Sensory nerve quantitative detection method and system based on electroencephalogram characteristics

Abstract

The invention discloses a sensory nerve quantitative detection method and system based on electroencephalogram characteristics, and relates to the technical field of electroencephalogram measurement, and the method comprises the steps: building detection files in one-to-one correspondence with users; respectively wearing the current stimulation module and the electroencephalogram detection module on the wrist and the head of the user; current stimulation is applied to a user step by step through a current stimulation module, and gamma waves in electroencephalogram signals of the user are collected in real time through an electroencephalogram detection module; gamma waves in the electroencephalogram signals of the user are analyzed, and whether the intensity of current applied to the user by the current stimulation module exceeds a physical stimulation sensing threshold value of the subject or not is judged; and analyzing severity grades of sensory disorders. The method has the advantages that detection is carried out through brain wave signals without relying on subjective feedback of a patient, interference of factors such as emotion and psychology is avoided, and even a special patient who cannot carry out self-perception report can complete sensory nerve quantitative detection through objective physiological reaction of brain waves.

Description

Technical Field

[0001] The present invention relates to the technical field of electroencephalogram measurement, and particularly to a method and system for quantitatively detecting sensory nerves based on electroencephalogram features. Background Art

[0002] Quantitative Sensory Testing (QST) is a method of measuring the thresholds of a subject's light touch (or pressure sense), vibration sense, temperature sense (cold sense, warm sense), and pain sense (cold pain sense and heat pain sense) using physical stimuli, and evaluating the sensory functions of the subject's peripheral nerves and central nervous system. It is mainly used to detect abnormalities in sensory functions such as pain, touch, temperature sense, and vibration sense. Conventional quantitative sensory testing methods mainly rely on standardized stimuli and the subjective responses of patients, and are used for diagnosing peripheral neuropathy, central nervous system diseases, and pain management. This method has insurmountable defects. First, conventional methods cannot accurately quantify the physical stimuli received by patients. The subjective reports of patients are not precise and objective enough, and are easily affected by factors such as emotions, psychological states, and pain tolerance. When evaluating nerve function, they cannot accurately reflect the severity of sensory impairment. Second, the detection process is time-consuming. Medical staff need to manually adjust the current intensity level by level, and this process usually requires multiple attempts to determine the patient's sensory threshold. For each level of adjustment, medical staff need to observe the patient's reaction and record data, and the process is relatively cumbersome. Finally, conventional quantitative sensory testing cannot be performed on special patients who may be unable to report self-perception due to disorders of consciousness, language disorders, or being too young.

[0003] Current Perception Threshold (CPT) measurement is a method of evaluating sensory nerve function by electrical stimulation that has gradually become mainstream in recent years. Electrical signals of different frequencies can activate different types of nerve fibers, thereby evaluating the functional states of different nerve fibers. CPT is commonly used to evaluate the conduction ability of sensory nerves and can help in the early identification of nerve damage, especially small fiber nerve damage, which is crucial for the early intervention of chronic diseases such as diabetes or drug-induced nerve damage. This detection can also evaluate an individual's pain sensitivity, especially when treating pain diseases (such as fibromyalgia, chronic pain syndrome, etc.), helping doctors quantify changes in pain perception. When performing current perception threshold measurement, external physical stimuli not only activate the peripheral nervous system but also affect the cortical activity of the brain through nerve conduction. The brain's response to these stimuli often manifests as changes in the electroencephalogram (EEG). These electroencephalogram changes provide additional accessible information for nerve quantitative detection, helping to more comprehensively understand the response state of the nervous system.

