An asynchronous electroencephalogram signal acquisition system and method

Abstract

The application discloses an asynchronous electroencephalogram signal acquisition system and method, and belongs to the technical field of biomedical engineering and neuroscience, and comprises: a front-end asynchronous trigger electroencephalogram signal acquisition circuit, which is used for realizing asynchronous acquisition of electroencephalogram signals and improving the quality of the electroencephalogram signals, so as to facilitate subsequent analysis and processing; and an embedded parallel logic control module, which is used for realizing signal processing and data flow management; an electroencephalogram asynchronous physiological calculation chip is arranged in the embedded parallel logic control module, and is used for improving the speed and accuracy of data processing and physiological calculation of emotion, and realizing emotion classification. The asynchronous electroencephalogram signal acquisition system and method can efficiently acquire and process electroencephalogram signals, so as to analyze emotional states.

Description

Technical Field

[0001] This invention relates to the fields of biomedical engineering and neuroscience, and in particular to an asynchronous electroencephalogram (EEG) signal acquisition system and method. Background Technology

[0002] Electroencephalography (EEG) is a technique that records changes in electrical potentials caused by the activity of neurons in the brain by placing electrodes on the scalp. It is a key tool for monitoring and analyzing the functional state of the brain. The signals captured by EEG technology can reflect the brain's activity under different emotional and cognitive states, with monitoring of emotional function being particularly important. Through detailed analysis of changes in EEG signals, brain electrical patterns associated with specific emotional states can be identified, thus providing a scientific basis for the diagnosis and assessment of emotional dysfunction. Changes in emotional states are often accompanied by changes in the intensity of activity in specific brain regions, which are manifested in EEG as fluctuations in different frequencies (such as alpha waves, beta waves, and theta waves) and amplitudes.

[0003] However, most current EEG signal systems employ a simultaneous acquisition method. Simultaneous acquisition refers to collecting data from multiple EEG electrodes at a fixed frequency under the same clock control. This method has significant limitations: it cannot flexibly respond to rapid changes in brain activity and typically generates a large amount of redundant data. This not only increases the burden of data storage and processing but also leads to missed crucial diagnostic information during analysis. Secondly, continuous use of high sampling rates significantly increases system energy consumption, which is detrimental to portable or long-term monitoring devices.

[0004] In contrast, asynchronous acquisition technology does not perform sampling at fixed time intervals, but rather triggers sampling based on actual changes in EEG signals. The core advantage of this method lies in its ability to automatically adjust sampling timing based on the activity and importance of the data source, thereby significantly reducing the storage and processing of invalid data while improving the relevance and accuracy of data acquisition. Furthermore, this strategy greatly improves the system's energy efficiency, enabling it to maintain low-power operation during long-term monitoring.

[0005] In the development of EEG technology, traditional EEG signal processing systems often use microprocessors or microcontrollers as the core of data processing. These systems rely on a central processing unit (CPU) and process data sequentially, facing major problems including slow processing speed, limited parallel processing capabilities, and relatively high power consumption. These limitations are particularly pronounced in EEG monitoring applications that require processing large amounts of data streams, especially during multi-channel data acquisition, where the CPU may struggle to process data from all channels in real time, leading to latency or data loss. In contrast, Field-Programmable Gate Arrays (FPGAs), as highly flexible digital logic devices, offer several significant advantages. First, they provide high parallel processing capabilities; FPGAs can be configured to process multiple data streams simultaneously, for example, processing multiple EEG signals from different brain regions in parallel. This characteristic makes FPGAs particularly suitable for multi-lead EEG signal devices, effectively improving data processing speed and accuracy. Second, they feature low-power design; FPGAs are designed to allow unused logic blocks to be turned off during inactivity, significantly reducing overall power consumption. This feature is particularly crucial for mobile or portable medical devices that require long-term battery power, ensuring continuous operation without frequent power replacements. Then there's programmability and customizability: unlike traditional hard-coded logic chips, FPGAs allow for on-demand configuration of hardware logic, providing immense flexibility for EEG signal processing. Developers can quickly reconfigure FPGAs based on specific monitoring needs or algorithm updates, adapting to new application scenarios without replacing hardware. Rapid prototyping is also possible; in the early stages of physiological computing chip development, FPGAs can quickly implement and test new processing algorithms and chip designs. This flexibility and cost-effectiveness make FPGAs play a vital role in the early development of physiological computing chips.

