CN115171658A

CN115171658A - Training method, training device, training apparatus, and storage medium

Info

Publication number: CN115171658A
Application number: CN202210661128.7A
Authority: CN
Inventors: 张志林; 李胜楠; 杨伟平; 梁栋; 吴景龙
Original assignee: Shenzhen Institute of Advanced Technology of CAS
Current assignee: Shenzhen Institute of Advanced Technology of CAS
Priority date: 2022-06-13
Filing date: 2022-06-13
Publication date: 2022-10-11
Also published as: WO2023240951A1

Abstract

The application provides a training method, a training device, training equipment and a storage medium, and relates to the technical field of computers. The training method comprises the following steps: randomly displaying a perception decision task, wherein the perception decision task comprises a visual classification task, an auditory classification task and a visual-auditory classification task, the visual classification task comprises respectively classifying the M first pictures, the auditory classification task comprises respectively classifying the N first sounds, the visual-auditory classification task comprises respectively classifying the L audiovisual stimulation pairs, and each audiovisual stimulation pair comprises a second picture and a second sound corresponding to a target in the second picture; collecting behavior response data generated when a user completes a perception decision task; and determining a training result according to the behavior response data, wherein the training result comprises the accuracy of the perception decision task completed by the user. The training method provided by the application can be used for training in a mode of combining multiple channels of vision, hearing and vision and hearing, and can be used for effectively improving the perception decision-making capability of individuals.

Description

Training method, training device, training apparatus, and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a training method, an apparatus, a device, and a storage medium.

Background

Individuals experience some degree of cognitive decline during aging, and Alzheimer's Disease (AD) is a progressive degenerative disease of the nervous system with underlying disease. Clinically, it is manifested as general dementia such as dysmnesia, aphasia, disuse, agnosia, impairment of visual spatial skills, dysfunction in execution, and changes in personality and behavior.

For the elderly, alzheimer's disease reduces their perceptual decision-making abilities, resulting in poor performance in decision-making tasks that require the input of higher cognitive abilities, such as sensory perception, attention to resources and memory.

In the prior art, most of researches on perception decision of Alzheimer's disease patients only stay in a single visual channel, such as training individuals on a visual level. However, such single channel training is too limited to effectively improve the perceptual decision-making ability of the individual.

Disclosure of Invention

In view of this, the present application provides a training method, a training apparatus, a training device, and a storage medium, which can effectively improve the perceptual decision-making capability of an individual.

In a first aspect, the present application provides a training method, comprising: randomly displaying a perception decision task, wherein the perception decision task comprises a visual classification task, an auditory classification task and a visual-auditory classification task, the visual classification task comprises respectively classifying M first pictures, the auditory classification task comprises respectively classifying N first sounds, the visual-auditory classification task comprises respectively classifying L audiovisual stimulus pairs, each audiovisual stimulus pair comprises a second picture and a second sound corresponding to a target in the second picture, and M is more than or equal to 2,N and more than or equal to 2,L and more than or equal to 2; collecting behavioral response data generated when a user completes the perception decision task; and determining a training result according to the behavior response data, wherein the training result comprises the accuracy of the perception decision task completed by the user.

In a possible implementation manner, the behavioral response data includes classification results corresponding to each classification task in the perceptual decision task and response time for completing each classification task, and the training method further includes: inputting the classification result and the reaction time into a preset drift diffusion model for processing to obtain a drift rate, a decision boundary and non-decision time; and evaluating the perception decision-making capability of the user according to the drift rate, the decision boundary and the non-decision time.

In one possible implementation, the training method further includes: and determining the health state of the user according to the perception decision-making capability of the user.

In one possible implementation, the training method further includes: acquiring M preset pictures; adjusting the basic attribute of each preset picture to obtain M first pictures; and constructing the visual classification task according to the M first pictures.

In one possible implementation, the training method further includes: acquiring N preset sounds; adjusting the sound attribute of each preset sound to obtain N first sounds; an auditory classification task is constructed from the N first sounds.

In one possible implementation, the training method further includes: determining L second pictures in the M first pictures; determining L second sounds in the N first sounds; matching the L second pictures with the L second sounds to obtain L audiovisual stimulation pairs; and constructing a visual-auditory classification task according to the L visual-auditory stimulation pairs.

In one possible implementation, the training method further includes: determining a stimulation intensity corresponding to each first picture and each first sound respectively, wherein the stimulation intensity is used for reflecting the respective corresponding correct rate when each first picture and each first sound are classified;

selecting a picture with the stimulation intensity as a first stimulation intensity and a picture with the stimulation intensity as a second stimulation intensity from the M first pictures;

selecting a sound with the first stimulation intensity and a sound with the second stimulation intensity from the N first sounds;

constructing a perception decision task of first stimulation intensity according to the picture of the first stimulation intensity and the sound of the first stimulation intensity;

and constructing a perception decision task of the second stimulation intensity according to the picture of the second stimulation intensity and the sound of the second stimulation intensity.

In a second aspect, the present application provides a training apparatus comprising:

the display unit is used for randomly displaying a perception decision task, the perception decision task comprises a visual classification task, an auditory classification task and a visual-auditory classification task, the visual classification task comprises respectively classifying M first pictures, the auditory classification task comprises respectively classifying N first sounds, the visual-auditory classification task comprises respectively classifying L visual-auditory stimulation pairs, each visual-auditory stimulation pair comprises a second picture and a second sound corresponding to a target in the second picture, and M is not less than 2,N and not less than 2,L and not less than 2;

the acquisition unit is used for acquiring behavioral response data generated when a user completes a perception decision task;

and the determining unit is used for determining a training result according to the behavior response data, wherein the training result comprises the accuracy of the perception decision task completed by the user.

In a third aspect, the present application provides a training apparatus, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the training method according to any one of the above-mentioned first aspects when executing the computer program.

In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the training method of any of the above-mentioned first aspects.

In a fifth aspect, embodiments of the present application provide a computer program product, which when run on a processor, causes the processor to execute the training method according to any one of the above-mentioned first aspects.