[0004] Conventional quantitative detection methods for sensory nerves cannot accurately quantify the pain stimuli received by patients. The subjective reports of patients are not precise and objective enough, and are easily affected by factors such as emotions, psychological states, and pain tolerance, and cannot accurately reflect the degree of nerve damage when evaluating nerve function. Secondly, the detection process is time-consuming. Medical staff need to manually adjust the current intensity step by step. This process usually requires multiple attempts to determine the patient's sensory threshold. For each level of adjustment, medical staff need to observe the patient's reaction and record data, and the process is rather cumbersome. Finally, conventional quantitative detection of sensory nerves faces special patient groups. For example, patients may be unable to make self-perception reports due to disorders of consciousness, language disorders, or being too young, resulting in the inability of traditional methods to accurately evaluate their nerve function. Summary of the Invention

[0005] To solve the above technical problems, a sensory nerve quantitative detection method and system based on electroencephalogram (EEG) features are provided. The present invention proposes a low-granularity non-linear automatic current intensity adjustment strategy. By non-linearly adjusting the current intensity and using the degree of EEG wave activity as the termination basis, the EEG signals under current stimuli of different intensities can be quickly measured, and the EEG signal acquisition process of patients can be quickly and automatically completed.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A sensory nerve quantitative detection method based on EEG features, comprising:

[0008] Establish a detection file corresponding to each user. The diagnostic file at least includes: user name, gender, age, contact information, medical history record, detection date, and detection time;

[0009] Wear a current stimulation module and an EEG detection module on the user's wrist and head respectively;

[0010] Apply current stimuli to the user step by step through the current stimulation module, and collect the gamma waves in the user's EEG signals in real time through the EEG detection module;

[0011] Analyze the gamma waves in the user's EEG signals to determine whether the current intensity applied to the user by the current stimulation module exceeds the physical stimulation perception threshold of the subject. If so, terminate the detection and output the user's EEG data. If not, continue the detection;

[0012] Process the user's EEG data to obtain high-quality EEG signals in the target frequency band, and analyze the severity grading of sensory disorders based on the high-quality EEG signals.

[0013] As a preference of this solution, the step of applying current stimuli to the user step by step through the current stimulation module specifically includes:

[0014] Set the initial current intensity, and set the current intensity adjustment interval to 0.05 seconds;

[0015] Adjust the current intensity based on the low-granularity non-linear automatic adjustment current intensity strategy. The specific steps for adjusting the current intensity are as follows:

[0016] Set the initial current increase change amplitude, denoted as ΔI max ;

[0017] The subsequent current increase change amplitude for each step is:

[0018] ΔI(t) = max(ΔI max *e -kt , 0.01)

[0019] where ΔI(t) is the current change amplitude at time step t, k is the decreasing rate, and t is the time step.

[0020] As an optimization of this solution, the specific steps for analyzing the gamma waves in the user's EEG signal and determining whether the current intensity applied by the current stimulation module to the user exceeds the physical stimulation perception threshold of the subject are as follows:

[0021] Taking ten cycles as the grouping standard, collect the amplitude data of ten consecutive cycles of Gamma waves and form an amplitude array;

[0022] Calculate the ratio of the sum of the amplitudes in two adjacent amplitude arrays and determine whether the ratio is greater than the amplitude change threshold. If so, it is determined that the current intensity has exceeded the physical stimulation perception threshold of the subject, and the experiment is terminated. If not, no response is made.

[0023] Furthermore, a sensory nerve quantitative detection system based on EEG characteristics is also proposed to implement the above-mentioned sensory nerve quantitative detection method based on EEG characteristics, including: a user detection file construction module, a current stimulation module, an EEG signal acquisition module, a stimulation threshold determination module, an EEG signal processing module, and a nerve quantitative analysis module, where:

[0024] The user detection file construction module is used to create a digital file containing the user's name, gender, age, contact information, medical history record, detection date, and detection time;

[0025] The current stimulation module includes a wearable electrode array configured to apply a stepped increasing current stimulation to the subject according to a preset non-linear adjustment strategy;

[0026] The EEG signal acquisition module integrates a high-precision Gamma wave sensor array to capture the electrophysiological signals of the subject's head in real time and extract the time-frequency characteristics of the Gamma band;

[0027] The stimulation threshold determination module dynamically determines whether the physical stimulation perception threshold is exceeded by analyzing the change rate of the Gamma wave amplitude in consecutive cycles, and triggers a detection termination instruction;

[0028] The EEG signal processing module processes the EEG data of the user to obtain high-quality EEG signals in the target frequency band, and analyzes the severity grading of sensory disorders based on the high-quality EEG signals.