[0006] In the fields of EEG signal processing and emotional state decoding, wearable devices are becoming increasingly important due to their continuous monitoring capabilities and comfort. However, traditional wearable devices suffer from several technical challenges, particularly in data processing and power management. Currently, most wearable devices collect EEG signals, digitize the data, and transmit it to the cloud for processing. While this approach leverages powerful cloud computing resources, it also presents several key issues: First, continuous data transmission significantly increases device power consumption. Especially when the device relies on batteries or energy harvesters, frequent data transmissions accelerate energy consumption and limit usage time. Second, after sensors collect data, it needs to be transmitted to the cloud for processing, and the results are then sent back to the device. This round-trip latency can lead to unacceptable delays in applications requiring immediate response, impacting user experience and device usability. In contrast, dedicated physiological computing chips offer high integration, low power consumption, and sufficient computing power, enabling local processing and decision-making within the sensing nodes. This is more suitable for wearable devices in terms of data transmission volume, system latency, and power consumption.

[0007] In the diagnosis and analysis of EEG-based emotional dysfunction, traditional algorithms mainly include Support Vector Machines (SVM), Artificial Neural Networks (ANN), and deep learning models such as Convolutional Neural Networks (CNN). These methods have achieved some success in emotional state analysis, but they typically require significant computational resources and often rely on high-performance CPUs and GPUs. On portable devices, the implementation and application of these algorithms may be limited by both power consumption and processing power. Compared to traditional algorithms, spiking neural network (SNN)-based algorithms offer a processing approach closer to that of biological nervous systems. SNNs work by simulating the spiking activity of neurons to process information, a method particularly suitable for processing EEG signals. The main advantages of SNNs include: firstly, low power consumption, as SNNs can operate with very low energy consumption because they only consume energy when neurons are activated, making them particularly suitable for power-constrained mobile and portable devices; secondly, high timeliness, as their event-driven nature makes SNNs suitable for real-time or near-real-time data processing, enabling them to quickly respond to emotional changes in EEG signals; and thirdly, hardware compatibility, as SNNs can be implemented directly in hardware, with further improvements in processing speed and efficiency achieved by designing dedicated chips for neuromorphic physiological computing. Summary of the Invention

[0008] The purpose of this invention is to provide an asynchronous electroencephalogram (EEG) signal acquisition system and method that can efficiently acquire and process EEG signals for emotional state analysis.

[0009] To achieve the above objectives, the present invention provides an asynchronous electroencephalogram (EEG) signal acquisition system, comprising:

[0010] The front-end asynchronously triggered EEG signal acquisition circuit is used to realize asynchronous acquisition of EEG signals and improve the quality of EEG signals for subsequent analysis and processing.

[0011] An embedded parallel logic control module is used to implement signal processing and data flow management; the embedded parallel logic control module is equipped with an asynchronous EEG physiological computing chip to improve the speed and accuracy of data processing and the physiological calculation of emotions, thereby realizing emotion classification.

[0012] Preferably, the EEG signal acquisition circuit includes EEG signal sensing electrodes, an EEG signal preprocessing circuit, and an asynchronous ADC circuit; the EEG signal sensing electrodes are flexible dry electrodes made of graphene / liquid metal to improve signal reception quality and subject comfort; the EEG signal preprocessing circuit includes a low-noise amplifier, a low-pass filter, a 50Hz power frequency notch filter, and a band-pass filter connected in sequence; the asynchronous ADC circuit includes a digital-to-analog converter, a multiplexer, a comparator, digital logic control, a counter, and digital logic gates, connected according to signal logic control.

[0013] Preferably, the embedded parallel logic control module uses an FPGA as its core and has built-in configuration and logic control for the asynchronous ADC circuit.

[0014] Preferably, the asynchronous EEG physiological computing chip includes a data cache module, a pulse coding module, and a spiking neural network classifier.

[0015] Preferably, the data caching module adopts a direct mapping caching architecture and a high-speed static random access memory; the pulse coding module uses a rate-based coding strategy combined with dynamically adjusted parameters; the spiking neural network classifier includes an input layer, a linear transformation layer, and a Leaky Integrate-and-Fire neuron layer, implementing a learning and optimization strategy.

[0016] This invention also provides an asynchronous electroencephalogram (EEG) signal acquisition method, comprising the following steps:

[0017] S1. Collect and analyze electroencephalogram (EEG) signals;

[0018] S2. After the processed EEG data is input, it is encoded by the pulse coding module to generate pulses;

[0019] S3. Input the generated pulses into a spiking neural network classifier for analysis and processing, and finally output the emotion classification category.