The training method provided by the application randomly shows the perception decision task to the user, and trains the user based on the perception decision task. In the training process, behavior response data generated when the user completes the perception decision task is collected, and a training result, such as the accuracy of the perception decision task completed by the user, can be determined according to the behavior response data. Because the perception decision task comprises classification tasks on a plurality of channels of vision, hearing and vision and hearing, the perception decision task is used for training a user, so that information storage and coding of the user in a high-order cognition process can be accelerated, the reaction speed of the user can be improved, the formation of perception decision can be promoted, and the perception decision capability of an individual can be effectively improved.

Drawings

FIG. 1 is a schematic flow chart diagram of a training method provided by an exemplary embodiment of the present application;

fig. 2 is a schematic diagram of a first picture provided in an embodiment of the present application;

FIG. 3 is a schematic diagram of a first sound provided by an embodiment of the present application;

FIG. 4 is a schematic diagram of a pair of audio-visual stimuli provided by an embodiment of the present application;

FIG. 5 is a detailed flow chart of a training method shown in yet another exemplary embodiment of the present application;

FIG. 6 is a detailed flow chart of a training method shown in yet another exemplary embodiment of the present application;

FIG. 7 is a schematic view of an exercise device provided in an embodiment of the present application;

fig. 8 is a schematic diagram of a training apparatus according to another embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

Individuals experience some degree of cognitive decline during aging, and Alzheimer's Disease (AD) is a progressive neurodegenerative disease with occult onset. Clinically, it is manifested as generalized dementia such as dysmnesia, aphasia, disuse, agnosia, visual-spatial skill impairment, executive dysfunction, and personality and behavioral changes.

Perceptual decision-making is a continuous layered cognitive operation that translates perceptual information into target oriented and reacts, and includes decision-making from the encoding of sensory information (e.g., information generated by an objective that acts directly on the sense organ), the accumulation of decision information, and the application of decision rules to the final generation of behavioral responses. For example, a user sees a picture, judges that the content in the picture is an animal, and selects an animal option from preset options, and the whole process is called perceptual decision.

For the elderly, alzheimer's disease reduces their perceptual decision-making abilities, resulting in poor performance on decision-making tasks that require high-level cognitive abilities, such as sensory perception, attention, and memory.

In view of this, the present application provides a training method, a training apparatus, a training device and a storage medium. The user is trained by randomly presenting a perceptual decision task to the user and based on the perceptual decision task. In the training process, behavior response data generated when the user completes the perception decision task is collected, and a training result, such as the accuracy of the perception decision task completed by the user, can be determined according to the behavior response data. Because the perception decision task comprises classification tasks on a plurality of channels of vision, hearing and vision and hearing, the perception decision task is used for training a user, so that information storage and coding of the user in a high-order cognition process can be accelerated, the reaction speed of the user can be improved, the formation of perception decision can be promoted, and the perception decision capability of an individual can be effectively improved.

The technical solution of the present application will be described in detail below with specific examples. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments.

The embodiment of the application provides training software. The training software can be installed in training equipment, and the training equipment can be equipment capable of displaying pictures and having an audio playing function, such as intelligent mobile phones, tablet computers, desktop computers, notebook computers, robots, intelligent wearable equipment and the like. The training software provided by the application can be used for training a user and testing the perception decision-making capability of the user before or after training.

Referring to fig. 1, fig. 1 is a schematic flow chart of a training method according to an exemplary embodiment of the present application. The training method as shown in fig. 1 may include: s101 to S103 are as follows:

s101: and randomly displaying the perception decision task.

The perceptual decision task includes a visual classification task, an auditory classification task, and a visual-auditory classification task.

The visual classification task comprises the step of respectively classifying M first pictures, wherein M is more than or equal to 2. Where M represents the number of first pictures, e.g., M may be a positive integer greater than or equal to 2. The first picture may be a picture containing any object, for example, the first picture may be a picture containing a face, a picture containing a car, a picture containing an animal, a picture containing a plant, a picture containing a building, a picture containing food, a picture containing a commodity, a picture containing an electronic device, a picture containing a musical instrument, or the like. Different kinds of first pictures can be added according to the actual training requirement, which is only an exemplary illustration here, and is not limited to this.

Referring to fig. 2, fig. 2 is a schematic view of a first picture provided in the present application. As shown in fig. 2, fig. 2 illustrates a first picture in the visual classification task, the first picture being a picture containing a face.

The acquisition channel of the first picture is not limited. For example, the first picture may be obtained by shooting, by collecting in a network, by drawing, or the like.

The auditory classification task comprises the step of classifying N first sounds respectively, wherein N is larger than or equal to 2. Where N represents the number of first sounds, e.g., N may be a positive integer greater than or equal to 2. The first sound may be audio including any sound, for example, the first sound may be audio including a human sound, audio including a car sound, audio including an animal sound, audio including an electronic device sound, audio including an instrument sound, or the like. Different kinds of first sounds may be added according to actual training requirements, which is only an exemplary illustration and is not limited in this respect.

Referring to fig. 3, fig. 3 is a schematic diagram of a first sound provided by the embodiment of the present application. As shown in fig. 3, fig. 3 illustrates a first sound in the auditory classification task, which is audio containing the sound of a character, particularly a girl speaking sound.

The acquisition channel of the first sound is not limited. For example, the first sound may be obtained by recording, may be collected in a network, and the like.

The visual-auditory classification task comprises the step of respectively classifying L visual-auditory stimulation pairs, wherein each visual-auditory stimulation pair comprises a second picture and a second sound corresponding to a target in the second picture, and L is larger than or equal to 2. Where L represents the number of pairs of audiovisual stimuli, e.g., L may be a positive integer greater than or equal to 2.

The second picture may be a picture containing an arbitrary object, for example, the second picture may be a picture containing a face, a picture containing a car, a picture containing an animal, a picture containing a musical instrument, or the like. Accordingly, the second sound may be audio containing a human voice, audio containing a car voice, audio containing an animal voice, audio containing an instrument voice, or the like.

Illustratively, the picture containing the face and the audio of the sound corresponding to the face form an audiovisual stimulation pair, the picture containing the car and the audio of the sound corresponding to the car form an audiovisual stimulation pair, the picture containing the animal and the audio of the sound corresponding to the animal form an audiovisual stimulation pair, and the picture containing the musical instrument and the audio of the sound corresponding to the musical instrument form an audiovisual stimulation pair.