[0029] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0030] The present invention provides a method and system for quantitative detection of sensory nerves based on EEG features. The present invention proposes a low-granularity non-linear automatic adjustment current intensity strategy. By non-linearly adjusting the current intensity, rapid acquisition is achieved without loss of accuracy in the sensitive current intensity interval, and the activity level of brain waves is used as the basis for terminating the measurement. The brain wave signals and corresponding intensity information under current stimulation of different intensities are collected for subsequent quantitative detection and classification. This solution detects through electroencephalogram (EEG) signals, does not rely on the subjective feedback of patients, and avoids the interference of factors such as emotions and psychology. Even for special patients who are unable to report self-perception, quantitative detection of sensory nerves can be completed through the objective physiological response of brain waves. Description of the Drawings

[0031] Figure 1 It is a flow chart of a method for quantitative detection of sensory nerves based on EEG features proposed in Embodiment 1;

[0032] Figure 2 It is a flow chart of a method for optimizing and implementing the setting of the initial current intensity proposed in Embodiment 2. Detailed Embodiments

[0033] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. In the following description, Embodiment 1:

[0034] Refer to Figure 1 As shown, a method for quantitative detection of sensory nerves based on EEG features includes:

[0035] Establish a detection file corresponding to the user one by one. The diagnosis file includes at least: user name, gender, age, contact information, medical history record, detection date, and detection time;

[0036] Wear the current stimulation module and the EEG detection module on the user's wrist and head respectively;

[0037] Apply current stimulation to the user step by step through the current stimulation module, and collect the gamma waves in the user's EEG signals in real time through the EEG detection module;

[0038] In this step, the specific implementation means are as follows: set the initial current intensity to 0.01 mA, and set the current intensity adjustment interval to 0.05 seconds;

[0039] Adjust the current intensity based on the low-granularity non-linear automatic current intensity adjustment strategy. The specific current intensity adjustment steps are as follows:

[0040] Set the initial current increase change amplitude, denoted as ΔI max , ΔI max According to the analysis of experimental data, take half of the minimum value between the minimum current intensity of electrical stimulation perception in historical test data and the initial current intensity;

[0041] The subsequent current increase change amplitude for each step is:

[0042] ΔI(t) = max(ΔI max *e -kt , 0.01)

[0043] where ΔI(t) is the current change amplitude at time step t, k is the decreasing rate, and t is the time step;

[0044] Analyze the gamma waves in the user's EEG signal to determine whether the current intensity applied by the current stimulation module to the user exceeds the physical stimulation perception threshold of the subject. If so, terminate the detection and output the user's EEG data. If not, continue the detection;

[0045] Specifically, in this step, taking ten cycles as the grouping standard, collect the amplitude data of ten consecutive cycles of gamma waves and form an amplitude array;

[0046] Calculate the ratio of the sum of amplitudes within two adjacent amplitude arrays and determine whether the ratio is greater than the amplitude change threshold. If so, it is determined that the current intensity has exceeded the physical stimulation perception threshold of the subject, and the experiment is terminated. If not, no response is made and the detection continues;

[0047] It can be understood that the gamma waves in the EEG signal are highly correlated with the pain perception of the human body. When the human body is subjected to pain stimulation, the amplitude of the gamma waves increases significantly. Therefore, the present invention uses the amplitude change of gamma waves as the basis for terminating the experiment: where FList is the amplitude array of ten consecutive cycles of gamma waves, FList i is the i-th amplitude array, Flist i-1Let \(A_{i - 1}\) be the \((i - 1)\)-th amplitude array, and \(C\) be the threshold of the amplitude change of the gamma wave. When the ratio of the sum of two adjacent amplitude arrays of the gamma wave is greater than \(C\), it is determined that the current intensity has exceeded the physical stimulus perception threshold of the subject, and the experiment is terminated. The determination of the amplitude change threshold requires collecting data through an induction experiment, analyzing the data time curve after normalization, and finding the normalized value corresponding to the point with the maximum curve slope.

