Post-traumatic stress disorder identification device, method and storage medium

Abstract

The present application relates to a post-traumatic stress disorder identification device, method and storage medium, the device includes: a data acquisition module, a model application module and a classification recognition module; the data acquisition module is used to collect multiple physiological data of the target user under different stimulation strategies at multiple different stages, the multiple physiological data include the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtain multiple feature data of the target user based on the multiple physiological data; the model application module is used to input the multiple feature data into the trained lightgbm model to obtain the classification information output by the trained lightgbm model; the classification recognition module is used to identify whether the target user is a patient with post-traumatic stress disorder according to the classification information. The above device can automatically identify patients with post-traumatic stress disorder by means of a machine learning algorithm, thereby improving the accuracy of identifying patients with post-traumatic stress disorder.

Description

Technical Field

[0001] The present application relates to the technical field of computer data processing, and in particular to a post-traumatic stress disorder identification device, method and storage medium. Background Art

[0002] Post-traumatic stress disorder (PTSD) refers to a delayed and persistent mental disorder caused by a person experiencing, witnessing, or encountering a serious threat to life, serious injury, or threat to physical integrity. People with PTSD are prone to mental breakdown in mild cases, mental illness that affects physical and mental health in severe cases, and even more serious cases may cause self-harm. Therefore, it is necessary to identify the symptoms of PTSD early to avoid the harm caused by PTSD to people.

[0003] Currently, post-traumatic stress disorder is initially screened by self-filling questionnaires. However, when filling out the questionnaire, users selectively choose answers to questions due to social expectations or personal needs, which may result in missed screening or incorrect screening. Summary of the invention

[0004] Based on this, it is necessary to provide a post-traumatic stress disorder identification device, method and storage medium to address the above technical problems, which can automatically identify patients with post-traumatic stress disorder through machine learning algorithms and improve the accuracy of identifying patients with post-traumatic stress disorder.

[0005] A post-traumatic stress disorder identification device comprises a data acquisition module, a model application module and a classification identification module; the data acquisition module is used to collect multiple physiological data of a target user under different stimulation strategies at multiple different stages, the multiple physiological data comprising the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtain multiple feature data of the target user based on the multiple physiological data; the model application module is used to input the multiple feature data into a trained LightGBM model to obtain classification information output by the trained LightGBM model; the classification identification module is used to identify whether the target user is a patient with post-traumatic stress disorder according to the classification information.

[0006] In one embodiment, the data acquisition module includes: a first data acquisition unit, which is used to collect first physiological data under multiple different stimulation strategies in the first stage, the multiple different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes a probe stimulus but does not include an aversive stimulus; a second data acquisition unit, which is used to collect second physiological data under multiple different stimulation strategies in the second stage, the multiple different stimulation strategies in the second stage are a second conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, and the second conditioned stimulus sequence is different from the first conditioned stimulus sequence; a third data acquisition unit, which is used to collect third physiological data under multiple different stimulation strategies in the third stage, the multiple different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of multiple third conditioned stimuli and multiple fourth conditioned stimuli, the third conditioned stimulus and the fourth conditioned stimulus both include a probe stimulus and do not include an aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; wherein the multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

[0007] In one embodiment, the aversive stimulus is compressed air of a specific pressure given to the target user's neck, and the probe stimulus is a short popping sound given to the target user; the first conditioned stimulus is a first basic conditioned stimulus given to the target user and compressed air of a specific pressure given to the target user's neck, and a short popping sound given to the target user; and / or, the second conditioned stimulus is a second basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the third conditioned stimulus is a third basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the fourth conditioned stimulus is a fourth basic conditioned stimulus displayed to the target user and a short popping sound given to the target user; wherein, the first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus are different, and the first basic conditioned stimulus and the third basic conditioned stimulus are the same.

[0008] In one embodiment, the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the first conditioned stimulus sequence is different from the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the second conditioned stimulus sequence; and / or, the sum of the numbers of the first conditioned stimuli and the second conditioned stimuli in the first conditioned stimulus sequence or the second conditioned stimulus sequence is greater than the sum of the numbers of the third conditioned stimuli and the fourth conditioned stimuli in the third conditioned stimulus sequence.

[0009] In one embodiment, the data acquisition module also includes a feature extraction unit, which includes: a first feature extraction subunit, which is used to obtain physiological feature data of each stage based on the physiological data of each stage in the first stage, the second stage and the third stage; a second feature extraction subunit, which is used to obtain differential feature data of the same conditional stimulus based on the physiological feature data of each stage; a third feature extraction subunit, which is used to obtain physiological feature data of each trial based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage; a fourth feature extraction subunit, which is used to obtain physiological feature data corresponding to before and after probe stimulation and physiological feature data corresponding to before and after aversive stimulation based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage; the multiple feature data include physiological feature data of each stage, differential feature data of the same conditional stimulus, physiological feature data of each trial, and physiological feature data corresponding to before and after probe stimulation and physiological feature data corresponding to before and after aversive stimulation.

[0010] In one embodiment, the system further includes a model construction module and a model training module, the model construction module includes a first sample collection unit and a construction unit, and the model training module includes a second sample collection unit and a training unit; the first sample collection unit is used to collect a plurality of first sample physiological data of a first sample user under different stimulation strategies at a plurality of different stages, the plurality of first sample physiological data include the physiological data of electrooculogram and the physiological data of electrocardiogram of the first sample user, and obtain a plurality of first sample feature data of the first sample user based on the plurality of first sample physiological data; the construction unit is used to input the plurality of first sample feature data into an untrained LightGBM model, and adjust the untrained LightGBM model by adjusting the first sample physiological data. a target LightGBM model is constructed by the number of sub-evaluators of the LightGBM model and the value of feature contribution; a second sample collection unit is used to collect multiple second sample physiological data of multiple second sample users under different stimulation strategies at multiple different stages, the multiple second sample physiological data include the physiological data of electrooculogram and the physiological data of electrocardiogram of the second sample users, and obtain multiple second sample feature data of the second sample users based on the multiple second sample physiological data; a training unit is used to perform model training on the target LightGBM model according to the multiple second sample feature data and by a ten-fold cross validation method to obtain a trained LightGBM model.

[0011] In one of the embodiments, the model building module also includes: a first feature screening unit, used to calculate the Pearson correlation of any two first sample feature data among multiple first sample feature data, and delete any first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain screened first sample feature data; a construction unit, used to input the screened first sample feature data into an untrained LightGBM model, and construct a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model.

[0012] In one embodiment, the model building module also includes a second feature screening unit; the second feature screening unit is used to eliminate the first sample feature data that meets the preset error feature rule from the multiple first sample feature data to obtain the multiple first sample feature data after elimination; wherein the first feature screening unit is used to calculate the Pearson correlation of any two first sample feature data from the multiple first sample feature data after elimination, and delete any first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain the screened first sample feature data.

