Depression population determination system and method based on asynchronous electroencephalogram

Abstract

The invention discloses a depression crowd judgment system and method based on asynchronous electroencephalogram, and the system comprises an asynchronous electroencephalogram collection module, an asynchronous electroencephalogram signal preprocessing module, an asynchronous electroencephalogram time sequence feature extraction module, an asynchronous electroencephalogram depression crowd judgment model and a feedback module. The output end of the asynchronous electroencephalogram collection module is connected with the input end of the asynchronous electroencephalogram signal preprocessing module, the output end of the asynchronous electroencephalogram signal preprocessing module is connected with the input end of the asynchronous electroencephalogram time sequence feature extraction module, and the output end of the asynchronous electroencephalogram time sequence feature extraction module is connected with the input end of the asynchronous electroencephalogram depression crowd judgment model. And the output end of the asynchronous electroencephalogram depression crowd judgment model is connected with the input end of the feedback module. According to the depression population judgment system and method based on the asynchronous electroencephalogram, accurate judgment and early recognition of depression are achieved by collecting and analyzing time sequence characteristics of the electroencephalogram signals and combining a machine learning algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of electroencephalogram (EEG) signal analysis and mental health monitoring, and in particular to a system and method for determining depressed people based on asynchronous EEG. Background Art

[0002] Currently, the diagnosis of depression mainly relies on clinical symptom assessment, psychological questionnaires, and physician judgment. Although these methods can effectively screen patients for depression, they have the following limitations:

[0003] High Subjectivity: Traditional depression assessment methods rely on the physician's subjective judgment, and results may be affected by factors such as physician experience and patient self-report. Inadequate Identification Sensitivity: Many patients with depression may mask their mood swings due to social and family pressures, resulting in the inability of traditional diagnostic methods to accurately identify early or mild depression. Lack of Physiological Indicators: Existing depression diagnostic methods lack objective physiological indicators to support them, resulting in an inaccurate basis for pathological research and early diagnosis. To achieve more objective and accurate depression diagnosis, scientists have begun to focus on electroencephalogram (EEG) signals, particularly the potential of EEG temporal characteristics for identifying depression. Although EEG, as a non-invasive, real-time physiological signal detection technology, has been applied in neurological disease research, existing EEG-based depression assessment methods have the following issues: Limitations of Synchronous Acquisition Methods: Most existing EEG acquisition methods rely on synchronous acquisition mechanisms, which cannot effectively capture the asynchrony between different brain regions when processing emotional information. This makes EEG temporal analysis unable to reflect actual EEG activity. Lack of high-precision timing analysis: Existing methods often lack detailed analysis of EEG signal timing, making it difficult to extract key timing features related to depression from asynchronous signals. Large individual differences: The EEG activity of patients with depression varies significantly from person to person, and traditional methods fail to account for these differences, resulting in low recognition rates and accuracy.

[0004] Therefore, there is an urgent need for a new depression identification method that can accurately capture asynchronous activities in EEG signals and combine it with high temporal precision analysis methods to improve the sensitivity and accuracy of diagnosis. Summary of the Invention

[0005] The purpose of this invention is to provide a system and method for identifying depressed people based on asynchronous EEG, which can realize accurate determination and early identification of depression by collecting and analyzing the time series characteristics of EEG signals and combining machine learning algorithms, especially in the application of asynchronous EEG signal collection and high time series precision analysis.

[0006] The present invention provides a system and method for determining depressed people based on asynchronous EEG, comprising an asynchronous EEG acquisition module, an asynchronous EEG signal preprocessing module, an asynchronous EEG time series feature extraction module, an asynchronous EEG depression population determination model and a feedback module. The asynchronous EEG acquisition module comprises a specific event stimulation unit and an asynchronous sampling mechanism unit; the output end of the asynchronous EEG acquisition module is connected to the input end of the asynchronous EEG signal preprocessing module, the output end of the asynchronous EEG signal preprocessing module is connected to the input end of the asynchronous EEG time series feature extraction module, the output end of the asynchronous EEG time series feature extraction module is connected to the input end of the asynchronous EEG depression population determination model, and the output end of the asynchronous EEG depression population determination model is connected to the input end of the feedback module.

[0007] Preferably, the specific event stimulation unit includes positive, negative and neutral audio stimulation; the asynchronous sampling mechanism unit includes an asynchronous brain-like chip and an analog-to-digital converter ADC.