[0048] Process the electroencephalogram data of the user to obtain high-quality electroencephalogram signals in the target frequency band, and analyze the severity level of sensory impairment based on the high-quality electroencephalogram signals.

[0049] Embodiment 2:

[0050] Refer to Figure 2 As shown, on the basis of Embodiment 1, this embodiment further gives an optimized implementation scheme for setting the initial current intensity, which specifically includes:

[0051] Classify the historical test data based on gender, age, and medical history records respectively to obtain several data sets;

[0052] Based on the collected detection file of the user, determine the corresponding data set of the user, denoted as the corresponding data set;

[0053] Determine the current intensity corresponding to the minimum current intensity of electrostimulation perception of each historical test data in the corresponding data set, denoted as the sample current intensity;

[0054] Calculate the average value and standard deviation of all sample current intensities corresponding to the corresponding data set, and use the value obtained by subtracting 6 times the standard deviation from the average value as the initial current intensity;

[0055] It can be understood that in the initial stage of the system operation, due to the lack of sufficient detection data for sample analysis, at this time, to ensure the full coverage during detection, the initial current intensity needs to be as small as possible to reduce the missed detection rate. Therefore, in Embodiment 1, the initial current intensity is given as 0.01 mA. However, due to the differences among different test individuals, it may lead to the problem of a longer test duration for users with a high physical stimulus perception threshold. Therefore, based on the system accumulating sufficient detection data as samples, this scheme proposes a more optimized scheme for determining the initial current intensity. By classifying the characteristics of the detection objects corresponding to the historical detection data, for example, using age group + gender + whether suffering from rheumatic diseases as the classification standard, classifying the historical data according to this standard, and selecting a more similar group based on the user's situation during the measurement test, and determining the most optimized initial current intensity based on the data in the group;

[0056] In this solution, the initial current intensity based on data analysis adopts the 6-sigma criterion, that is, the initial current intensity is determined by subtracting 6 standard deviations from the average value. Based on the normal distribution criterion, the probability that the minimum current intensity of the user's electrical stimulation perception is above this initial current intensity is 99.99%. Based on the Chebyshev lower bound calculation, the probability that the minimum current intensity of the user's electrical stimulation perception is above this initial current intensity is 98.61%. By using this initial current intensity, the applied current intensity can reach the minimum current intensity of the user's electrical stimulation perception fastest, thereby effectively shortening the test duration and improving the detection efficiency.

[0057] Example 3:

[0058] In this example, based on Example 1, it is further pointed out that: by processing the user's electroencephalogram data, high-quality electroencephalogram signals in the target frequency band are obtained, and analyzing the severity level of sensory impairment based on the high-quality electroencephalogram signals specifically includes:

[0059] Construct a sensory nerve quantitative detection model, which takes the electroencephalogram signals collected at different time steps as input and outputs the probability that the electroencephalogram signal at this time step belongs to when the electrical stimulation is perceived;

[0060] The sensory nerve quantitative detection model is specifically:

[0061] The temporal convolutional network module is composed of three causal convolutional layers with filter sizes of 128, and 128 respectively. Each layer integrates the dilated convolution mechanism and residual connections to extract the temporal features of the nerve electrical signals;

[0062] The contrastive learning module pre-trains the features output by the temporal convolutional network through a contrastive loss function, and the contrastive loss is calculated based on the Euclidean distance of positive and negative sample pairs;

[0063] The classification module includes a fully connected layer and a Softmax classifier for downstream classification tasks;

[0064] The model training adopts a two-stage training method: in the pre-training stage, only the temporal convolutional network and the contrastive learning module are optimized, and in the downstream training stage, the parameters of the temporal convolutional network are frozen and the classification module is optimized;

[0065] Among them, in the temporal convolutional network module:

[0066] The causal convolution satisfies that the output y(t) satisfies where x(t) is the input sequence, w(i) is the convolution kernel, and k is the size of the convolution kernel, ensuring that the output y(t) of the causal convolution only depends on the current and historical inputs;

[0067] The dilated convolution function is: Among them, d is the dilation factor, and k is the size of the convolutional kernel. Dilated convolution expands the receptive field by introducing gaps in the convolutional kernel, enabling it to capture longer-term dependencies without increasing the computational complexity;