[0013] A method for identifying post-traumatic stress disorder includes: collecting multiple physiological data of a target user under different stimulation strategies at multiple different stages, the multiple physiological data including the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtaining multiple feature data of the target user based on the multiple physiological data; inputting the multiple feature data into a trained LightGBM model to obtain classification information output by the trained LightGBM model; and identifying whether the target user is a patient with post-traumatic stress disorder according to the classification information.

[0014] A computer-readable storage medium stores a computer program, which implements the steps of the above method when executed by a processor.

[0015] The above-mentioned PTSD identification device, method and storage medium collect multiple physiological data of the target user under different stimulation strategies at different stages, the multiple physiological data include the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtain multiple feature data of the target user based on the multiple physiological data, input the multiple feature data into the trained LightGBM model, obtain the classification information output by the trained LightGBM model, and identify whether the target user is a patient with PTSD based on the classification information. Therefore, patients with PTSD can be automatically identified through the trained LightGBM model, avoiding the missed screening or misscreening of patients with PTSD caused by manual operation, and improving the accuracy of identifying patients with PTSD. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a structural block diagram of a post-traumatic stress disorder identification device in one embodiment;

[0017] Figure 2 A structural block diagram of a data acquisition module in one embodiment;

[0018] Figure 3 is a structural block diagram of a feature extraction unit in one embodiment;

[0019] Figure 4 A structural block diagram of a model building module in one embodiment;

[0020] Figure 5 A diagram of an application environment of a method for identifying post-traumatic stress disorder in an embodiment;

[0021] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0023] The present application provides a post-traumatic stress disorder identification device, which is applied to a server. Figure 1 As shown, in one embodiment, a post-traumatic stress disorder identification device 100 includes:

[0024] The data acquisition module 102 is used to collect multiple physiological data of the target user under different stimulation strategies at different stages, the multiple physiological data include the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtain multiple feature data of the target user based on the multiple physiological data.

[0025] In this embodiment, a plurality of different test stages are set, and different stimulation strategies are used in each stage to stimulate the target user, and physiological data of the target user under each stimulation strategy is collected. Among them, the stimulation strategy can be to present different pictures to the target user, or to provide sound stimulation or air blowing stimulation. The plurality of physiological data include physiological data of electrooculography and physiological data of electrocardiography of the target user. The physiological data of electrooculography can include physiological data of the upper part of the right eye, physiological data of the lower part of the right eye, and physiological data of the right side of the right eye, and the physiological data of electrocardiography can include physiological data of the left side of the chest.

[0026] For example, four sets of electrophysiological signals, including three-lead electrooculogram and one-lead electrocardiogram, are collected from the target user. Among them, the electrooculogram collection positions are the upper part of the right eye, the lower part of the right eye, and the right side of the right eye, and the electrocardiogram collection position is the left chest. The target user sits in front of the computer monitor, wears the electrophysiological signal collection device, and starts the test software after the signal is stable. The test software is used to present one or more pictures to the target user to stimulate the target user. While the target user is watching the pictures, the target user can also be stimulated by sound, such as a short blasting sound. Or, the target user can be stimulated by gas of a certain pressure. Set multiple test stages, record the electrophysiological signals of the target user in each stage, and obtain multiple physiological data of the target user under different stimulation strategies in multiple different stages. After all tests are completed for about 10 seconds, the signal collection is ended.

[0027] Furthermore, multiple characteristic data of the target user are obtained based on multiple physiological data. It can be that multiple characteristic data of the target user are directly extracted from multiple physiological data. For example, when the physiological data is represented by an electrophysiological signal, the standard deviation, mean, amplitude, etc. of the electrophysiological signal are taken as characteristic data. It can also be that multiple characteristic data of the target user are obtained by analyzing multiple physiological data. For example, when the physiological data is represented by an electrophysiological signal, an index related to the autocorrelation and change trend of the electrophysiological signal is obtained by performing nonlinear changes on the electrophysiological signal, and the index is used as one of the characteristic data. And, multiple heart rate variability features of each stage are extracted from the electrophysiological signal, and various heart rate variability features are used as corresponding characteristic data. The typing dimension features of each trial in each stage are extracted from the electrophysiological signal, and the typing dimension features are used as one of the characteristic data. It should be noted that this part is only used as an example of part of it, and the specific method of obtaining multiple characteristic data of the target user based on multiple physiological data is not limited to the example in this part.

[0028] The model application module 104 is used to input multiple feature data into the trained LightGBM model to obtain classification information output by the trained LightGBM model.

[0029] In this embodiment, a trained LightGBM model is used to analyze and process multiple feature data of the target user to obtain classification information of the target user. The trained LightGBM model is obtained by pre-building and training the model using sample feature data of post-traumatic stress disorder. The classification information of the target user is the output data of the trained LightGBM model, which represents the classification of the target user, that is, information representing that the target user is a patient with post-traumatic stress disorder or information representing that the target user is not a patient with post-traumatic stress disorder.

[0030] The classification identification module 106 is used to identify whether the target user is a patient with post-traumatic stress disorder according to the classification information.

[0031] In this embodiment, the classification information output by the trained lightgbm model can indicate that the target user is a patient with post-traumatic stress disorder, or that the target user is not a patient with post-traumatic stress disorder. By means of information recognition, it can be determined based on the classification information whether the target user is a patient with post-traumatic stress disorder.

[0032] The above-mentioned PTSD identification device collects multiple physiological data of the target user under different stimulation strategies at different stages, the multiple physiological data include the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtains multiple feature data of the target user based on the multiple physiological data, inputs the multiple feature data into the trained LightGBM model, obtains the classification information output by the trained LightGBM model, and identifies whether the target user is a patient with PTSD based on the classification information. Therefore, patients with PTSD can be automatically identified through the trained LightGBM model, avoiding missed screening or misscreening of patients with PTSD caused by manual operation, and improving the accuracy of identifying patients with PTSD.