Referring to fig. 4, fig. 4 is a schematic diagram of an audiovisual stimulation pair provided in an embodiment of the present application. As shown in fig. 4, fig. 4 illustrates an audiovisual stimulus pair in the audiovisual classification task, the audiovisual stimulus pair comprising a second picture and a second sound corresponding to the target in the second picture. The second picture is a picture containing a car, and the second sound is audio containing car sound, specifically audio containing car whistling sound.

The channel of acquisition of the pair of audiovisual stimuli is not limited. For example, the second picture of the pair of audiovisual stimuli may be selected from the first picture, may be retaken, may be gathered in a network, or may be drawn. The second sound in the pair of audiovisual stimuli may be selected from the first sound, recorded, or gathered in a network.

In one possible implementation, after the user starts the training software installed on the training device, the user selects a training option and starts to randomly display a perceptual decision task in a display interface of the training device. When the user selects the training option, the user can click manually, can operate by remote control, and can control by voice.

Illustratively, the display interface of the training device centrally presents a gaze point, the presentation duration of which can be set by itself, for example, 2000ms, and then starts to randomly present the visual classification task, the auditory classification task, and the visual-auditory classification task. One way of presentation may be to present one of the tasks after the presentation of the other task is completed until all the tasks are presented. For example, a visual classification task is displayed first, after all the M first pictures in the visual classification task are displayed, an auditory classification task is displayed, and after all the N first sounds in the auditory classification task are displayed, a visual-auditory classification task is displayed until all the L visual-auditory stimulation pairs in the visual-auditory classification task are displayed.

It should be noted that the display order of the visual classification task, the auditory classification task, and the visual and auditory classification task is not limited. For example, the presentation sequence may be a visual classification task, an auditory classification task, a visual-auditory classification task, an auditory classification task, or the like, which is not limited thereto.

Another exemplary presentation manner may be that a visual classification task, an auditory classification task, and a visual and auditory classification task are alternately presented, that is, M first pictures, N first sounds, and L pairs of audiovisual stimuli are alternately presented until all tasks are presented. For example, a plurality of first pictures are displayed, a plurality of audiovisual stimulus pairs are displayed, a plurality of first sounds are displayed, a plurality of first pictures are displayed, a plurality of sounds are displayed, and the like until all tasks are displayed.

For another example, a first picture is displayed, followed by a first sound, followed by a pair of audiovisual stimuli, followed by a first sound, followed by a first picture, and so on. This is merely an example and is not intended to be limiting.

It should be noted that, in order to ensure the effectiveness of the training result, when the first picture in the visual classification task and the second picture in the visual and auditory classification task are displayed, the pictures are all presented on a background with uniform color, and the presented visual angles are the same. For example, all appear on a background of gray, white, etc., and all appear at viewing angles of 8 ° × 8 °. The description is given for illustrative purposes only and is not intended to be limiting.

S102: and acquiring behavioral response data generated by the user when the perception decision task is completed.

Illustratively, in a display interface of the training device, in addition to displaying each perceptual decision task, options corresponding to each perceptual decision task are also displayed, in the process of randomly displaying each perceptual decision task, a user performs selection operations on each classification task, data generated by the selection operations are behavior response data, and the training device collects the behavior response data.

For example, what is currently shown in the display interface is the first picture in the visual classification task, and two options are displayed side by side below, or above, or to the right, or to the left of the first picture. When the first picture is a picture containing a face, the right choice is to click and display the left option in the two options side by side; when the first picture is a picture containing a car, the right option of the two options is correctly selected to be clicked and displayed side by side.

As another example, the current training device plays the first sound in the auditory classification task and the two options are displayed side-by-side in the display interface. When the first sound is a physical sound, the right choice is to click and display the left option of the two options side by side; when the first sound is a car sound, the right option of the two options is clicked and displayed side by side. It is worth mentioning that the distance between the user and the training device can be set and adjusted by the user during the training process. For example, the user is 60 centimeters from the display interface and the speaker.

In the training process, the user makes different selections for different perception decision tasks according to the own ability, namely, selects the option which the user thinks is correct. The training device records the user's selections made for each classification task.

In a possible implementation manner, the user performs the selection operation on each classification task, and the selection operation can be realized through a mouse. For example, for a first picture containing a face, a first sound containing a character sound, a second picture containing a face, and a second sound containing a character sound corresponding to the second picture, correct selections of these classification tasks are all clicking a left mouse button. For a first picture containing a car, a first sound containing the sound of the car, a second picture containing the car and a second sound corresponding to the second picture and containing the sound of the car, correct selection of the classification tasks is to click a right mouse button.

For example, a second picture in the visual and auditory classification task is displayed in the current display interface, and the training device plays a second sound corresponding to the second picture. The second picture is a second picture containing a face, the second sound is a second sound containing a character sound, and the right selection is to click a left mouse button.

In the training process, the user makes different selections for different perception decision tasks according to own ability, namely, the user clicks a left button or a right button of a mouse. The training device records the user's selections made for each classification task.

Optionally, in a possible implementation manner, in order to ensure the effectiveness of the training, when the perceptual decision task is displayed, two adjacent classification tasks may be displayed based on a preset time interval, and the display duration of each classification task is a preset duration. For example, the preset time interval between two adjacent classification tasks may be 1200 to 1500ms, and the presentation time duration of each classification task may be 300ms.

For example, the preset time interval between two adjacent pairs of audiovisual stimuli may be 1200 to 1500ms, and the presentation time duration of each pair of audiovisual stimuli may be 300 to 500ms. The description is given for illustrative purposes only and is not intended to be limiting.

S103: and determining a training result according to the behavior reaction data.

The training results include the accuracy with which the user completed the perceptual decision task.

Illustratively, the behavior reaction data is data generated by a user performing selection operation on each classification task, the behavior reaction data corresponding to each classification task is compared with the correct selection corresponding to the task, and a training result is determined according to the comparison result.

For example, each correct choice is scored, and if no choice or wrong choice is scored, a score can be obtained according to the behavior response data of the user, and the proportion of the score to the total score (the corresponding score when all tasks are selected correctly) is calculated to obtain the correct rate of the perception decision task completed by the user. The description is given for illustrative purposes only and is not intended to be limiting.