[0020] Preferably, in S2, the pulse coding module introduces the parameter para. th For parameter parath Optimization is performed by dynamically adjusting the pulse threshold of each EEG processing channel. The pulse threshold of each EEG processing channel is calculated using the following formula:

[0021]

[0022] Where N is the sample size, l is the mean of the rate of change of the signal in the sample, r is the standard deviation of the rate of change of the signal in the sample, and V th (k) is the variable threshold of the k-th input channel;

[0023] On each EEG processing channel, a positive pulse is generated when the rate of change of the signal exceeds the corresponding variable threshold; a negative pulse is generated when the rate of change of the signal exceeds the variable threshold in the negative direction.

[0024] Preferably, the steps for optimizing the parameters are as follows: First, calculate the statistical descriptive features from the pulse sequence of each EEG processing channel. After feature selection, calculate the median pulse emission rate and the median time difference between pulses. Second, evaluate the feature performance under different settings using an SVM classifier. Use a Bayesian optimization method to find the optimal value, evaluate the performance of the feature set, and adjust it based on the obtained classification accuracy. Use leave-one-out cross-validation to ensure the generalization ability of the evaluation process, i.e., leave one sample as the test set each time, and use all other samples for training.

[0025] Preferably, in S3, after the generated pulses are input into the input layer of the spiking neural network classifier, the input data is mapped to a 128-dimensional hidden space in the first linear transformation layer for linear transformation. The output data is then processed by the Leaky Integrate and Fire neuron layer to simulate the dynamic behavior of biological neurons, including the integration and leakage of membrane potential, and the firing process when a threshold is reached. The output data is then input into the second linear transformation layer, which maps the 128-dimensional features to a dimension corresponding to the number of output categories. Finally, the output data is input into the second Leaky Integrate and Fire neuron layer to further integrate the features and directly use them for the classification task. The entire network is trained by minimizing the cross-entropy loss and the classification performance is optimized by updating the weights. Finally, the emotion classification category is output.

[0026] Therefore, the beneficial effects of the asynchronous EEG signal acquisition system and method described above in this invention are as follows:

[0027] (1) It can improve signal reception quality and subject comfort: By using flexible dry electrodes, the present invention not only enhances signal reception but also improves wearing comfort, making it suitable for long-term monitoring of EEG signals.

[0028] (2) Optimizes data acquisition efficiency and storage: By combining precise signal preprocessing and level-crossing asynchronous ADC, this invention ensures data accuracy and reliability. It samples only when the signal passes a preset level threshold, effectively reducing invalid data acquisition and decreasing data storage and transmission requirements. Furthermore, this mechanism provides higher temporal resolution when signals change rapidly, facilitating the accurate recording of rapidly changing events in EEG signals, such as spikes and fluctuations, thus optimizing the performance of portable EEG devices for long-term monitoring.

[0029] (3) Emotional disorder classification based on spiking neural network: The spiking neural network classification method used in this invention can be directly deployed on hardware, especially on dedicated chips designed for small portable devices, effectively supporting efficient and accurate emotional state classification.

[0030] (4) It possesses technical advantages in asynchronous acquisition and physiological computation: The asynchronous acquisition mechanism allows for the capture of asynchronous signals from different brain regions, improving the accuracy of monitoring activity in different brain regions and contributing to in-depth neuroscience research. Its event-triggered asynchronous mechanism responds to specific neurophysiological signals in real time, accurately recording the temporal relationships of neural activity. Furthermore, because it does not rely on a synchronous clock, the asynchronous ADC and dedicated chip reduce power consumption, extend device battery life, and improve data processing efficiency and system anti-interference capabilities, enhancing overall system stability and reliability.

[0031] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the composition of an asynchronous electroencephalogram (EEG) signal acquisition system according to the present invention;

[0033] Figure 2 This is a schematic diagram of the electroencephalogram (EEG) signal preprocessing circuit of an asynchronous EEG signal acquisition system according to the present invention;

[0034] Figure 3 This is a schematic diagram of the asynchronous ADC circuit of an asynchronous EEG signal acquisition system according to the present invention;

[0035] Figure 4 This is a flowchart illustrating an asynchronous electroencephalogram (EEG) signal acquisition method according to the present invention.

[0036] Figure 5 This is the parameter para of the present invention. th The optimization strategy diagram. Detailed Implementation

[0037] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0038] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0039] Example 1

[0040] like Figure 1 As shown, this embodiment provides an asynchronous electroencephalogram (EEG) signal acquisition system, including:

[0041] The front-end asynchronously triggered EEG signal acquisition circuit is used to achieve asynchronous acquisition of EEG signals and improve the quality of EEG signals for subsequent analysis and processing.