[0033] In one embodiment, Figure 2 As shown, the data acquisition module includes a first data acquisition unit 202 , a second data acquisition unit 204 and a third data acquisition unit 206 . The first data acquisition unit 202 is used to acquire first physiological data under multiple different stimulation strategies in the first stage, wherein the multiple different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes a probe stimulus but does not include an aversive stimulus; the second data acquisition unit 204 is used to acquire second physiological data under multiple different stimulation strategies in the second stage, wherein the multiple different stimulation strategies in the second stage are a second conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, and the second conditioned stimulus sequence is different from the first conditioned stimulus sequence; the third data acquisition unit 206 is used to acquire third physiological data under multiple different stimulation strategies in the third stage, wherein the multiple different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of multiple third conditioned stimuli and multiple fourth conditioned stimuli, the third conditioned stimulus and the fourth conditioned stimulus both include a probe stimulus and do not include an aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; wherein the multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

[0034] Among them, the aversive stimulus is compressed air of a specific pressure given to the target user's neck, and the probe stimulus is a short blasting sound given to the target user; the first conditioned stimulus is giving the target user the first basic conditioned stimulus and giving the target user's neck compressed air of a specific pressure, and giving the target user a short blasting sound; and / or, the second conditioned stimulus is giving the target user the second basic conditioned stimulus and giving the target user a short blasting sound; and / or, the third conditioned stimulus is giving the target user the third basic conditioned stimulus and giving the target user a short blasting sound; and / or, the fourth conditioned stimulus is showing the target user the fourth basic conditioned stimulus and giving the target user a short blasting sound; wherein the first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus are different, and the first basic conditioned stimulus and the third basic conditioned stimulus are the same.

[0035] In this embodiment, the first conditioned stimulus, the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are four different conditioned stimuli. Each conditioned stimulus includes a basic conditioned stimulus and an additional stimulus, and the additional stimulus can be an aversive stimulus or a probe stimulus.

[0036] The basic conditioned stimulus includes a first basic conditioned stimulus, a second basic conditioned stimulus, a third basic conditioned stimulus and a fourth basic conditioned stimulus. The first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus are different, and the first basic conditioned stimulus and the third basic conditioned stimulus are the same. Among them, the basic condition is to provide visual environment stimulation to the target user, the first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus respectively provide different visual environment stimulation to the target user, and the first basic conditioned stimulus and the third basic conditioned stimulus provide the same visual environment stimulation to the target user.

[0037] The probe stimulus is a short explosive sound that appears in every trial during the test; the aversive stimulus is a compressed air blown toward the neck at a certain pressure, which only appears in a specified trial. Here, a trial refers to the appearance of a conditioned stimulus to the end during the test.

[0038] In this embodiment, physiological data collected based on a specific conditioned stimulus combination structure paradigm. The paradigm consists of a first stage, a second stage and a third stage, which are recorded as stage 1, stage 2 and stage 3. Different conditioned stimulus sequences are selected from the alternative conditioned stimulus sequence library in each stage. The first conditioned stimulus sequence corresponds to the first conditioned stimulus sequence in stage 1, the second conditioned stimulus sequence corresponds to the second conditioned stimulus sequence in stage 2, and the third conditioned stimulus sequence corresponds to the third conditioned stimulus sequence in stage 3. Among them, the conditioned stimulus contained in the second conditioned stimulus sequence is the same as that contained in the first conditioned stimulus sequence, that is, both contain multiple first conditioned stimuli and multiple second conditioned stimuli, but the order of the first conditioned stimulus and the second conditioned stimulus in the second conditioned stimulus sequence is different from that in the first conditioned stimulus sequence. That is, the difference between the conditioned stimulus sequences in stage 1 and stage 2 is only that the first conditioned stimulus and the second conditioned stimulus appear in different order. In addition, the sum of the number of the first conditioned stimulus and the second conditioned stimulus in stage 1 and / or stage 2 is greater than the sum of the number of the third conditioned stimulus and the fourth conditioned stimulus in the third conditioned stimulus sequence in stage 3. For example, both stage 1 and stage 2 have 10 groups of conditioned stimuli, and stage 3 has 6 groups of conditioned stimuli. Since the target user becomes more tired as time goes on during the test, the target user's physiological data collected in the third stage can be more accurate by setting the total number of conditioned stimuli in the third stage to be smaller than the total number of conditioned stimuli in the first two stages.

[0039] Therefore, through the above method, multiple physiological data under different stimulation strategies at different stages can be collected.

[0040] In one embodiment, the data acquisition module further includes a feature extraction unit. Figure 3 As shown, the feature extraction unit includes a first feature extraction subunit 302, a second feature extraction subunit 304, a third feature extraction subunit 306 and a fourth feature extraction subunit 308. The plurality of feature data includes physiological feature data of each stage, differential feature data of the same conditional stimulation, physiological feature data of each trial, and physiological feature data corresponding to before and after the probe stimulation and physiological feature data corresponding to before and after the disgust stimulation.

[0041] The first feature extraction subunit 302 is used to obtain physiological feature data of each stage based on the physiological data of each stage in the first stage, the second stage and the third stage.

[0042] In one aspect, the multiple characteristic data include characteristic data of electrocardiogram and characteristic data of electrooculogram. The multiple characteristic data acquired by the data acquisition module include characteristic data of electrocardiogram and characteristic data of electrooculogram at each stage. It can be that the characteristic data of electrocardiogram at each stage is acquired based on the electrocardiological physiological data of each stage in the first stage, the second stage and the third stage, such as multiple heart rate variability characteristics at each stage. It can also be that the mean of various conditional stimuli at each stage is calculated based on the characteristic data of electrocardiogram and characteristic data of electrooculogram of each trial as the physiological characteristic data of each conditional stimulus at each stage.

[0043] The second feature extraction subunit 304 is used to obtain differential feature data of the same conditional stimulus according to the physiological feature data of each stage.

[0044] In one aspect, the multiple characteristic data acquired by the data acquisition module include differential characteristic data corresponding to the same conditional stimulus in multiple stages. Among them, the differential characteristic data refers to the difference in physiological characteristic data corresponding to the same conditional stimulus, and the physiological characteristic data can be represented by the mean of the above-mentioned various conditional stimuli. For example, the multiple differential characteristic data include the difference between the physiological characteristic data corresponding to the first conditional stimulus in the second stage and the physiological characteristic data corresponding to the first conditional stimulus in the first stage, and the difference between the physiological characteristic data corresponding to the second conditional stimulus in the second stage and the physiological characteristic data corresponding to the second conditional stimulus in the first stage. In addition, in this embodiment, the third conditional stimulus and the first conditional stimulus can also be treated as the same conditional stimulus. That is, the multiple differential characteristic data include the difference between the physiological characteristic data corresponding to the third conditional stimulus in the third stage and the physiological characteristic data corresponding to the first conditional stimulus in the second stage, and the difference between the physiological characteristic data corresponding to the third conditional stimulus in the third stage and the physiological characteristic data corresponding to the first conditional stimulus in the first stage.

[0045] For example, multiple differential characteristic data include the physiological characteristic data corresponding to the first conditioned stimulus in stage 2 minus the physiological characteristic data corresponding to the first conditioned stimulus in stage 1, the physiological characteristic data corresponding to the third conditioned stimulus in stage 3 minus the physiological characteristic data corresponding to the first conditioned stimulus in stage 2, the physiological characteristic data corresponding to the third conditioned stimulus in stage 3 minus the physiological characteristic data corresponding to the first conditioned stimulus in stage 1, and the physiological characteristic data corresponding to the fourth conditioned stimulus in stage 2 minus the physiological characteristic data corresponding to the fourth conditioned stimulus in stage 1.