[0068] The residual connection directly passes the input to the output for addition, enhancing the stability during model training and the flow of gradients during backpropagation;

[0069] In the contrastive learning module:

[0070] The contrastive loss function adopts a formula based on the Euclidean distance, where y i indicates whether the sample pair is a positive sample pair or a negative sample pair, 1 represents a positive sample pair, 0 represents a negative sample pair, and D(x i , x j ) represents the Euclidean distance between the sample features output by the temporal convolutional network module, m is the preset margin distance, and L is the model loss value;

[0071] Among them, the positive sample pair is the electroencephalogram data when electrical stimulation is perceived under two different time windows, and the negative sample pair is the electroencephalogram data when electrical stimulation is perceived and not perceived;

[0072] Through contrastive learning for model pre-training, the goal is to learn high-quality feature representations rather than directly performing classification. The loss function of the model makes the feature distances of positive sample pairs as close as possible, while making the feature distances of negative sample pairs as far apart as possible. The high-quality feature embedding space learned by the model during pre-training can be used for downstream classification tasks. During the pre-training stage, the classification module does not participate in the training;

[0073] Classification module:

[0074] The loss function is where N is the number of samples, y i is the true label (0 or 1) of the i-th sample, p i is the predicted probability that the i-th sample is the true label 1, and L is the model loss value;

[0075] During the downstream task training stage, the pre-trained representations are used for the classification module. At this time, only the classification module of the model will be optimized. The features and labels of the training data are input into the classifier module, and the loss is calculated and the classifier is optimized.

[0076] The electroencephalogram signals collected by the user at different time steps are used as a batch and input into the sensory nerve quantitative detection model to obtain the probability that each time-step corresponding electroencephalogram signal belongs to the case when electrical stimulation is perceived. Then, it is judged whether the probability is greater than the probability threshold. If so, this segment of electroencephalogram signal is marked as "true", that is, the electrical stimulation corresponding to this segment of electroencephalogram signal has been perceived. If not, no response is made;

[0077] Among them, the set probability threshold is 0.6, which is used as the classification criterion. If the output value of the sensory nerve quantitative detection model is greater than 0.6, it is judged as "true", that is, the electrical stimulation corresponding to this segment of EEG signal has been perceived;

[0078] Among them, the collected EEG signals are subjected to data noise separation. The low-frequency drift and high-frequency noise are removed through a filtering algorithm, and the signals in a specific frequency band are extracted. Combined with independent component analysis (ICA), the artifacts such as electromyogram and eye movement are separated and removed. The filtering algorithms used include high-pass filtering, low-pass filtering, band-pass filtering, and notch filtering. The noise-separated EEG data is segmented according to a time window, and the high-quality EEG signals in the target frequency band are retained. The high-quality EEG signals in the target frequency band are input into the sensory nerve quantitative detection model;

[0079] Output the current intensity corresponding to all EEG signals marked as "true", select the current intensity when it is first "true" in the classification result as the minimum current intensity of electrical stimulation perception, and classify the severity of the sensory disorder of the experimental personnel according to the minimum current intensity of electrical stimulation perception.

[0080] Example 4:

[0081] Based on the sensory nerve quantitative detection method based on EEG features proposed in Examples 1 to 3, this example proposes a sensory nerve quantitative detection system based on EEG features, including a user detection file construction module, a current stimulation module, an EEG signal acquisition module, a stimulation threshold determination module, an EEG signal processing module, and a nerve quantitative analysis module, where:

[0082] The user detection file construction module is used to create a digital file containing the user's name, gender, age, contact information, medical history record, detection date, and detection time;

[0083] The current stimulation module includes a wearable electrode array, which is configured to apply a stepped increasing current stimulation to the subject according to a preset non-linear adjustment strategy. The current stimulation module can directly set the initial current intensity to 0.01 mA, or determine the initial current intensity through the initial current intensity determination method proposed in Example 2;

[0084] The EEG signal acquisition module integrates a high-precision Gamma wave sensor array to capture the electrophysiological signals of the subject's head in real time and extract the time-frequency features of the Gamma band;