[0008] Preferably, the asynchronous EEG time series feature extraction module includes a time series analysis method.

[0009] Preferably, the asynchronous EEG depression population determination model adopts an RCNN-LSTM fusion deep learning model.

[0010] Preferably, a method for determining depression based on asynchronous EEG comprises the following steps:

[0011] Step S1: In the asynchronous EEG acquisition module, brain area activities are induced by stimulating a specific event unit in the asynchronous EEG acquisition module, and EEG signals of different brain areas are accurately acquired by an asynchronous sampling mechanism unit in the asynchronous EEG acquisition module;

[0012] The subjects were stimulated with 6 segments of positive, negative, and neutral audio. They closed their eyes and listened to 18 audio segments, each lasting 1-6 seconds. After a specific event was induced at each EEG acquisition point, the event feature matching unit of the asynchronous brain-like chip matched the event feature, controlled the pulse generator to generate an enable signal, activated the analog-to-digital converter (ADC) to acquire EEG signals for 6 seconds, and stopped playing the current audio segment after all channels were activated. After the acquisition was completed, the next audio segment was played, and the cycle continued until all audio segments were played.

[0013] Step S2: in the asynchronous EEG signal preprocessing module, the collected EEG signal is subjected to denoising, filtering and segmentation processing;

[0014] Step S3: In the asynchronous EEG timing feature extraction module, based on the timing analysis method, extract the key timing features of the EEG signal;

[0015] Step S4: In the asynchronous EEG depression population determination model, the RCNN-LSTM fusion deep learning model is used to determine depression based on the extracted time series features;

[0016] RCNN is used to extract intermediate layer features and emotional information from EEG data and time series information. LSTM is used to fuse time series features and model temporal characteristics. Cross-validation experiments are performed to optimize the combination of brain regions and audio stimulation data segments to obtain the optimal classification results.

[0017] Step S5: In the feedback module, the diagnosis results are output in real time, personalized treatment recommendations are provided to patients, and the system is linked with other medical devices to form an intelligent diagnosis and treatment closed loop.

[0018] Preferably, in step S1, the 18 audio segments are all selected from the International Emotional Sound System Database IADS-2.

[0019] Preferably, in step S1, high-precision timing calibration of EEG acquisition points is performed using an asynchronous brain-like chip.

[0020] Therefore, the present invention adopts the above-mentioned asynchronous EEG-based depressed population determination system and method, and realizes accurate determination and early identification of depression by collecting and analyzing the timing characteristics of EEG signals and combining machine learning algorithms, especially in the application of asynchronous EEG signal acquisition and high timing precision analysis.

[0021] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is an overall system block diagram of a system and method for determining depressed people based on asynchronous EEG in the present invention;

[0023] Figure 2 This is a schematic diagram of the experimental process of audio emotion-induced asynchronous EEG acquisition in a system and method for determining depressed people based on asynchronous EEG in the present invention;

[0024] Figure 3 This is a schematic diagram of the audio stimulation process of a system and method for determining depressed people based on asynchronous EEG in the present invention;

[0025] Figure 4 This is a schematic diagram of an RCNN-LSTM fusion deep learning model of a system and method for determining depression based on asynchronous EEG in the present invention;

[0026] Figure 5 This is a schematic diagram of the overall process of a system and method for determining depressed people based on asynchronous EEG of the present invention. DETAILED DESCRIPTION

[0027] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0028] Unless otherwise defined, technical or scientific terms used in the present invention shall have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs.

[0029] The words "first", "second" and similar terms used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Words such as "include" or "comprise" mean that the elements or objects preceding the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0030] Example 1

[0031] like Figure 1-Figure 5 As shown, the present invention provides a system and method for identifying depressed individuals based on asynchronous EEG, comprising an asynchronous EEG acquisition module, an asynchronous EEG signal preprocessing module, an asynchronous EEG time series feature extraction module, an asynchronous EEG depression identification model, and a feedback module. The asynchronous EEG acquisition module includes a specific event stimulation unit and an asynchronous sampling mechanism unit. The specific event stimulation unit includes positive, negative, and neutral audio stimulation. The asynchronous sampling mechanism unit includes an asynchronous brain-inspired chip and an analog-to-digital converter (ADC).