[0046] The third feature extraction subunit 306 is used to obtain physiological feature data of each trial based on the physiological data of each trial in each of the first stage, the second stage and the third stage.

[0047] In one aspect, the multiple characteristic data acquired by the data acquisition module include physiological characteristic data of each trial in each stage. It can be that based on the electrocardiographic physiological data of each stage in the first stage, the second stage and the third stage, the electrophysiological signal of each trial in each stage is subjected to nonlinear changes to obtain the electrocardiographic characteristic data such as the index, standard deviation, mean and classification dimension characteristics related to the autocorrelation and change trend of the electrophysiological signal, and based on the electroocular electrophysiological data of each stage in the first stage, the second stage and the third stage, the electroocular characteristic data such as the index, standard deviation, mean and classification dimension characteristics related to the autocorrelation and change trend of the electrophysiological signal are subjected to nonlinear changes to obtain the electrophysiological characteristic data such as the index, standard deviation, mean and classification dimension characteristics related to the autocorrelation and change trend of the electrophysiological signal.

[0048] The fourth feature extraction subunit 308 is used to obtain the physiological feature data corresponding to before and after the probe stimulation and the physiological feature data corresponding to before and after the disgust stimulation based on the physiological data of each trial in each of the first stage, the second stage and the third stage.

[0049] In one aspect, the multiple characteristic data acquired by the data acquisition module include physiological characteristic data before and after probe stimulation of each trial in each stage, physiological characteristic data before and after aversion stimulation of each trial, wherein the physiological characteristic data are represented by the mean and / or amplitude of the collected electrophysiological signals. For example, the mean and amplitude of the electrophysiological signals corresponding to one or more pre-set time points before probe stimulation and aversion stimulation of each trial in each stage, and the mean and amplitude of the electrophysiological signals corresponding to one or more pre-set time points after probe stimulation and aversion stimulation of each trial in each stage are calculated respectively, as the physiological characteristic data before and after stimulation of different stimuli.

[0050] Alternatively, the plurality of characteristic data acquired by the data acquisition module further include first trend characteristic data indicating changes in the electrophysiological signal before and after the probe stimulation in each stage and second trend characteristic data indicating changes in the electrophysiological signal before and after the aversion stimulation in each stage. The first trend characteristic data may be the ratio of the electrophysiological signal before the probe stimulation to the electrophysiological signal after the probe stimulation, and the second trend characteristic data may be the ratio of the electrophysiological signal before the aversion stimulation to the electrophysiological signal after the aversion stimulation.

[0051] Alternatively, the multiple feature data acquired by the data acquisition module also include the mean value of the electrophysiological signal of the probe stimulation and the mean value of the electrophysiological signal of the aversive stimulation under each conditional stimulation in each stage, so as to constitute the feature data of each conditional stimulation combination.

[0052] Alternatively, the multiple characteristic data acquired by the data acquisition module also include the ratio of the characteristic data corresponding to each trial to the characteristic data of the probe stimulation or aversion stimulation corresponding thereto. For example, the mean and amplitude of the electrophysiological signal corresponding to one or more pre-set time points before the probe stimulation and before the aversion stimulation of each trial are calculated respectively, as well as the mean and amplitude of the electrophysiological signal corresponding to one or more pre-set time points after the probe stimulation and after the aversion stimulation of each trial in each stage, as physiological characteristic data before and after stimulation of different stimuli. Further, the mean of the physiological characteristic data of the probe stimulation in each stage and the mean of the physiological characteristic data of the aversion stimulation can be calculated. For example, the ratio of the amplitude of the electrophysiological signal collected at any time point after the probe stimulation to the mean of the electrophysiological signal after the probe stimulation in the corresponding stage is used as one of the characteristic data. By analogy, the ratio of the amplitude of the electrophysiological signal corresponding to each conditional stimulus of each trial to the mean of the corresponding conditional stimulus of the corresponding stage is obtained.

[0053] In one embodiment, a post-traumatic stress disorder identification device further includes a model building module and a model training module. Figure 4 As shown, the model building module includes a first sample collection unit 402 and a building unit 404 , and the model training module includes a second sample collection unit 406 and a training unit 408 .

[0054] The first sample acquisition unit 402 is used to collect multiple first sample physiological data of the first sample user under different stimulation strategies at multiple different stages, the multiple first sample physiological data include the electrooculographic physiological data and the electrocardiographic physiological data of the first sample user, and obtain multiple first sample feature data of the first sample user based on the multiple first sample physiological data.

[0055] In this embodiment, the first sample user is different from the target user, and the multiple first sample physiological data of the first sample user are used as sample data for constructing the lightgbm model. The multiple first sample physiological data of the first sample user collected have the same attributes as the multiple physiological data of the target user used when the model is applied. The multiple first sample physiological data can also refer to the relevant description of the multiple physiological data of the target user mentioned above.

[0056] The construction unit 404 is used to input the plurality of first sample feature data into an untrained LightGBM model, and construct a target LightGBM model by adjusting the number of sub-evaluators and feature contribution values ​​of the untrained LightGBM model.

[0057] In this embodiment, an untrained LightGBM model is used to perform multi-step screening on the collected multiple first sample feature data, and the first sample feature data whose feature contribution is less than the specified contribution threshold is gradually deleted by continuously adjusting the number of sub-evaluators and the size of the feature contribution in the LightGBM model; the number of sub-evaluators can be gradually reduced from 1000 to between 100 and 200, and the contribution threshold must be specified for each number of sub-evaluators, because the more sub-evaluators there are, the higher the contribution of the same first sample feature data, which is in direct proportion; at the same time, since there are many first sample feature data to be screened initially, in order to prevent some of the first sample feature data from not being used in the process of building the model, resulting in a contribution of 0, a higher number of sub-evaluators must be set in the initial stage. Finally, the target LightGBM model is obtained. The number of sub-evaluators in the target LightGBM model is fixed, so a target LightGBM model suitable for this application is constructed. For example, more than 35,000 first sample feature data were obtained in the early stage through the above method. After adjusting the number of sub-evaluators and the size of feature contribution, the target lightgbm model was obtained. At this time, the first sample feature data screened out by the target lightgbm model was 35.

[0058] The second sample acquisition unit 406 is used to collect multiple second sample physiological data of multiple second sample users under different stimulation strategies at multiple different stages, the multiple second sample physiological data include the electrooculographic physiological data and the electrocardiographic physiological data of the second sample users, and obtain multiple second sample feature data of the second sample users based on the multiple second sample physiological data.