[0085] The stimulation threshold determination module dynamically determines whether the physical stimulation perception threshold is exceeded by analyzing the change rate of the Gamma wave amplitude in consecutive cycles and triggers a detection termination instruction;

[0086] The electroencephalogram signal processing module processes based on the electroencephalogram data of the user to obtain high-quality electroencephalogram signals in the target frequency band, and analyzes the severity grading of sensory disorders based on the high-quality electroencephalogram signals:

[0087] Specifically, a sensory nerve quantitative detection model is integrated in the electroencephalogram signal processing module. The sensory nerve quantitative detection model includes a temporal convolutional network module, a contrast learning module, and a classification module. The model includes three temporal convolutional networks with filter sizes of 128, 256, and 128 respectively. The temporal convolutional network is based on convolutional operations and uses causal convolution, dilated convolution, and residual connection mechanisms to capture long-term dependencies in time series data. The main goal of causal convolution is to ensure that the convolutional operation only uses data at the current time point and its previous time points, and does not use future data. For causal convolution, the output y(t) only depends on the current and previous inputs: where x(t) is the input sequence, w(i) is the convolutional kernel, and k is the size of the convolutional kernel. Dilated convolution expands the receptive field by introducing gaps in the convolutional kernel, so that it can capture longer dependencies without increasing the computational cost. The dilated convolution function is: where d is the dilation factor and k is the size of the convolutional kernel. Residual connection directly passes the input to the output for addition, improving the stability during model training and the fluidity of the gradient during backpropagation. The contrast learning module uses contrastive loss to maximize the similarity between positive sample pairs and minimize the similarity between negative sample pairs. The contrastive loss function adopts a formula based on the Euclidean distance: where y i represents whether the sample pair is a positive sample pair (1 for positive and 0 for negative); D(x i , x j ) represents the Euclidean distance between the sample features output by the temporal convolutional network module, and m is a preset margin distance. Positive sample pairs are EEG data when electrical stimulation is perceived under two different time windows, and negative sample pairs are EEG data when electrical stimulation is perceived and not perceived. The classification module is composed of a fully connected layer and a Softmax layer, and the classification layer is trained using the cross-entropy loss function. The loss function is: where N is the number of samples, y i is the true label (0 or 1) of the i-th sample, and p iis the predicted probability that the i-th sample is the true label 1. The model training is divided into two stages: the pre-training stage and the downstream task training stage. The model is pre-trained through contrastive learning, and the goal is to learn high-quality feature representations rather than directly performing classification. The loss function of the model makes the feature distances of positive sample pairs as close as possible, while the feature distances of negative sample pairs are as far apart as possible. The high-quality feature embedding space learned by the model during pre-training can be used for downstream classification tasks. During the pre-training stage, the classification module does not participate in the training. The downstream task training stage is to use the pre-trained representation for the classification module. At this time, only the classification module of the model will be optimized. The features and labels of the training data are input into the classifier module, and the loss is calculated and the classifier is optimized;

[0088] The electroencephalogram signal processing module takes the electroencephalogram signals collected by the experimenter at different time steps as the same batch and inputs them into the above model. The model simultaneously classifies and predicts the electroencephalogram signals at all time steps. The value output by the model (ranging from 0 to 1) represents the probability that a certain segment of electroencephalogram signal belongs to the perception of an electrical stimulus. A threshold of 0.6 is set as the classification criterion. If the output value is greater than 0.6, it is judged as "true", that is, the electrical stimulus corresponding to this segment of electroencephalogram signal has been perceived. Record the current intensity corresponding to the electroencephalogram signal with the classification result of "true", and select the current intensity at the first time when the classification result is "true" as the minimum current intensity of electroencephalogram signal perception. The severity of sensory impairment of the experimenter is graded according to the minimum current intensity of electroencephalogram signal perception.