[0032] The output of the asynchronous EEG acquisition module is connected to the input of the asynchronous EEG signal preprocessing module to transmit the acquired raw asynchronous EEG signals to the preprocessing module. The output of the asynchronous EEG signal preprocessing module is connected to the input of the asynchronous EEG time series feature extraction module to feed the preprocessed signals into the feature extraction module; the asynchronous EEG time series feature extraction module includes a time series analysis method.

[0033] The output of the asynchronous EEG time series feature extraction module is connected to the input of the asynchronous EEG depression identification model, providing the model with time series feature data. The asynchronous EEG depression identification model uses an RCNN-LSTM fusion deep learning model. The output of the asynchronous EEG depression identification model is connected to the input of the feedback module, transmitting the identification results to the feedback module.

[0034] Step S1: In the asynchronous EEG acquisition module, brain area activities are induced by stimulating the specific event unit in the asynchronous EEG acquisition module, and EEG signals of different brain areas are accurately acquired by the asynchronous sampling mechanism unit in the asynchronous EEG acquisition module.

[0035] The subjects were stimulated with 6 segments of positive, negative and neutral audio. In step S1, all 18 audio segments were selected from the International Emotional Sound System Database IADS-2. The subjects closed their eyes and listened to the 18 audio segments, each lasting 1-6 seconds. After each EEG acquisition point induced a specific event (such as a spike signal, specific frequency band power, or multiple neural pulses), the event feature matching unit of the asynchronous brain-like chip matched the pulse generator to generate an enable signal, activating the analog-to-digital converter ADC to acquire 6 seconds of EEG signals. After all channels were activated, the current audio segment was stopped, and the next audio segment was played after the acquisition was completed. The cycle continued until all audio segments were played. In step S1, the EEG acquisition points were calibrated with high precision timing by the asynchronous brain-like chip.

[0036] Step S2: In the asynchronous EEG signal preprocessing module, the collected EEG signal is subjected to denoising, filtering and segmentation processing to remove interference and ensure signal purity.

[0037] Step S3: In the asynchronous EEG timing feature extraction module, based on the timing analysis method, the key timing features of the EEG signal are extracted, including the time sequence of brain area excitation, phase change, amplitude change, etc.

[0038] Step S4: The asynchronous EEG depression identification model uses convolutional neural networks (CNNs) and long short-term memory (LSTM) networks from deep learning as the foundation. An RCNN-LSTM fusion deep learning model is designed to improve the recognition rate of depressive disorders. This RCNN-LSTM fusion deep learning model is then used to identify depression based on extracted temporal features.

[0039] The RCNN extracts intermediate features and emotional information from EEG data and temporal information. A recurrent convolutional neural network (RCNN), based on an improved traditional CNN model, is used to extract intermediate features and the characteristics of the information contained in the EEG data and temporal information of audio stimulation from different brain regions. This allows for the full exploration of the temporal characteristics of EEG signals from different brain regions and the inherent emotional information of EEG signals under different audio stimulations. At the same time, attention is paid to the temporal correlation information in the EEG signal input from each brain region, enabling this temporal information to be better reflected in the activation of each module and the correlation mechanism within the model.

[0040] By using LSTM to fuse time series features and model temporal characteristics, the method comprehensively considers the temporal characteristics of asynchronous EEG signals, more comprehensively preserving their abstract and deep features. Finally, through continuous cross-validation experiments, different combinations of brain regions and audio stimulation data segments were tried to find the optimal combination. This combination was optimized through cross-validation experiments to achieve the best classification results.

[0041] Step S5: In the feedback module, the diagnosis results are output in real time, personalized treatment recommendations are provided to patients, and the system is linked with other medical devices to form an intelligent diagnosis and treatment closed loop.