[0042] An embedded parallel logic control module is used for signal processing and data flow management. This module incorporates an asynchronous EEG physiological computing chip to improve the speed and accuracy of data processing and the physiological computation of emotions, enabling emotion classification.

[0043] The EEG signal acquisition circuit includes EEG signal sensing electrodes, an EEG signal preprocessing circuit, and an asynchronous ADC circuit. The EEG signal sensing electrodes are flexible dry electrodes made of graphene / liquid metal to improve signal reception quality and subject comfort.

[0044] like Figure 2As shown, the EEG signal preprocessing circuit includes a low-noise amplifier, a low-pass filter, a 50Hz power frequency notch filter, and a band-pass filter connected in sequence. The main purpose of this circuit is to improve the quality of the EEG signal for subsequent analysis and processing. The circuit first receives the raw EEG signal, which contains electrical activity information collected from the human brain. The weak raw EEG signal is amplified by a gain of 10 to 50 times through the low-noise amplifier. The amplified signal first passes through the low-pass filter to remove high-frequency noise, retaining the frequency components more critical for subsequent analysis. Subsequently, the signal passes through the 50Hz power frequency filter, specifically designed to eliminate interference from the power line. This step is particularly critical when using EEG equipment in industrial environments, as power frequency noise is a common source of interference. Finally, the signal enters the band-pass filter. This step sets a specific frequency range, allowing only theta waves (4-7Hz) and alpha waves (8-12Hz) to pass through. These two frequency ranges of EEG signals are very important when analyzing brain activity because they are associated with different brain functional states.

[0045] like Figure 3 As shown, the asynchronous ADC circuit is mainly a Level-Crossing ADC circuit designed based on the asynchronous sampling principle. It includes a digital-to-analog converter, a multiplexer, a comparator, digital logic control, a counter, and digital logic gates, which are connected according to signal logic control.

[0046] The sampling time of this asynchronous ADC circuit is triggered by changes in the input signal itself, rather than by an external clock. Specifically, the sampling frequency is dynamically adjusted by monitoring the crossover between the analog signal and a preset voltage threshold, thereby achieving efficient and accurate analog-to-digital conversion. The main components of the circuit and their functions are as follows: 1-bit DAC: This DAC is used to set the high (V) H ) and low (V L The voltage threshold of the ADC is adjusted according to the feedback control signal. These thresholds critically determine the triggering timing of the ADC sampling process and are crucial for sensitive response to signal changes. Compared with traditional multi-bit DACs, the simplified 1-bit DAC significantly reduces power consumption and system complexity, making it particularly suitable for power-constrained biomedical devices that require long-term operation. Multiplexer (MUX): The MUX is controlled by the logic output of the lower comparator (CMPD). The CMPD converts the V output of the DAC... ON Compared with the reference voltage, where V M For V H and V L The average value. Based on the comparison results, the MUX dynamically selects the output V. H or V L In conjunction with V ONThe signal is sent to the upper comparator CMPU. This ensures that the comparison window of the upper comparator is set correctly to capture signal level crossovers. Comparators (CMPD and CMPU): CMPD is responsible for monitoring the DAC output V. ON With reference voltage V M The relationship between the signal and the threshold crossing is monitored. The CMPU switches the comparison window based on the control signal output from the MUX. When the signal crosses the set threshold, it triggers data acquisition and CLC digital logic control. RLC and CLC logic: Based on the characteristics of EEG signals, two level crossing triggering methods are defined: continuous level crossing (CLC) and repetitive level crossing (RLC). The RLC logic controls the MUX to output rising or falling signals to the counter and generates an RLC cross pulse C. R The CLC logic controls the DAC logic, generating pulse C triggered by the CLC crossover. C These pulses form a change signal through a logic OR gate. CLC triggers an update of the entire system, while RLC triggers only refresh the counter. Up / Down Counter (U / DCounter): This counter records the number and direction of level crossing events; this data is used to generate a digital representation of the analog signal.

[0047] The embedded parallel logic control module uses an FPGA as its core and has built-in configuration and logic control for the asynchronous ADC circuit, enabling efficient signal processing and data flow management.