[0089] In summary, the advantages of the present invention are as follows: A low-granularity non-linear automatic current intensity adjustment strategy is proposed. By non-linearly adjusting the current intensity, rapid acquisition is achieved without losing the accuracy of the sensitive current intensity range, and the electroencephalogram wave activity level is used as the termination measurement basis to collect electroencephalogram signals and corresponding intensity information under different intensity current stimuli for subsequent quantitative detection and classification; The present invention proposes a sensory nerve quantitative detection model. Based on the relationship between external stimuli and electroencephalogram signals, the model performs sensory nerve quantitative detection on the electroencephalogram signal data characteristics under different current intensities. The model uses a temporal convolutional network and contrastive learning to capture the long-term dependence relationship of the input sequence, extracts electroencephalogram wave characteristics in different states, and uses a softmax classifier to detect and classify the extracted high-quality temporal characteristics, and finally completes the grading of the severity of sensory impairment of the experimenter.

[0090] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification is only the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection required by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for quantitatively detecting sensory nerves based on electroencephalogram features, characterized in that Including: Establish a detection file corresponding to each user, and the diagnosis file at least includes: user name, gender, age, contact information, medical history record, detection date and detection time; Wear the current stimulation module and the electroencephalogram detection module on the user's wrist and head respectively; Gradually apply current stimulation to the user through the current stimulation module, and collect gamma waves in the user's electroencephalogram signal in real time through the electroencephalogram detection module; Analyze the gamma waves in the user's electroencephalogram signal to judge whether the current intensity applied by the current stimulation module to the user exceeds the physical stimulation perception threshold of the subject. If so, terminate the detection and output the user's electroencephalogram data. If not, continue the detection; Process the user's electroencephalogram data to obtain high-quality electroencephalogram signals in the target frequency band, and analyze the severity classification of sensory disorders based on the high-quality electroencephalogram signals.

2. The sensory nerve quantitative detection method based on electroencephalogram features according to claim 1, characterized in that, The step of gradually applying current stimulation to the user through the current stimulation module specifically includes: Set the initial current intensity to 0.01 mA, and the current intensity adjustment interval to 0.05 seconds; Adjust the current intensity based on the low-granularity nonlinear automatic adjustment current intensity strategy. The specific current intensity adjustment steps are: Set the initial current increase variation range, denoted as ΔI max ; The subsequent increase in current for each step is: ΔI(t) = max(ΔI max *e -kt , 0.01) Where, ΔI(t) is the current change amplitude at time step t, k is the decreasing rate, and t is the time step.

3. The sensory nerve quantitative detection method based on EEG features according to claim 2, wherein The initial current intensity can also be determined by the following method: Classify the historical test data based on gender, age, and medical history record respectively to obtain several data sets; Based on the collected user's detection file, determine the data set corresponding to the user, denoted as the corresponding data set; Determine the current intensity corresponding to the minimum current intensity of electrostimulation perception of each historical test data in the corresponding data set, denoted as the sample current intensity; Calculate the average value and standard deviation of all sample current intensities corresponding to the corresponding data set, and use the value obtained by subtracting 6 times the standard deviation from the average value as the initial current intensity.

4. The sensory nerve quantitative detection method based on EEG features according to claim 1, wherein The specific steps of analyzing the gamma waves in the user's electroencephalogram signal to judge whether the current intensity applied by the current stimulation module to the user exceeds the physical stimulation perception threshold of the subject are: Taking ten cycles as the grouping standard, collect the amplitude data of ten consecutive cycles of Gamma waves and form an amplitude array; Calculate the ratio of the sum of amplitudes in two adjacent amplitude arrays and judge whether the ratio is greater than the amplitude change threshold. If so, it is determined that the current intensity has exceeded the physical stimulation perception threshold of the subject, and the experiment is terminated. If not, no response is made.