[0042] Therefore, the present invention adopts the aforementioned asynchronous EEG-based depression identification system and method to accurately reflect the asynchronous characteristics of EEG. Unlike traditional synchronous acquisition methods, the present invention adopts an asynchronous EEG acquisition mechanism, inducing brain activity through specific events such as audio and visual stimulation, and collecting data based on the triggering timing of the events, which can accurately reflect the true timing characteristics of EEG. Through asynchronous EEG acquisition technology, the present invention can more accurately reflect the asynchronous activity of different brain regions during information processing, overcoming the limitations of traditional synchronous acquisition methods. High-precision timing analysis: By using advanced timing analysis methods (such as LSTM) to extract fine timing features from EEG signals, a more accurate physiological basis is provided for the early diagnosis of depression. Using advanced timing analysis methods to perform detailed analysis of EEG signals can reveal the EEG characteristics of patients with depression at the timing level, providing higher accuracy than existing methods. Personalized determination: Incorporating individual differences in EEG, the present invention uses machine learning algorithms to train a personalized depression determination model, which can effectively improve diagnostic accuracy and sensitivity. The depression determination model of the present invention can make personalized judgments based on the differences in EEG characteristics of each individual, reducing the dependence on standardized EEG models and improving the sensitivity and accuracy of diagnosis.

[0043] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can still be modified or replaced by equivalents, and these modifications or equivalent replacements cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A system for determining depression based on asynchronous EEG, characterized by: It includes an asynchronous EEG acquisition module, an asynchronous EEG signal preprocessing module, an asynchronous EEG time series feature extraction module, an asynchronous EEG depression population determination model and a feedback module. The asynchronous EEG acquisition module includes a specific event stimulation unit and an asynchronous sampling mechanism unit; the output end of the asynchronous EEG acquisition module is connected to the input end of the asynchronous EEG signal preprocessing module, the output end of the asynchronous EEG signal preprocessing module is connected to the input end of the asynchronous EEG time series feature extraction module, the output end of the asynchronous EEG time series feature extraction module is connected to the input end of the asynchronous EEG depression population determination model, and the output end of the asynchronous EEG depression population determination model is connected to the input end of the feedback module.

2. The system for determining depression based on asynchronous EEG according to claim 1, characterized in that: The specific event stimulation unit includes positive, negative and neutral audio stimulation; the asynchronous sampling mechanism unit includes an asynchronous brain-like chip and an analog-to-digital converter ADC.

3. The system for determining depression based on asynchronous EEG according to claim 1, characterized in that: The asynchronous EEG time series feature extraction module includes a time series analysis method.

4. The system for determining depression based on asynchronous EEG according to claim 1, characterized in that: The asynchronous EEG depression population judgment model adopts the RCNN-LSTM fusion deep learning model.

5. A method for determining depression based on asynchronous EEG according to any one of claims 1 to 4, characterized in that: The following steps are involved: Step S1: In the asynchronous EEG acquisition module, brain area activities are induced by stimulating the specific event unit in the asynchronous EEG acquisition module, and EEG signals of different brain areas are accurately collected by the asynchronous sampling mechanism unit in the asynchronous EEG acquisition module. The subjects were stimulated with 6 segments of positive, negative, and neutral audio. They closed their eyes and listened to 18 audio segments, each lasting 1-6 seconds. After a specific event was induced at each EEG acquisition point, the event feature matching unit of the asynchronous brain-like chip matched the event feature, controlled the pulse generator to generate an enable signal, activated the analog-to-digital converter (ADC) to acquire EEG signals for 6 seconds, and stopped playing the current audio segment after all channels were activated. After the acquisition was completed, the next audio segment was played, and the cycle continued until all audio segments were played. Step S2: in the asynchronous EEG signal preprocessing module, the collected EEG signal is subjected to denoising, filtering and segmentation processing; Step S3: In the asynchronous EEG timing feature extraction module, based on the timing analysis method, extract the key timing features of the EEG signal; Step S4: In the asynchronous EEG depression population determination model, the RCNN-LSTM fusion deep learning model is used to determine depression based on the extracted time series features; RCNN is used to extract intermediate layer features and emotional information from EEG data and time series information. LSTM is used to fuse time series features and model temporal characteristics. Cross-validation experiments are performed to optimize the combination of brain regions and audio stimulation data segments to obtain the optimal classification results. Step S5: In the feedback module, the diagnosis results are output in real time, personalized treatment recommendations are provided to patients, and the system is linked with other medical devices to form an intelligent diagnosis and treatment closed loop.

6. The method of the asynchronous EEG-based depression population determination system according to claim 5, characterized in that: In step S1 , 18 audio segments are selected from the International Affective Sound System Database IADS-2.

7. The method of the asynchronous EEG-based depression population determination system according to claim 5, characterized in that: In step S1, high-precision timing calibration of EEG acquisition points is performed using an asynchronous brain-like chip.

Images