[0059] In this embodiment, after the target LightGBM model is constructed, the target LightGBM model needs to be trained so that the LightGBM model obtained after training is suitable for the identification of post-traumatic stress disorder in this application. Here, multiple second sample physiological data of multiple second sample users under different stimulation strategies at multiple different stages are collected. Among them, the multiple second sample physiological data of the second sample user are sample data for training the target LightGBM model, which have the same attributes as the multiple physiological data of the target user used when the model is applied. For multiple second sample physiological data, please refer to the above-mentioned relevant description of the multiple physiological data of the target user.

[0060] The training unit 408 is used to perform model training on the target LightGBM model according to the plurality of second sample feature data and adopt a ten-fold cross validation method to obtain a trained LightGBM model.

[0061] In this embodiment, the collected second sample feature data is divided into two parts, one part is used as the model input data of the target LightGBM model, and the other part is used as the prediction result of the model output reference of the target LightGBM model. The allocation method adopts the ten-fold cross-validation method, thereby training the target LightGBM model to obtain a trained LightGBM model.

[0062] In one embodiment, the model building module also includes: a first feature screening unit, used to calculate the Pearson correlation of any two first sample feature data among multiple first sample feature data, and delete any first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain screened first sample feature data; a construction unit, used to input the screened first sample feature data into an untrained LightGBM model, and construct a target LightGBM model by adjusting the number of sub-evaluators and feature contribution values ​​of the untrained LightGBM model.

[0063] In this embodiment, before inputting a plurality of first sample feature data into an untrained LightGBM model, the plurality of first sample feature data are screened by removing redundant first sample feature data using Pearson correlation. Specifically, the Pearson correlation between the first sample feature data is calculated, and one of the two first sample feature data whose correlation exceeds a specified threshold is retained. The threshold can be set within the interval [0.6, 0.9]. Therefore, the generalization effect of the model can be improved.

[0064] In one embodiment, the model building module also includes a second feature screening unit; the second feature screening unit is used to eliminate the first sample feature data that meets the preset error feature rules from the multiple first sample feature data to obtain the multiple first sample feature data after elimination; wherein the first feature screening unit is used to calculate the Pearson correlation of any two first sample feature data from the multiple first sample feature data after elimination, and delete any one of the first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain the screened first sample feature data.

[0065] In this embodiment, before removing redundant first sample feature data by using the Pearson correlation method, the first sample feature data with errors are removed, and the removal method is: first sample feature data that meets the preset error feature rule is removed to obtain multiple first sample feature data after removal. For example, the first sample feature data that is calculated incorrectly or cannot be calculated among the multiple first sample feature data is deleted, thereby reducing the efficiency of subsequent Pearson correlation calculations.

[0066] Therefore, by constructing and training the trained lightgbm model in the above manner, a sensitivity of 0.89 and a specificity of 0.84 can be obtained. Among them, the sensitivity of 0.89 means that the trained lightgbm model can identify 89% of patients with post-traumatic stress disorder, and the specificity of 0.84 means that the trained lightgbm model can identify 84% of patients without post-traumatic stress disorder. In addition, the correlation between the pre- and post-tests of the trained lightgbm model reached 0.71, that is, the correlation of the results of the same user before and after was 0.71. It can be seen that the obtained trained lightgbm model has a high accuracy in identifying post-traumatic stress disorder, and the use of the trained lightgbm model can improve the accuracy of identifying patients with post-traumatic stress disorder.

[0067] Each module in the above-mentioned post-traumatic stress disorder identification device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0068] The present application also provides a method for identifying post-traumatic stress disorder, which is applied to Figure 5 Taking the server 502 in the example as an example, the method includes the following steps: collecting multiple physiological data of the target user under different stimulation strategies at multiple different stages, the multiple physiological data include the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtaining multiple feature data of the target user based on the multiple physiological data; inputting the multiple feature data into the trained LightGBM model to obtain classification information output by the trained LightGBM model; and identifying whether the target user is a patient with post-traumatic stress disorder according to the classification information.

[0069] In one embodiment, the above-mentioned collecting multiple physiological data of the target user under different stimulation strategies at multiple different stages includes: collecting first physiological data under multiple different stimulation strategies in the first stage, the multiple different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes a probe stimulus but does not include an aversive stimulus; collecting second physiological data under multiple different stimulation strategies in the second stage, the multiple different stimulation strategies in the second stage are a second conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, and the second conditioned stimulus sequence is different from the first conditioned stimulus sequence; collecting third physiological data under multiple different stimulation strategies in the third stage, the multiple different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of multiple third conditioned stimuli and multiple fourth conditioned stimuli, the third conditioned stimulus and the fourth conditioned stimulus both include a probe stimulus and do not include an aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; wherein the multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

[0070] In one embodiment, the aversive stimulus is compressed air of a specific pressure given to the target user's neck, and the probe stimulus is a short popping sound given to the target user; the first conditioned stimulus is a first basic conditioned stimulus given to the target user and compressed air of a specific pressure given to the target user's neck, as well as a short popping sound given to the target user; and / or, the second conditioned stimulus is a second basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the third conditioned stimulus is a third basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the fourth conditioned stimulus is a fourth basic conditioned stimulus displayed to the target user and a short popping sound given to the target user; wherein, the first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus are different, and the first basic conditioned stimulus and the third basic conditioned stimulus are the same.

[0071] In one embodiment, the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the first conditioned stimulus sequence is different from the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the second conditioned stimulus sequence; and / or, the sum of the numbers of the first conditioned stimuli and the second conditioned stimuli in the first conditioned stimulus sequence or the second conditioned stimulus sequence is greater than the sum of the numbers of the third conditioned stimuli and the fourth conditioned stimuli in the third conditioned stimulus sequence.

[0072] In one embodiment, the above-mentioned acquisition of multiple characteristic data of the target user based on multiple physiological data includes: acquiring physiological characteristic data of each stage based on the physiological data of each stage in the first stage, the second stage and the third stage; acquiring differential characteristic data of the same conditioned stimulus according to the physiological characteristic data of each stage; acquiring physiological characteristic data of each trial based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage; acquiring physiological characteristic data corresponding to before and after probe stimulation and physiological characteristic data corresponding to before and after aversive stimulation based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage; the multiple characteristic data include physiological characteristic data of each stage, differential characteristic data of the same conditioned stimulus, physiological characteristic data of each trial, physiological characteristic data corresponding to before and after probe stimulation, and physiological characteristic data corresponding to before and after aversive stimulation.