[0048] The embedded parallel logic control module incorporates an asynchronous EEG physiological computing chip developed using 40nm process technology. This chip enhances the speed and accuracy of data processing and enables physiological calculations of emotions, facilitating emotion classification. It includes a data cache module, a pulse coding module, and a spiking neural network classifier. The data cache module employs a 2MB direct-mapped cache architecture and high-speed static random-access memory (SRAM), supporting a 32-bit wide data interface. In the SRAM architecture, each read / write operation directly accesses a fixed storage location, resulting in faster data access and reduced processing latency.

[0049] The pulse coding module uses a rate-based coding strategy, combined with dynamically adjusted parameter para. th It effectively adjusts the pulse threshold of each channel to adapt to varying signal strengths, thereby improving the accuracy and efficiency of signal encoding.

[0050] The spiking neural network classifier integrates a multi-layer neural network, including an input layer, multiple linear transformation layers, and Leaky Integrate-and-Fire (LIF) neuron layers, to implement a learning and optimization strategy. At the network's output layer, this classifier determines the final emotion category by comparing pre-activation scores, achieving accurate emotion classification.

[0051] This system effectively acquires brain response signals under various external stimuli, such as motor intervention, neuroelectric stimulation, and music intervention. When external stimuli elicit neural activity responses in specific brain regions, the electrical signals from these regions are acquired through EEG signal sensing electrodes and transmitted to an EEG signal preprocessing circuit. This circuit amplifies and filters the weak raw neural electrical signals by 10 to 50 times. Multiple independent LC-ADCs in the asynchronous ADC circuit acquire the preprocessed EEG signals asynchronously in parallel. Each LC-ADC unit dynamically adjusts its sampling threshold and frequency based on the actual signal level changes, triggering sampling only when a critical level change occurs. This sampling mechanism significantly improves the time efficiency of data acquisition and reduces unnecessary data storage. The processed signals are transmitted in real-time via a Serial Peripheral Interface (SPI) to an embedded parallel logic control module constructed from an FPGA. The embedded parallel logic control module is responsible for synchronously receiving, buffering, and further analyzing these signals, including pulse coding, pulse feature extraction, and emotion classification using an SNN classifier. Finally, the emotion classification results are output to the host application software via a Universal Asynchronous Receive / Transmit (UART) interface. The application software is responsible for visualizing and displaying the emotion classification results so that users can intuitively understand and analyze the data.

[0052] like Figure 4 As shown, this embodiment also provides an asynchronous EEG signal acquisition method, including the following steps:

[0053] S1. Collect and analyze electroencephalogram (EEG) signals.

[0054] S2. After the processed EEG data is input, it is encoded by the pulse coding module to generate pulses. The pulse coding module introduces the parameter para. th For parameter para th Optimization is performed by dynamically adjusting the pulse threshold of each EEG processing channel. The pulse threshold of each EEG processing channel is calculated using the following formula:

[0055]

[0056] Where N is the sample size, l is the mean of the rate of change of the signal in the sample, r is the standard deviation of the rate of change of the signal in the sample, and V th (k) is the variable threshold of the k-th input channel.

[0057] On each EEG processing channel, a positive pulse is generated when the rate of change of the signal exceeds the corresponding variable threshold; a negative pulse is generated when the rate of change of the signal exceeds the variable threshold in the negative direction.

[0058] For parameter para thThe optimization steps are as follows Figure 5 As shown, the specific steps are as follows: First, statistical descriptive features are calculated from the pulse sequences of each EEG processing channel. After feature selection, the median pulse emission rate and the median time difference between pulses are calculated. Second, the feature performance under different settings is evaluated using an SVM classifier. The optimal value is found using a Bayesian optimization method, and the performance of the feature set is evaluated. The settings are adjusted based on the obtained classification accuracy. The leave-one-out cross-validation method is used to ensure the generalization ability of the evaluation process, that is, one sample is left as the test set each time, and all other samples are used for training.

[0059] S3. The generated pulses are input into a spiking neural network classifier (containing an input layer, multiple linear transformation layers, and Leaky Integrate-and-Fire (LIF) neuron layers) for analysis and processing, and finally the emotion classification category is output.

[0060] The generated pulses are input into the input layer of a spiking neural network classifier. In the first linear transformation layer, the input data is mapped to a 128-dimensional hidden space for linear transformation. The output data is then processed by a LeakyIntegrate and Fire neuron layer to simulate the dynamic behavior of biological neurons, including the integration and leakage of membrane potential, and the firing process when a threshold is reached. The output data is then input into a second linear transformation layer, which maps the 128-dimensional features to dimensions corresponding to the number of output categories. Finally, the output data is input into a second LeakyIntegrate and Fire neuron layer for further feature integration and direct application to the classification task. The entire network is trained by minimizing cross-entropy loss and its classification performance is optimized through weight updates, ultimately outputting an emotion classification category.

