04-01 天津大学 一种基于偏定向相干的疲劳脑网络分析方法
CN108764043A (zh) * 2018-04-26 2018-11-06 太原理工大学 基于动态功能连接的熵值的脑电分类方法
CN109524112A (zh) * 2018-12-26 2019-03-26 杭州电子科技大学 一种基于部分定向相干法的脑功能网络构建方法
CN109770924A (zh) * 2019-01-24 2019-05-21 五邑大学 一种基于广义一致性构建脑功能网络与相关向量机的疲劳分类方法
CN109793528A (zh) * 2019-01-28 2019-05-24 华南理工大学 一种基于动态脑功能网络的情绪分类方法
CN109893126A (zh) * 2019-03-21 2019-06-18 杭州电子科技大学 基于脑功能网络特征的癫痫发作预测方法
US20190261881A1 (en) * 2014-07-22 2019-08-29 Leonidas D. Iasemidis Biomarkers for Determining Susceptibility to SUDEP
CN110584684A (zh) * 2019-09-11 2019-12-20 五邑大学 一种驾驶疲劳相关的eeg功能连接动态特性的分析方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6916527B2 (ja) * 2018-05-25 2021-08-11 国立大学法人広島大学 感性評価装置、感性評価方法および感性多軸モデル構築方法
CN109259764B (zh) * 2018-07-27 2021-05-14 昆明理工大学 一种确定动态脑功能网络阈值方法
CN109363671B (zh) * 2018-10-30 2021-10-01 中国人民解放军战略支援部队信息工程大学 基于ssvep与erp融合的情绪动态脑网络图的构建方法
CN110236534B (zh) * 2019-05-20 2020-10-09 北京航空航天大学 一种基于脑电相关性网络弹性的专注力评估方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190261881A1 (en) * 2014-07-22 2019-08-29 Leonidas D. Iasemidis Biomarkers for Determining Susceptibility to SUDEP
CN104473636A (zh) * 2014-12-30 2015-04-01 天津大学 一种基于偏定向相干的疲劳脑网络分析方法
CN108764043A (zh) * 2018-04-26 2018-11-06 太原理工大学 基于动态功能连接的熵值的脑电分类方法
CN109524112A (zh) * 2018-12-26 2019-03-26 杭州电子科技大学 一种基于部分定向相干法的脑功能网络构建方法
CN109770924A (zh) * 2019-01-24 2019-05-21 五邑大学 一种基于广义一致性构建脑功能网络与相关向量机的疲劳分类方法
CN109793528A (zh) * 2019-01-28 2019-05-24 华南理工大学 一种基于动态脑功能网络的情绪分类方法
CN109893126A (zh) * 2019-03-21 2019-06-18 杭州电子科技大学 基于脑功能网络特征的癫痫发作预测方法
CN110584684A (zh) * 2019-09-11 2019-12-20 五邑大学 一种驾驶疲劳相关的eeg功能连接动态特性的分析方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
ADKINSON JOSHUA A.; LIU RUI; VLACHOS IOANNIS; IASEMIDIS LEONIDAS: "Connectivity Analysis for Epileptogenic Focus Localization", 2016 32ND SOUTHERN BIOMEDICAL ENGINEERING CONFERENCE (SBEC), 11 March 2016 (2016-03-11), pages 3 - 4, XP032895960, ISBN: 978-1-5090-2132-1, DOI: 10.1109/SBEC.2016.38 *
LUIZ A. BACCALA ; K. SAMESHIMA ; D. Y. TAKAHASHI: "Generalized Partial Directed Coherence", DIGITAL SIGNAL PROCESSING, 2007 15TH INTERNATIONAL CONFERENCE ON, 1 July 2007 (2007-07-01), Pi, pages 163 - 166, XP031125520, ISBN: 978-1-4244-0881-8 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114818786A (zh) * 2022-04-06 2022-07-29 五邑大学 通道筛选方法、情绪识别方法、系统及存储介质
CN114818786B (zh) * 2022-04-06 2024-03-01 五邑大学 通道筛选方法、情绪识别方法、系统及存储介质

Also Published As

Publication number Publication date
CN111150392A (zh) 2020-05-15

Similar Documents

Publication Publication Date Title
WO2021159571A1 (zh) 一种有向动态脑功能网络多类情绪识别构建方法及其装置
Kaya et al. 1D-local binary pattern based feature extraction for classification of epileptic EEG signals
WO2021046949A1 (zh) 一种驾驶疲劳相关的eeg功能连接动态特性的分析方法
Kaziha et al. A convolutional neural network for seizure detection
CN114305333B (zh) 一种基于脑网络和深度学习的脑电信号识别方法
CN114224360B (zh) 一种基于改进emd-ica的eeg信号处理方法、设备及存储介质
JP2018067287A (ja) センサ信号の自己学習による異常検出
Olokodana et al. Real-time automatic seizure detection using ordinary kriging method in an edge-IoMT computing paradigm
Al-Salman et al. An efficient approach for EEG sleep spindles detection based on fractal dimension coupled with time frequency image
CN111839506A (zh) 脑力负荷检测方法及装置
Wang et al. Biomedical time series clustering based on non-negative sparse coding and probabilistic topic model
CN115919330A (zh) 基于多层次se注意力和图卷积的eeg情绪状态分类方法
Jiang et al. Analytical comparison of two emotion classification models based on convolutional neural networks
Khosla et al. EEG‐based automatic multi‐class classification of epileptic seizure types using recurrence plots
Yazid et al. Variable step dynamic threshold local binary pattern for classification of atrial fibrillation
Gandhi et al. Denoising time series data using asymmetric generative adversarial networks
CN116350239A (zh) 一种脑电信号专注度分类方法及系统
CN110584662A (zh) 基于脑电信息的自闭症风险量化方法、装置和存储介质
Lopez et al. HAPPILEE: the Harvard automated processing pipeline in low electrode electroencephalography, a standardized software for low density EEG and ERP data
Grilo Jr et al. Artifact removal for emotion recognition using mutual information and Epanechnikov kernel
CN113855020A (zh) 用于情绪识别的方法及装置、计算机设备、存储介质
EP4339965A1 (en) Method and system for eeg motor imagery classification
Chakrabarty et al. Denoising and stability using independent component analysis in high dimensions–visual inspection still required
Noorbasha et al. Removal of motion artifacts from EEG records by overlap segmentation SSA with modified grouping criteria for portable or wearable applications
CN116035594B (zh) 一种基于分割-降噪网络的脑电伪迹去除方法

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20918787

Country of ref document: EP

Kind code of ref document: A1