In the embodiment, the perceptual decision task is randomly displayed to the user, and the user is trained based on the perceptual decision task. In the training process, behavior response data generated when the user completes the perception decision task is collected, and a training result, such as the accuracy of the perception decision task completed by the user, can be determined according to the behavior response data. Because the perception decision task comprises classification tasks on a plurality of channels of vision, hearing and vision and hearing, the perception decision task is used for training a user, so that information storage and coding of the user in a high-order cognition process can be accelerated, the reaction speed of the user can be improved, the formation of perception decision can be promoted, and the perception decision capability of an individual can be effectively improved.

Referring to fig. 5, fig. 5 is a detailed flowchart of a training method according to another exemplary embodiment of the present application, where the training method shown in fig. 5 may include: s201 to S205 are as follows:

s201: and randomly displaying the perception decision task.

S202: and acquiring behavioral response data generated when the user completes a perception decision task.

S203: and determining a training result according to the behavior reaction data.

S201 to S203 are identical to S101 to S103 in the embodiment corresponding to fig. 1, and refer to the description of S101 to S103 in the embodiment corresponding to fig. 1, which is not repeated herein.

The behavior response data comprises classification results corresponding to all classification tasks in the perception decision task and response time for completing all classification tasks.

Illustratively, the classification result corresponding to each classification task is the selection operation made by the user. Such as the user clicking the left of the two options displayed side-by-side, the user clicking the right of the two options displayed side-by-side, the user clicking the left mouse button, and the user clicking the right mouse button.

The reaction time to complete each classification task is determined by the time at which the presentation of each classification task is started and the time at which the user makes a selection. For example, for a certain classification task, timing is started when the classification task is displayed, the user finishes immediately after making a selection, and the recorded time is the reaction time corresponding to the classification task.

S204: and inputting the classification result and the reaction time into a preset drift diffusion model for processing to obtain a drift rate, a decision boundary and non-decision time.

The preset drift diffusion model simulates the decision process in the classification task, each selection of the user is represented as an upper boundary and a lower boundary, and the perception decision process continuously accumulates evidence over time until the evidence reaches one of the two boundaries, and then causes corresponding behavior reaction.

The drift rate, decision boundary and non-decision time are different parameters obtained by processing the classification result and the reaction time by the drift diffusion model. These different parameters respectively map the cognitive processes behind the perceptual decision process behavior. In particular, drift rate is used to describe the speed of information accumulation, decision boundaries are used to describe the reaction boundaries that need to be reached before a reaction is made, and non-decision times are used to describe the times used to describe sensory coding and motor reactions.

The distribution of different reactions can influence the numerical value of each parameter in the drift diffusion model, so that the potential cognitive process of a user in the process of cross-channel perception decision can be reflected by calculating the specific parameters of the drift diffusion model under different conditions, and the training effect of the user can be determined.

In a possible implementation manner, in order to prevent a user from quickly guessing and determining a classification result and further causing a deviation of a training result, before inputting the classification result and the reaction time into a preset drift diffusion model for processing, data with the reaction time less than the preset reaction time is removed. For example, data with reaction times less than 300ms are rejected.

Optionally, after the data with the reaction time less than the preset reaction time are removed, the standard deviation can be calculated according to all the remaining reaction times, and then the data with the reaction time exceeding the preset standard deviation range are removed. For example, data with reaction times outside of plus or minus 2.5 standard deviations are discarded. The description is given for illustrative purposes only and is not intended to be limiting.

The function to which the preset drift diffusion model is applied is as follows:

in the above equation (1), f (t) is a conditional probability distribution with respect to t. The function f (t | v, a, z) can be split into two parts, i.e. f (t) prior and f (t) likelihood, according to Bayesian theory. Prior refers to the subjective guess of probability distribution by a user without knowing the parameters of the drift diffusion model, and likelihood refers to the parameters of the drift diffusion model calculated under the condition of obtaining the probability distribution of behavior reaction data.

Therefore, the emphasis of the drift diffusion model is to find the parameter values under the likelihood condition. Since the complexity of the formula cannot directly find the parameter value, markov Chain Monte Carlo (MCMC) is required. The MCMC algorithm may obtain the functional features by means of successive sampling, whereby the parameters of the population are deduced from the samples. Therefore, the likelihood part in bayes is calculated by the MCMC algorithm, thereby estimating the parameter distribution.

Illustratively, an HDDM toolkit in computer programming language (python) may be employed that provides a hierarchical bayesian parameter estimation of the drift-diffusion model, allowing simultaneous estimation of each tested drift-diffusion model parameter, resulting in drift rate, decision boundaries and time-to-no-decision.

Optionally, the classification result and the reaction time are input into a preset drift diffusion model for processing, and besides the drift rate, the decision boundary and the non-decision time, parameters such as a relative starting point, inter-training variation of the drift rate, inter-training variation of the non-decision time and the like can be obtained.

Where relative starting points are used to describe the starting preference for reaction selection. The inter-training variation from a starting point is represented as a uniformly distributed range of average relative starting points, which is used to describe the distribution of actual starting points for a particular training. The inter-training variation in drift rate is represented as a positive-too-distributed standard deviation, with the mean being the drift rate, used to describe the actual drift rate distribution for a particular training. The inter-training variation of the non-decision time appears as a uniformly distributed range of average non-decision times for describing the distribution of actual non-decision times in the training.

S205: and evaluating the perception decision-making capability of the user according to the drift rate, the decision boundary and the non-decision time.

Illustratively, parameters such as drift rate, decision boundary, and non-decision time correspond to different index ranges respectively. For example, the index range corresponding to the drift rate may be greater than-5 and less than 5, the index range corresponding to the decision boundary may be greater than 0.5 and less than 2, and the index range corresponding to the non-decision time may be greater than 0.1 and less than 0.5.

Illustratively, if the drift rate, the decision boundary and the non-decision time of the user are all in the respective corresponding index ranges, the perceptual decision-making capability of the user is evaluated to be strong. And if two of the drift rate, the decision boundary and the non-decision time of the user are in the respective corresponding index ranges, estimating the perception decision capability of the user and the like. And if the drift rate, the decision boundary and the non-decision time of the user are within the corresponding index range, or the drift rate, the decision boundary and the non-decision time of the user are not within the corresponding index range, evaluating the perception decision capability of the user. The description is given for illustrative purposes only and is not intended to be limiting.