[0073] In one embodiment, the above-mentioned method for identifying post-traumatic stress disorder also includes a model building step, which includes: collecting multiple first sample physiological data of the first sample user under different stimulation strategies at multiple different stages, the multiple first sample physiological data include the physiological data of the electrooculogram and the physiological data of the electrocardiogram of the first sample user, and obtaining multiple first sample feature data of the first sample user based on the multiple first sample physiological data; inputting the multiple first sample feature data into an untrained LightGBM model, and constructing a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model; collecting multiple second sample physiological data of the second sample users under different stimulation strategies at multiple different stages, the multiple second sample physiological data include the physiological data of the electrooculogram and the physiological data of the electrocardiogram of the second sample users, and obtaining multiple second sample feature data of the second sample users based on the multiple second sample physiological data; training the target LightGBM model according to the multiple second sample feature data and using a ten-fold cross-validation method to obtain a trained LightGBM model.

[0074] In one embodiment, before the above-mentioned step of inputting multiple first sample feature data into an untrained LightGBM model, it also includes: calculating the Pearson correlation of any two first sample feature data among the multiple first sample feature data, and deleting any first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain the screened first sample feature data; the above-mentioned step of inputting multiple first sample feature data into the untrained LightGBM model includes: inputting the screened first sample feature data into the untrained LightGBM model, and constructing a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model.

[0075] In one embodiment, before the step of calculating the Pearson correlation of any two first sample feature data among the multiple first sample feature data, it also includes: eliminating the first sample feature data that meet the preset error feature rule among the multiple first sample feature data to obtain the multiple first sample feature data after elimination; the above-mentioned calculation of the Pearson correlation of any two first sample feature data among the multiple first sample feature data includes: calculating the Pearson correlation of any two first sample feature data among the multiple first sample feature data after elimination, and deleting any one of the first sample feature data among any two first sample feature data whose Pearson correlation exceeds the specified threshold to obtain the screened first sample feature data.

[0076] For more definitions of the technical features in the post-traumatic stress disorder identification method, please refer to the above description of the corresponding modules or units in the post-traumatic stress disorder identification device, which will not be repeated here.

[0077] In some embodiments, a computer-readable storage medium is provided on which a computer program is stored. In different embodiments, when the computer program is executed by a processor, the steps performed by each functional module or functional unit of the post-traumatic stress disorder identification device described in the above embodiments are respectively implemented.

[0078] Those of ordinary skill in the art can understand that all or part of the process of implementing the method for acquiring depressive symptom features based on facial expression changes can be completed by instructing related hardware through a computer program, and the aforementioned computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it may include the process of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0079] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to connect to an external device to receive physiological data of one or more users input by the external device. When the computer program is executed by the processor, a method for identifying post-traumatic stress disorder is implemented.

[0080] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0081] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: collecting multiple physiological data of a target user under different stimulation strategies at multiple different stages, the multiple physiological data including the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtaining multiple feature data of the target user based on the multiple physiological data; inputting the multiple feature data into a trained LightGBM model to obtain classification information output by the trained LightGBM model; and identifying whether the target user is a patient with post-traumatic stress disorder based on the classification information.

[0082] In one embodiment, when the processor executes a computer program to implement the above-mentioned step of collecting multiple physiological data of the target user under different stimulation strategies at multiple different stages, the following steps are specifically implemented: collecting first physiological data under multiple different stimulation strategies in the first stage, the multiple different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes a probe stimulus but does not include an aversive stimulus; collecting second physiological data under multiple different stimulation strategies in the second stage, the multiple different stimulation strategies in the second stage are a second conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, and the second conditioned stimulus sequence is different from the first conditioned stimulus sequence; collecting third physiological data under multiple different stimulation strategies in the third stage, the multiple different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of multiple third conditioned stimuli and multiple fourth conditioned stimuli, the third conditioned stimulus and the fourth conditioned stimulus both include a probe stimulus and do not include an aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; wherein the multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

[0083] In one embodiment, the aversive stimulus is compressed air of a specific pressure given to the target user's neck, and the probe stimulus is a short popping sound given to the target user; the first conditioned stimulus is a first basic conditioned stimulus given to the target user and compressed air of a specific pressure given to the target user's neck, as well as a short popping sound given to the target user; and / or, the second conditioned stimulus is a second basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the third conditioned stimulus is a third basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the fourth conditioned stimulus is a fourth basic conditioned stimulus displayed to the target user and a short popping sound given to the target user; wherein, the first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus are different, and the first basic conditioned stimulus and the third basic conditioned stimulus are the same.

[0084] In one embodiment, the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the first conditioned stimulus sequence is different from the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the second conditioned stimulus sequence; and / or, the sum of the numbers of the first conditioned stimuli and the second conditioned stimuli in the first conditioned stimulus sequence or the second conditioned stimulus sequence is greater than the sum of the numbers of the third conditioned stimuli and the fourth conditioned stimuli in the third conditioned stimulus sequence.

[0085] In one embodiment, when the processor executes a computer program to implement the above-mentioned step of obtaining multiple characteristic data of the target user based on multiple physiological data, the following steps are specifically implemented: based on the physiological data of each stage in the first stage, the second stage and the third stage, the physiological characteristic data of each stage are obtained; based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage, the physiological characteristic data of each trial are obtained; based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage, the physiological characteristic data corresponding to before and after the probe stimulation and the physiological characteristic data corresponding to before and after the aversive stimulation are obtained; the multiple characteristic data include the physiological characteristic data of each stage, the differential characteristic data of the same conditional stimulation, the physiological characteristic data of each trial, the physiological characteristic data corresponding to before and after the probe stimulation, and the physiological characteristic data corresponding to before and after the aversive stimulation.

[0086] In one embodiment, when the processor executes the computer program, the following steps are also implemented: collecting multiple first sample physiological data of the first sample user under different stimulation strategies at multiple different stages, the multiple first sample physiological data include the physiological data of the electrooculogram and the physiological data of the electrocardiogram of the first sample user, and obtaining multiple first sample feature data of the first sample user based on the multiple first sample physiological data; inputting the multiple first sample feature data into an untrained LightGBM model, and constructing a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model; collecting multiple second sample physiological data of the second sample users under different stimulation strategies at multiple different stages, the multiple second sample physiological data include the physiological data of the electrooculogram and the physiological data of the electrocardiogram of the second sample users, and obtaining multiple second sample feature data of the second sample users based on the multiple second sample physiological data; training the target LightGBM model according to the multiple second sample feature data and using a ten-fold cross validation method to obtain a trained LightGBM model.

[0087] In one embodiment, when the processor executes the computer program, the following steps are also implemented: calculating the Pearson correlation of any two first sample feature data among multiple first sample feature data, and deleting any first sample feature data among any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain screened first sample feature data; when the processor executes the computer program to implement the above-mentioned step of inputting multiple first sample feature data into an untrained LightGBM model, the following steps are specifically implemented: inputting the screened first sample feature data into the untrained LightGBM model, and constructing a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model.