5. A method for quantitatively detecting sensory nerves based on EEG features according to claim 1, characterized in that, The analysis of the severity classification of sensory disorders based on the high-quality electroencephalogram signal specifically includes: Construct a quantitative sensory nerve detection model, which takes the electroencephalogram signals collected at different time steps as input and outputs the probability that the electroencephalogram signal at this time step belongs to the time when the electrical stimulation is perceived; The electroencephalogram (EEG) signals collected by the user at different time steps are used as a batch input into the sensory nerve quantitative detection model to obtain the probability that the EEG signal collected at each time step belongs to the state of perceiving an electrical stimulus. It is then determined whether the probability is greater than the probability threshold. If so, the segment of EEG signal is marked as "true", indicating that the electrical stimulus corresponding to this segment of EEG signal has been perceived. If not, no response is made. The current intensities corresponding to all the EEG signals marked as "true" are output. The current intensity at the first occurrence of "true" in the classification results is selected as the minimum current intensity for the perception of the electrical stimulus. Based on the minimum current intensity for the perception of the electrical stimulus, the severity of the sensory impairment of the experimental subject is graded.

6. The quantitative detection method of sensory nerves based on EEG features according to claim 5, characterized in that The sensory nerve quantitative detection model specifically includes: A temporal convolutional network module, which consists of three causal convolutional layers with filter sizes of 128, 256, and 128 respectively. Each layer integrates an extended convolutional mechanism and a residual connection to extract the temporal features of the nerve electrical signals. A contrastive learning module that pre-trains the features output by the temporal convolutional network through a contrastive loss function. The contrastive loss is calculated based on the Euclidean distance between positive and negative sample pairs. A classification module, which includes a fully connected layer and a Softmax classifier for downstream classification tasks. The model training adopts a two-stage training method: in the pre-training stage, only the temporal convolutional network and the contrastive learning module are optimized. In the downstream training stage, the parameters of the temporal convolutional network are frozen and the classification module is optimized.

7. A quantitative detection method for sensory nerves based on EEG features according to claim 6, characterized in that In the temporal convolutional network module: The causal convolution satisfies that the output y(t) satisfies where x(t) is the input sequence, w(i) is the convolution kernel, and k is the size of the convolution kernel, ensuring that the output y(t) of the causal convolution depends only on the current and historical inputs; The extended convolution function is as follows: where d is the dilation factor and k is the size of the convolution kernel; The residual connection directly passes the input to the output for addition, enhancing the stability during model training and the flow of gradients during backpropagation.

8. A method for quantitatively detecting sensory nerves based on EEG features according to claim 6, characterized in that, In the contrastive learning module: The contrastive loss function adopts a formula based on the Euclidean distance, where y i indicates whether the sample pair is a positive sample pair or a negative sample pair, 1 represents a positive sample pair, 0 represents a negative sample pair, and D(x i , x j ) represents the Euclidean distance between the sample features output by the temporal convolutional network module, m is a preset boundary distance, and L is the model loss value; Among them, the positive sample pairs are the electroencephalogram data when an electrical stimulus is perceived under two different time windows, and the negative sample pairs are the electroencephalogram data when an electrical stimulus is perceived and when it is not perceived.

9. The sensory nerve quantitative detection method based on EEG features according to claim 6, wherein In the classification module: The loss function is where N is the number of samples, y i is the true label (0 or 1) of the i-th sample, and p i is the predicted probability that the i-th sample is the true label 1, and L is the loss value of the model.

10. A sensory nerve quantitative detection system based on EEG features for implementing the sensory nerve quantitative detection method based on EEG features according to any one of claims 1-9, including a user detection file construction module, a current stimulation module, an EEG signal acquisition module, a stimulation threshold determination module, an EEG signal processing module, and a nerve quantitative analysis module, where: The user detection file construction module is used to create a digital file containing the user's name, gender, age, contact information, medical history record, detection date, and detection time. The current stimulation module includes a wearable electrode array configured to apply a stepped increasing current stimulation to the subject according to a preset non-linear adjustment strategy. The EEG signal acquisition module integrates a high-precision Gamma wave sensor array to capture the electrophysiological signals of the subject's head in real time and extract the time-frequency features in the Gamma band. The stimulation threshold determination module dynamically determines whether the physical stimulation perception threshold is exceeded by analyzing the change rate of the Gamma wave amplitude in consecutive periods and triggers a detection termination instruction. The EEG signal processing module processes the user's electroencephalogram data to obtain high-quality EEG signals in the target frequency band and analyzes the severity of the sensory impairment based on the high-quality EEG signals.

Images