In this embodiment, the classification result and the reaction time of the user are processed through a preset drift diffusion model, so as to obtain a drift rate, a decision boundary and a non-decision time. The potential cognitive process of the user in the cross-channel perceptual decision process can be accurately reflected through parameters such as drift rate, decision boundary and non-decision time, and then the drift rate, the decision boundary and the non-decision time are analyzed, so that the perceptual decision capability of the user can be accurately evaluated.

Referring to fig. 6, fig. 6 is a detailed flowchart of a training method according to still another exemplary embodiment of the present application, where the training method shown in fig. 6 may include: s301 to S306. It should be noted that S301 to S305 in this embodiment are completely the same as S201 to S205 in the embodiment corresponding to fig. 5, and specific reference is made to the description of S201 to S205 in the embodiment corresponding to fig. 5, which is not repeated in this embodiment. S306 specifically comprises the following steps:

s306: and determining the health state of the user according to the perception decision-making capability of the user.

Illustratively, some disorders may reduce the perceptual decision-making ability of the user. For example, for the elderly, alzheimer's disease reduces their perceptual decision-making ability. And acquiring the perception decision-making capability of the healthy user, and taking the perception decision-making capability of the healthy user as a reference. And comparing the perception decision-making capability of the user obtained in the training process with the perception decision-making capability of the healthy user, and determining the health state of the user in the training according to the comparison result.

For example, the perceptual decision-making capability of a healthy user is strong, the perceptual decision-making capability of the user obtained in the training process is poor, and the health state of the user in the training is determined to be a non-health state. Specifically, the user of the training may be determined to be an alzheimer patient. The description is given for illustrative purposes only and is not intended to be limiting.

In the embodiment, the health state of the user can be accurately determined by comparing the perceptual decision-making capability of the user who trains this time with the perceptual decision-making capability of the healthy user. For example, accurate and timely discovery of Alzheimer's disease patients is facilitated, so that the Alzheimer's disease patients can be treated as early as possible.

Optionally, in a possible implementation manner, before randomly exhibiting the perceptual decision task, the training method provided by the present application may further include: acquiring M preset pictures; adjusting the basic attribute of each preset picture to obtain M first pictures; and constructing a visual classification task according to the M first pictures.

Exemplarily, the preset picture refers to an original first picture. Basic properties may include spatial frequency, contrast, brightness, pixels, size, sharpness, format, etc. of the picture. For example, several preset pictures are obtained, wherein one half of the pictures are pictures containing faces, and the other half of the pictures are pictures containing cars. The spatial frequency, contrast, brightness and pixel adjustment of these pictures are made uniform. For example, the pixels may all be adjusted to 670 × 670 pixels.

After the spatial frequency, the contrast, the brightness and the pixels are adjusted to be consistent, the definition of each picture is adjusted by adopting preset software (such as Matlab software) through a signal-to-noise ratio. For example, the sharpness of each picture is adjusted to 8 different levels of 30%, 32.5%, 35%, 37.5%, 40%, 42.5%, 45%, 50%, respectively, by the signal-to-noise ratio.

After the adjustment, M first pictures, for example, 240 first pictures, are obtained. And setting a correct option for each first picture according to the picture content specifically contained in each first picture, wherein if the correct selection corresponding to the first picture is to click and display the left option in the two options side by side, or the correct selection corresponding to the first picture is to click and display the right option in the two options side by side, or the correct selection corresponding to the first picture is to click a left mouse button, or the correct selection corresponding to the first picture is to click a right mouse button, and the like. And constructing and obtaining a visual classification task according to each first picture and the correct option corresponding to each first picture.

In the embodiment, the obtained first pictures are all adjusted in basic attributes, so that training deviation caused by the difference of the basic attributes of the pictures is effectively avoided, the basic attributes of the pictures are ensured not to influence the selection of a user, and the accuracy of a training result is improved.

Optionally, in a possible implementation manner, before randomly exhibiting the perceptual decision task, the training method provided by the present application may further include: acquiring N preset sounds; adjusting the sound attribute of each preset sound to obtain N first sounds; an auditory classification task is constructed from the N first sounds.

Illustratively, the preset sound refers to an original first sound. The sound properties may include the frequency, pitch, loudness, timbre, etc. of the sound. For example, several preset sounds are obtained, wherein half of the sounds are human sounds and the other half of the sounds are automobile sounds. The loudness and frequency of these sounds are adjusted to be consistent. For example, preset software (such as Matlab software) is used to normalize the sounds, and the loudness and frequency adjustment of the processed sounds are consistent. And then, the processed sounds are embedded in white noises with different loudness by using voice synthesis software to obtain first sounds with different signal-to-noise ratios.

For example, the loudness of the processed sound may be reduced to 50%, and the loudness-adjusted sound may be embedded in 8 white noises with different loudness by using the speech synthesis software, so as to obtain a plurality of first sounds with signal-to-noise ratios of 12.5%, 25%, 37.5%, 50%, 62.5%, 75%, 87.5%, and 100%, respectively. The loudness of the first sounds is uniform, e.g., 60dB for each of the first sounds.

The sharpness of each picture is adjusted by the signal-to-noise ratio. For example, the sharpness of each picture is adjusted to 8 different levels of 30%, 32.5%, 35%, 37.5%, 40%, 42.5%, 45%, 50%, respectively, by the signal-to-noise ratio.

After the adjustment, N first sounds, for example, 240 first sounds, are obtained. Setting a correct option for each first sound according to the sound content specifically contained in each first sound, for example, if the correct selection corresponding to the first sound is to click and display the left option in the two options side by side, or the correct selection corresponding to the first sound is to click and display the right option in the two options side by side, or the correct selection corresponding to the first sound is to click the left button of a mouse, or the correct selection corresponding to the first sound is to click the right button of the mouse, or the like. And constructing and obtaining an auditory sense classification task according to each first sound and the correct option corresponding to each first sound.

In the embodiment, the obtained first sounds are adjusted in sound attributes, so that training deviation caused by differences of the sound attributes is effectively avoided, the sound attributes of the sounds are guaranteed not to influence the selection of a user, and the accuracy of a training result is improved.