[0088] In one embodiment, when the processor executes the computer program, the following steps are also implemented: first sample feature data that meets a preset error feature rule from multiple first sample feature data are eliminated to obtain multiple first sample feature data after elimination; when the processor executes the computer program to implement the above-mentioned step of calculating the Pearson correlation of any two first sample feature data from the multiple first sample feature data, the following steps are specifically implemented: calculating the Pearson correlation of any two first sample feature data from the multiple first sample feature data after elimination, and deleting any first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain the screened first sample feature data.

[0089] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: collecting multiple physiological data of a target user under different stimulation strategies at multiple different stages, the multiple physiological data including the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtaining multiple feature data of the target user based on the multiple physiological data; inputting the multiple feature data into a trained LightGBM model to obtain classification information output by the trained LightGBM model; and identifying whether the target user is a patient with post-traumatic stress disorder based on the classification information.

[0090] In one embodiment, when the computer program is executed by the processor to implement the above-mentioned steps of collecting multiple physiological data of the target user under different stimulation strategies at multiple different stages, the following steps are specifically implemented: collecting first physiological data under multiple different stimulation strategies in the first stage, the multiple different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes a probe stimulus but does not include an aversive stimulus; collecting second physiological data under multiple different stimulation strategies in the second stage, the multiple different stimulation strategies in the second stage are a second conditioned stimulus sequence consisting of multiple first conditioned stimuli and multiple second conditioned stimuli, and the second conditioned stimulus sequence is different from the first conditioned stimulus sequence; collecting third physiological data under multiple different stimulation strategies in the third stage, the multiple different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of multiple third conditioned stimuli and multiple fourth conditioned stimuli, the third conditioned stimulus and the fourth conditioned stimulus both include a probe stimulus and do not include an aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; wherein the multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

[0091] In one embodiment, the aversive stimulus is compressed air of a specific pressure given to the target user's neck, and the probe stimulus is a short popping sound given to the target user; the first conditioned stimulus is a first basic conditioned stimulus given to the target user and compressed air of a specific pressure given to the target user's neck, as well as a short popping sound given to the target user; and / or, the second conditioned stimulus is a second basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the third conditioned stimulus is a third basic conditioned stimulus given to the target user and a short popping sound given to the target user; and / or, the fourth conditioned stimulus is a fourth basic conditioned stimulus displayed to the target user and a short popping sound given to the target user; wherein, the first basic conditioned stimulus, the second basic conditioned stimulus and the fourth basic conditioned stimulus are different, and the first basic conditioned stimulus and the third basic conditioned stimulus are the same.

[0092] In one embodiment, the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the first conditioned stimulus sequence is different from the order of the multiple first conditioned stimuli and the multiple second conditioned stimuli in the second conditioned stimulus sequence; and / or, the sum of the numbers of the first conditioned stimuli and the second conditioned stimuli in the first conditioned stimulus sequence or the second conditioned stimulus sequence is greater than the sum of the numbers of the third conditioned stimuli and the fourth conditioned stimuli in the third conditioned stimulus sequence.

[0093] In one embodiment, when the computer program is executed by the processor to implement the above-mentioned steps of obtaining multiple characteristic data of the target user based on multiple physiological data, the following steps are specifically implemented: based on the physiological data of each stage in the first stage, the second stage and the third stage, the physiological characteristic data of each stage are obtained; based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage, the physiological characteristic data of each trial are obtained; based on the physiological data of each trial in each stage in the first stage, the second stage and the third stage, the physiological characteristic data corresponding to before and after the probe stimulation and the physiological characteristic data corresponding to before and after the aversive stimulation are obtained; the multiple characteristic data include the physiological characteristic data of each stage, the differential characteristic data of the same conditional stimulation, the physiological characteristic data of each trial, and the physiological characteristic data corresponding to before and after the probe stimulation and the physiological characteristic data corresponding to before and after the aversive stimulation.

[0094] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: collecting multiple first sample physiological data of the first sample user under different stimulation strategies at multiple different stages, the multiple first sample physiological data include the physiological data of the electrooculogram and the physiological data of the electrocardiogram of the first sample user, and obtaining multiple first sample feature data of the first sample user based on the multiple first sample physiological data; inputting the multiple first sample feature data into an untrained LightGBM model, and constructing a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model; collecting multiple second sample physiological data of the second sample users under different stimulation strategies at multiple different stages, the multiple second sample physiological data include the physiological data of the electrooculogram and the physiological data of the electrocardiogram of the second sample users, and obtaining multiple second sample feature data of the second sample users based on the multiple second sample physiological data; training the target LightGBM model according to the multiple second sample feature data and using a ten-fold cross validation method to obtain a trained LightGBM model.

[0095] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: calculating the Pearson correlation of any two first sample feature data among multiple first sample feature data, and deleting any first sample feature data among any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain screened first sample feature data; when the computer program is executed by the processor to implement the above-mentioned step of inputting multiple first sample feature data into an untrained LightGBM model, the following steps are specifically implemented: inputting the screened first sample feature data into the untrained LightGBM model, and constructing a target LightGBM model by adjusting the number of sub-evaluators and the value of feature contribution of the untrained LightGBM model.

[0096] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: first sample feature data that meets a preset error feature rule from multiple first sample feature data are eliminated to obtain multiple first sample feature data after elimination; when the computer program is executed by the processor to implement the above-mentioned step of calculating the Pearson correlation of any two first sample feature data from the multiple first sample feature data, the following steps are specifically implemented: calculating the Pearson correlation of any two first sample feature data from the multiple first sample feature data after elimination, and deleting any first sample feature data from any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain the screened first sample feature data.

[0097] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0098] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0099] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A post-traumatic stress disorder identification device, characterized in that: It includes data collection module, model application module and classification and recognition module; The data acquisition module is used to collect multiple physiological data of the target user under different stimulation strategies at different stages, the multiple physiological data including the target user's electrooculographic physiological data and electrocardiographic physiological data, and obtain multiple feature data of the target user based on the multiple physiological data; The model application module is used to input the multiple feature data into the trained LightGBM model to obtain the classification information output by the trained LightGBM model; The classification identification module is used to identify whether the target user is a patient with post-traumatic stress disorder according to the classification information; Wherein, the data acquisition module includes: a first data acquisition unit, configured to acquire first physiological data under a plurality of different stimulation strategies in a first stage, wherein the plurality of different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of a plurality of first conditioned stimuli and a plurality of second conditioned stimuli, wherein the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes the probe stimulus but does not include the aversive stimulus; A second data acquisition unit is used to acquire second physiological data under a plurality of different stimulation strategies in a second stage, wherein the plurality of different stimulation strategies in the second stage are a second conditioned stimulation sequence consisting of a plurality of the first conditioned stimulations and a plurality of the second conditioned stimulations, and the second conditioned stimulation sequence is different from the first conditioned stimulation sequence; a third data acquisition unit, configured to acquire third physiological data under a plurality of different stimulation strategies in a third stage, wherein the plurality of different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of a plurality of third conditioned stimuli and a plurality of fourth conditioned stimuli, wherein the third conditioned stimulus and the fourth conditioned stimulus both include the probe stimulus and do not include the aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; The multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