[0061] Therefore, the present invention employs the above-mentioned asynchronous EEG signal acquisition system and method, which can efficiently acquire and process EEG signals for emotional state analysis.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An asynchronous electroencephalographic signal acquisition system, characterized by, The application relates to an asynchronous brain electrical signal acquisition system. The application comprises the following: A front-end asynchronous trigger brain electrical signal acquisition circuit is used for realizing asynchronous acquisition of brain electrical signals and improving the quality of the brain electrical signals, so that subsequent analysis and processing are facilitated. An embedded parallel logic control module is used for realizing signal processing and data stream management. An embedded parallel logic control module is provided with a brain electrical asynchronous physiological calculation chip, which is used for improving the speed and accuracy of data processing and physiological calculation of emotions, and realizing emotion classification. The brain electrical asynchronous physiological computing chip comprises a data cache module, a pulse coding module and a pulse neural network classifier; the data cache module adopts a direct mapping cache architecture and a high-speed static random access memory, the pulse coding module uses a strategy based on rate coding, and combines a dynamic adjustment parameter ; the pulse neural network classifier comprises an input layer, a linear transformation layer and a Leaky Integrate-and-Fire neuron layer, and implements a learning and optimization strategy. The brain electrical signal acquisition circuit comprises a brain electrical signal sensing electrode, a brain electrical signal preprocessing circuit and an asynchronous ADC circuit; the brain electrical signal sensing electrode is a graphene / liquid metal flexible dry electrode, so that the signal receiving quality and the comfort of a subject are improved; the brain electrical signal preprocessing circuit comprises a low-noise amplifier, a low-pass filter, a 50Hz power frequency trap and a band-pass filter which are sequentially connected; the asynchronous ADC circuit comprises a digital-to-analog converter, a multiplexer, a comparator, digital logic control, a counter and a digital logic gate, and is connected according to signal logic control. The asynchronous brain electrical signal acquisition system performs the following steps: S1, collecting brain electrical signals and performing analysis and processing; S2, after the processed brain electrical data are input, the data are encoded by a pulse coding module to generate pulses; In S2, the pulse coding module introduces a parameter , which is optimized to dynamically adjust the pulse threshold for each EEG processing channel, . Parameters The step of optimizing specifically comprises: firstly, calculating statistical description features from the pulse sequence of each EEG processing channel, calculating the pulse emission rate and the median of the inter-pulse time difference after feature selection; secondly, evaluating the performance of the features under different settings by using the SVM classifier, using the Bayesian optimization method to find the optimal value, evaluating the performance of the feature set, and adjusting according to the obtained classification accuracy; using the leave-one-out cross-validation method to ensure the generalization ability of the evaluation process, that is, leaving one sample as the test set each time, and using all the remaining samples for training; S3, the generated pulses are input into a pulse neural network classifier to be analyzed and processed, and finally emotion classification categories are output; ； in, N It is the sample size. l It is the mean of the rate of change of the signal in the sample. r It is the standard deviation of the rate of change of the signal in the sample. V th ( k ) is the first k Input channel variable threshold; The pulse threshold of each EEG processing channel is calculated by the following formula: On each EEG processing channel, when the signal change rate exceeds the corresponding variable threshold, a positive pulse is generated; when the signal change rate exceeds the variable threshold in the negative direction, a negative pulse is generated; 2. The asynchronous electroencephalogram signal acquisition system of claim 1, wherein: In S3, after the generated pulses are input by the input layer of the pulse neural network classifier, in the first linear transformation layer, the input data are mapped to a 128-dimensional hidden space, linear transformation of the data is performed, the output data are processed by a LeakyIntegrate and Fire neuron layer, the dynamic behavior of a biological neuron is simulated, including integration and leakage of membrane potential and the discharge process when the threshold is reached; the output data are input again into the second linear transformation layer, the 128-dimensional features are mapped to a dimension corresponding to the number of output categories, finally, the output data are input again into the second Leaky Integrate and Fire neuron layer, the features are further integrated and directly used for a classification task; the whole network is trained by minimizing cross-entropy loss, and the classification performance is optimized through weight updating, and finally emotion classification categories are output. The embedded parallel logic control module uses FPGA as the core and is internally provided with configuration and logic control of the asynchronous ADC circuit.

Images