Optionally, in a possible implementation manner, before randomly exhibiting the perceptual decision task, the training method provided by the present application may further include: determining L second pictures in the M first pictures; determining L second sounds in the N first sounds; matching the L second pictures with the L second sounds to obtain L audiovisual stimulation pairs; and constructing a visual-auditory classification task according to the L visual-auditory stimulation pairs.

For example, the second picture of the pair of audiovisual stimuli may be selected from the first picture and the second sound of the pair of audiovisual stimuli may be selected from the first sound. For example, L first pictures are selected from the M first pictures, and the L first pictures are determined as L second pictures. And selecting L first sounds from the N first sounds, and determining the L first sounds as L second sounds.

It can be understood that, since the second sound is the sound corresponding to the target in the second picture, in order to increase the speed of constructing the audiovisual classification task, when the second picture is determined, the picture with the sound corresponding to the target in the second picture is selected. For example, a certain first picture is selected from the M first pictures as a picture including a car, the sound corresponding to the car is exactly present in the N first sounds, the selected first picture is determined as a second picture, and the sound corresponding to the car in the N first sounds is determined as a second sound. The description is given for illustrative purposes only and is not intended to be limiting.

And matching the L selected second pictures with the L selected second sounds to obtain L audiovisual stimulation pairs. And setting a correct option for each audio-visual stimulation pair according to each audio-visual stimulation pair, wherein the correct selection of the audio-visual stimulation pair is to click and display the left option in the two options side by side, or the correct selection of the audio-visual stimulation pair is to click and display the right option in the two options side by side, or the correct selection of the audio-visual stimulation pair is to click a left mouse button, or the correct selection of the audio-visual stimulation pair is to click a right mouse button, and the like. And constructing and obtaining the visual and auditory classification task according to each visual and auditory stimulation pair and the correct options of each visual and auditory stimulation pair.

In the present embodiment, the second picture and the second sound in each of the pair of viewing stimuli are both selected from the first picture and the first sound. Since the fundamental property of the first picture and the sound property of the first sound are both adjusted, it is equivalent to the fundamental property of the second picture and the sound property of the second sound in each audiovisual stimulus pair being adjusted. Training deviation caused by the difference of the basic attribute and the sound attribute of each picture is effectively avoided, the basic attribute and the sound attribute of each picture are ensured not to influence the selection of a user, and therefore the accuracy of a training result is improved.

Optionally, in a possible implementation manner, in order to improve the accuracy of the training result, pre-training may be further included before the formal training. Specifically, the training method provided by the present application may further include: s401 to S405.

S401: and determining the stimulation intensity corresponding to each first picture and each first sound respectively.

S402: and selecting a picture with the stimulation intensity as the first stimulation intensity and a picture with the stimulation intensity as the second stimulation intensity from the M first pictures.

S403: and selecting the sound with the first stimulation intensity and the sound with the second stimulation intensity from the N first sounds.

The stimulation intensity is used for reflecting the respective corresponding correctness of each first picture and each first sound when being classified.

Illustratively, M first pictures are displayed in a display interface of the training device, a user performs selection operation on each first picture, and the selection operation corresponding to each first picture is compared with the correct selection corresponding to the first picture. And (4) scoring each correct selection, and if the correct selection is not selected or the selection is wrong and not scored, obtaining a score according to all selection operations of the user at this time, and calculating the proportion of the score to the total score (the corresponding score when all the first pictures are selected correctly) to obtain the accuracy of the pre-training of the user at this time.

And determining the stimulation intensity of the first picture selected by the user to be correct according to the correct rate. When the accuracy is a first threshold value, selecting the correct stimulation intensity of the first picture as a first stimulation intensity by the user; and when the accuracy is a second threshold value, the user selects the correct stimulation intensity of the first picture as a second stimulation intensity. Wherein the first threshold is greater than the second threshold and the first stimulation intensity is higher than the second stimulation intensity. For example, the first threshold is 90%, the second threshold is 70%, the first stimulus intensity is high, and the second stimulus intensity is low.

Illustratively, the accuracy of the pre-training is 90%, and the stimulation intensity of the user selecting the correct first picture is the first stimulation intensity, i.e. the high intensity. Or, the accuracy of the pre-training is 70%, and the stimulation intensity of the user selecting the correct first picture is the second stimulation intensity, that is, the low intensity. The description is given for illustrative purposes only and is not intended to be limiting.

Illustratively, N first sounds are displayed in a display interface of the training device, a user makes a selection operation on each first sound, and the selection operation corresponding to each first sound is compared with the correct selection corresponding to the first sound. And (4) scoring each correct selection, wherein the scores are not selected or not selected in error, a score can be obtained according to all the selection operations of the user at this time, and the proportion of the score to the total score (the corresponding score when all the first sounds are selected correctly) is calculated to obtain the accuracy of the pre-training of the user at this time.

Based on the accuracy, the stimulation intensity at which the user selects the correct first sound is determined. When the accuracy is a first threshold value, selecting the correct stimulation intensity of the first sound as a first stimulation intensity by the user; when the accuracy rate is the second threshold, the user selects the correct stimulation intensity of the first sound to be the second stimulation intensity. Wherein the first threshold is greater than the second threshold and the first stimulation intensity is higher than the second stimulation intensity. For example, the first threshold is 90%, the second threshold is 70%, the first stimulus intensity is high intensity, and the second stimulus intensity is low intensity.

Illustratively, the accuracy of the pre-training is 90%, and the stimulation intensity of the first sound selected by the user is the first stimulation intensity, i.e. the high intensity. Or, the accuracy of the pre-training is 70%, and the stimulation intensity of the first sound selected by the user is the second stimulation intensity, that is, the low intensity. The description is given for illustrative purposes only and is not intended to be limiting.

S404: and constructing a perception decision task of the first stimulation intensity according to the picture of the first stimulation intensity and the sound of the first stimulation intensity.

The perceptual decision task for the first stimulus intensity includes a visual classification task for the first stimulus intensity, an auditory classification task for the first stimulus intensity, and a visual-auditory classification task for the first stimulus intensity.

It is to be understood that the process of constructing the visual classification task of the first stimulus intensity, the auditory classification task of the first stimulus intensity, and the visual-auditory classification task of the first stimulus intensity is similar to the process of constructing the visual classification task, the auditory classification task, and the visual-auditory classification task described above.