2. The device according to claim 1, characterized in that The aversive stimulus is compressed air of a specific pressure given to the target user's neck, and the probe stimulus is a short explosive sound given to the target user; The first conditioned stimulus is to give the target user a first basic conditioned stimulus and give the target user's neck compressed air with a specific pressure, and give the target user a short blasting sound; and / or, The second conditioned stimulus is to give the target user a second basic conditioned stimulus and give the target user a short explosive sound; and / or, The third conditioned stimulus is to give the target user a third basic conditioned stimulus and give the target user a short explosive sound; and / or, The fourth conditioned stimulus is to present the fourth basic conditioned stimulus to the target user and give the target user a short explosive sound; Among them, the first basal conditioned stimulus, the second basal conditioned stimulus and the fourth basal conditioned stimulus are different, and the first basal conditioned stimulus and the third basal conditioned stimulus are the same.

3. The device according to claim 1, characterized in that The order of the first conditioned stimuli and the second conditioned stimuli in the first conditioned stimulus sequence is different from the order of the first conditioned stimuli and the second conditioned stimuli in the second conditioned stimulus sequence; and / or, The total number of the first conditioned stimuli and the second conditioned stimuli in the first conditioned stimulus sequence or the second conditioned stimulus sequence is greater than the total number of the third conditioned stimuli and the fourth conditioned stimuli in the third conditioned stimulus sequence.

4. The device according to claim 1, characterized in that The data acquisition module also includes a feature extraction unit, and the feature extraction unit includes: A first feature extraction subunit, configured to obtain physiological feature data of each stage based on the physiological data of each stage in the first stage, the second stage and the third stage; A second feature extraction subunit is used to obtain differential feature data of the same conditional stimulus according to the physiological feature data of each stage; A third feature extraction subunit is used to obtain physiological feature data of each trial based on the physiological data of each trial in each of the first stage, the second stage and the third stage; a fourth feature extraction subunit, configured to obtain the physiological feature data corresponding to before and after the probe stimulation and the physiological feature data corresponding to before and after the disgust stimulation based on the physiological data of each trial in each of the first stage, the second stage and the third stage; The multiple characteristic data include the physiological characteristic data of each stage, the difference characteristic data of the same conditional stimulation, the physiological characteristic data of each trial, the physiological characteristic data corresponding to before and after the probe stimulation, and the physiological characteristic data corresponding to before and after the disgust stimulation.

5. The device according to claim 1, characterized in that The device further comprises a model construction module and a model training module, wherein the model construction module comprises a first sample collection unit and a construction unit, and the model training module comprises a second sample collection unit and a training unit; The first sample collection unit is used to collect a plurality of first sample physiological data of the first sample user under different stimulation strategies at a plurality of different stages, the plurality of first sample physiological data including the physiological data of electrooculogram and the physiological data of electrocardiogram of the first sample user, and to obtain a plurality of first sample feature data of the first sample user based on the plurality of first sample physiological data; The construction unit is used to input the plurality of first sample feature data into an untrained LightGBM model, and construct a target LightGBM model by adjusting the number of sub-evaluators and feature contribution values ​​of the untrained LightGBM model; a second sample collection unit, configured to collect a plurality of second sample physiological data of a plurality of second sample users under different stimulation strategies at a plurality of different stages, wherein the plurality of second sample physiological data include electrooculographic physiological data and electrocardiographic physiological data of the second sample users, and to obtain a plurality of second sample feature data of the second sample users based on the plurality of second sample physiological data; A training unit is used to perform model training on the target LightGBM model according to the plurality of second sample feature data and adopt a ten-fold cross validation method to obtain the trained LightGBM model.

6. The device according to claim 5, characterized in that The model building module also includes: A first feature screening unit is used to calculate the Pearson correlation of any two first sample feature data among the plurality of first sample feature data, and delete any one of the first sample feature data among any two first sample feature data whose Pearson correlation exceeds a specified threshold, so as to obtain screened first sample feature data; The construction unit is used to input the screened first sample feature data into an untrained LightGBM model, and construct a target LightGBM model by adjusting the number of sub-evaluators and feature contribution values ​​of the untrained LightGBM model.

7. The device according to claim 6, characterized in that The model building module also includes a second feature screening unit; A second feature screening unit is used to remove the first sample feature data that meets a preset error feature rule from the plurality of first sample feature data to obtain a plurality of first sample feature data after removal; Among them, the first feature screening unit is used to calculate the Pearson correlation of any two first sample feature data among the multiple first sample feature data after elimination, and delete any first sample feature data among any two first sample feature data whose Pearson correlation exceeds a specified threshold to obtain the screened first sample feature data.

8. A method for identifying post-traumatic stress disorder, characterized in that: The method comprises: Collecting multiple physiological data of the target user under different stimulation strategies at different stages, the multiple physiological data including the target user's electrooculographic physiological data and electrocardiographic physiological data, and acquiring multiple feature data of the target user based on the multiple physiological data; Input the plurality of feature data into a trained lightgbm model to obtain classification information output by the trained lightgbm model; Identifying whether the target user is a patient with post-traumatic stress disorder according to the classification information; The collecting of multiple physiological data of the target user under different stimulation strategies at different stages includes: collecting first physiological data under a plurality of different stimulation strategies in a first stage, wherein the plurality of different stimulation strategies in the first stage are a first conditioned stimulus sequence consisting of a plurality of first conditioned stimuli and a plurality of second conditioned stimuli, wherein the first conditioned stimulus includes an aversive stimulus and a probe stimulus, and the second conditioned stimulus includes the probe stimulus but does not include the aversive stimulus; collecting second physiological data under a plurality of different stimulation strategies in a second stage, wherein the plurality of different stimulation strategies in the second stage are a second conditioned stimulation sequence consisting of a plurality of the first conditioned stimulations and a plurality of the second conditioned stimulations, and the second conditioned stimulation sequence is different from the first conditioned stimulation sequence; collecting third physiological data under a plurality of different stimulation strategies in the third stage, wherein the plurality of different stimulation strategies in the third stage are a third conditioned stimulus sequence consisting of a plurality of third conditioned stimuli and a plurality of fourth conditioned stimuli, the third conditioned stimulus and the fourth conditioned stimulus both contain the probe stimulus and do not contain the aversive stimulus, and the third conditioned stimulus, the second conditioned stimulus and the fourth conditioned stimulus are different; The multiple physiological data include the first physiological data, the second physiological data and the third physiological data.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 8 are implemented.

Images