The difference is that the visual classification task, the auditory classification task, and the visual-auditory classification task are constructed based on the first picture, the first sound, the second picture, and the second sound, and in the present embodiment, the picture based on the first stimulus intensity and the sound based on the first stimulus intensity are constructed. For the specific process, reference may be made to the process of constructing the visual classification task, the auditory classification task, and the visual and auditory classification task, which are not described herein again.

For example, the constructed visual classification task with the first stimulus intensity includes 50 pictures with the first stimulus intensity, the auditory classification task with the first stimulus intensity includes 50 sounds with the first stimulus intensity, and the visual-auditory classification task with the first stimulus intensity includes 50 pairs of audiovisual stimuli.

S405: and constructing a perception decision task of the second stimulation intensity according to the picture of the second stimulation intensity and the sound of the second stimulation intensity.

The perceptual decision task for the second stimulation intensity includes a visual classification task for the second stimulation intensity, an auditory classification task for the second stimulation intensity, and a visual-auditory classification task for the second stimulation intensity.

It is understood that the process of constructing the visual classification task of the second stimulus intensity, the auditory classification task of the second stimulus intensity, and the visual-auditory classification task of the second stimulus intensity is similar to the process of constructing the visual classification task, the auditory classification task, and the visual-auditory classification task described above.

The difference is that the visual classification task, the auditory classification task, and the visual-auditory classification task are constructed based on the first picture, the first sound, the second picture, and the second sound, and in the present embodiment, the picture based on the second stimulus intensity and the sound based on the second stimulus intensity are constructed. For the specific process, reference may be made to the above processes for constructing the visual classification task, the auditory classification task, and the visual and auditory classification tasks, which are not described herein again.

For example, the constructed visual classification task with the second stimulation intensity includes 50 pictures with the second stimulation intensity, the auditory classification task with the second stimulation intensity includes 50 sounds with the second stimulation intensity, and the visual-auditory classification task with the second stimulation intensity includes 50 audiovisual stimulation pairs.

Illustratively, different users (e.g., normal elderly and alzheimer patients) are trained using a perceptual decision task of a first constructed stimulation strength and a perceptual decision task of a second constructed stimulation strength. Collecting behavioral response data generated when a user completes a perception decision task of a first stimulation intensity and a perception decision task of a second stimulation intensity; and determining a target training result according to the behavior response data, wherein the target training result comprises the accuracy of the perception decision task with the first stimulation intensity and the accuracy of the perception decision task with the second stimulation intensity.

In the embodiment, the perception decision tasks with different stimulation intensities are constructed, the perception decision tasks with different stimulation intensities can be adopted for training aiming at different users, and the perception decision abilities of different users can be improved in a targeted manner.

Optionally, in a possible implementation manner, the training method provided by the present application may further include: and adjusting the difficulty of the perception decision task according to the training result, thereby more effectively improving the perception decision capability of the user.

For example, when the accuracy of the perceptual decision task completed by the user is greater than the preset accuracy, it is proved that the current training effect of the user is good, and the difficulty of the perceptual decision task can be increased. For example, the types of pictures and sounds in the perceptual decision task can be gradually increased, the preset time interval between two adjacent classification tasks can be shortened, and options corresponding to each classification task can be increased.

For another example, when the accuracy of the perceptual decision task completed by the user is less than or equal to the preset accuracy, it is proved that the current training effect of the user is not good, and the difficulty of the perceptual decision task can be reduced. For example, the types of pictures and sounds in the perceptual decision task may be reduced, the preset time interval between two adjacent classification tasks may be increased, and the like.

Optionally, in a possible implementation manner, the training method provided by the application can also use a competition model to study the influence of cross-channel on alzheimer patients. When both visual and auditory information occur, the individual reacts more rapidly than when the information is single channel information (e.g., visual or auditory). This phenomenon is called Redundant Signal Effect (RSE).

RSE can be explained by statistical stimulation (statistical stimulation), i.e. the individual reacts to a single-channel stimulation (visual stimulation, auditory stimulation) of the multi-sensory-channel stimulation (visual-auditory stimulation) that reaches the sensory threshold first, resulting in a faster response to stimulation even without integration of the dual-channel information. Through the training in the multi-channel combination mode, an individual can reach a sensation threshold limit in multi-sensation-channel stimulation (visual and auditory stimulation), and therefore the sensation decision-making capability of the individual is improved.

Fig. 7 is a schematic view of a training device provided in the embodiment of the present application, and as shown in fig. 7, the training device provided in the embodiment includes:

a presentation unit 510, configured to randomly present a perceptual decision task, where the perceptual decision task includes a visual classification task, an auditory classification task, and a visual-auditory classification task, the visual classification task includes respectively classifying M first pictures, the auditory classification task includes respectively classifying N first sounds, the visual-auditory classification task includes respectively classifying L audiovisual stimulus pairs, each audiovisual stimulus pair includes a second picture and a second sound corresponding to a target in the second picture, where M is greater than or equal to 2,N and greater than or equal to 2,L and greater than or equal to 2;

the acquisition unit 520 is used for acquiring behavioral response data generated when the perception decision task is completed by a user;

a determining unit 530, configured to determine a training result according to the behavior response data, where the training result includes a correct rate of the perceptual decision task completed by the user.

Optionally, the behavior response data includes classification results corresponding to each classification task in the perceptual decision task and response time for completing each classification task.

Optionally, the training apparatus further comprises:

the evaluation unit is used for inputting the classification result and the reaction time into a preset drift diffusion model for processing to obtain a drift rate, a decision boundary and non-decision time; and evaluating the perception decision-making capability of the user according to the drift rate, the decision boundary and the non-decision time.

Optionally, the training device further comprises:

and the state determining unit is used for determining the health state of the user according to the perception decision capability of the user.

Optionally, the training device further comprises:

the first construction unit is used for acquiring M preset pictures; adjusting the basic attribute of each preset picture to obtain M first pictures; and constructing the visual classification task according to the M first pictures.

Optionally, the training device further comprises:

the second construction unit is used for acquiring N preset sounds; adjusting the sound attribute of each preset sound to obtain N first sounds; constructing the auditory classification task according to the N first sounds.

Optionally, the training device further comprises:

a third constructing unit, configured to determine L second pictures in the M first pictures; determining L second sounds among the N first sounds; matching the L second pictures and the L second sounds to obtain L audiovisual stimulation pairs; and constructing the visual-auditory classification task according to the L visual-auditory stimulation pairs.

Optionally, the training device further comprises:

a third constructing unit, configured to determine a stimulation intensity corresponding to each of the first pictures and each of the first sounds, where the stimulation intensity is used to reflect a respective correctness rate when each of the first pictures and each of the first sounds are classified; selecting a picture with the stimulation intensity as a first stimulation intensity and a picture with the stimulation intensity as a second stimulation intensity from the M first pictures; selecting a sound with the first stimulation intensity and a sound with the second stimulation intensity from the N first sounds; constructing a perception decision task of first stimulation intensity according to the picture of the first stimulation intensity and the sound of the first stimulation intensity; and constructing a perception decision task of the second stimulation intensity according to the picture of the second stimulation intensity and the sound of the second stimulation intensity.

Referring to fig. 8, fig. 8 is a schematic view of a training apparatus according to another embodiment of the present application. As shown in fig. 8, the training apparatus 6 of this embodiment includes: a processor 60, a memory 61 and a computer program 62 stored in said memory 61 and executable on said processor 60. The processor 60, when executing the computer program 62, implements the steps in the above-described various training method embodiments, such as S101 to S103 shown in fig. 1. Alternatively, the processor 60, when executing the computer program 62, implements the functions of the units in the embodiments, such as the functions of the units 510 to 530 shown in fig. 7.

Illustratively, the computer program 62 may be divided into one or more units, which are stored in the memory 61 and executed by the processor 60 to accomplish the present application. The one or more units may be a series of computer instruction segments capable of performing specific functions, which are used to describe the execution of the computer program 62 in the training device 6. For example, the computer program 62 may be divided into a presentation unit, an acquisition unit, and a determination unit, each unit functioning as described above.

The training device may include, but is not limited to, a processor 60, a memory 61. It will be appreciated by those skilled in the art that fig. 8 is merely an example of a training device 6 and does not constitute a limitation of the device and may include more or less components than those shown, or some components in combination, or different components, for example the training device may also include input output devices, network access devices, buses, etc.

The Processor 60 may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 61 may be an internal storage unit of the training device, such as a hard disk or a memory of the device. The memory 61 may also be an external storage terminal of the training device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the training device. Further, the memory 61 may also include both an internal storage unit and an external storage terminal of the device. The memory 61 is used for storing the computer instructions and other programs and data required by the terminal. The memory 61 may also be used to temporarily store data that has been output or is to be output.

The embodiments of the present application further provide a computer storage medium, where the computer storage medium may be nonvolatile or volatile, and the computer storage medium stores a computer program, and when the computer program is executed by a processor, the computer program implements the steps in the training method embodiments.

The present application also provides a computer program product for causing a training apparatus to perform the steps of the above-described respective training method embodiments, when the computer program product is run on the training apparatus.

An embodiment of the present application further provides a chip or an integrated circuit, where the chip or the integrated circuit includes: and the processor is used for calling and running the computer program from the memory so that the training device provided with the chip or the integrated circuit executes the steps in each training method embodiment.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules, so as to perform all or part of the functions described above. Each functional unit and module in the embodiments may be integrated in one processing unit, or each unit may exist alone physically, or two or more units are integrated in one unit, and the integrated unit may be implemented in a form of hardware, or in a form of software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working processes of the units and modules in the system may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not cause the essential features of the corresponding technical solutions to depart from the spirit scope of the technical solutions of the embodiments of the present application, and are intended to be included within the scope of the present application.

Claims

1. A method of training, comprising:

randomly displaying a perception decision task, wherein the perception decision task comprises a visual classification task, an auditory classification task and a visual-auditory classification task, the visual classification task comprises respectively classifying M first pictures, the auditory classification task comprises respectively classifying N first sounds, the visual-auditory classification task comprises respectively classifying L visual-auditory stimulation pairs, each visual-auditory stimulation pair comprises a second picture and a second sound corresponding to a target in the second picture, and M is more than or equal to 2,N and more than or equal to 2,L and more than or equal to 2;

collecting behavioral response data generated by a user when the perception decision task is completed;

and determining a training result according to the behavioral response data, wherein the training result comprises the accuracy of the perception decision task completed by the user.

2. The training method according to claim 1, wherein the behavior response data includes classification results corresponding to each classification task in the perceptual decision task and response time for completing each classification task, and the training method further comprises:

inputting the classification result and the reaction time into a preset drift diffusion model for processing to obtain a drift rate, a decision boundary and non-decision time;

and evaluating the perception decision-making capability of the user according to the drift rate, the decision boundary and the non-decision time.

3. The training method of claim 2, wherein after evaluating the perceptual decision-making capability of the user based on the drift rate, the decision boundary, and the non-decision time, the training method further comprises:

and determining the health state of the user according to the perception decision-making capability of the user.

4. A training method as defined in claim 1, wherein prior to randomly exhibiting the perceptual decision task, the training method further comprises:

acquiring M preset pictures;

adjusting the basic attribute of each preset picture to obtain M first pictures;

and constructing the visual classification task according to the M first pictures.

5. The training method of claim 4, wherein prior to randomly exhibiting the perceptual decision task, the training method further comprises:

acquiring N preset sounds;

adjusting the sound attribute of each preset sound to obtain N first sounds;

constructing the auditory classification task according to the N first sounds.

6. A training method as defined in claim 5, wherein prior to randomly exhibiting the perceptual decision task, the training method further comprises:

determining L second pictures in the M first pictures;

determining L second sounds in the N first sounds;

matching the L second pictures and the L second sounds to obtain L audiovisual stimulation pairs;

and constructing the visual-auditory classification task according to the L visual-auditory stimulation pairs.

7. Training method according to any of claims 1 to 6, characterized in that it further comprises:

determining a stimulation intensity corresponding to each first picture and each first sound respectively, wherein the stimulation intensity is used for reflecting the respective corresponding correct rate when each first picture and each first sound are classified;

8. An exercise device, comprising:

the display unit is used for randomly displaying a perception decision task, the perception decision task comprises a visual classification task, an auditory classification task and a visual-auditory classification task, the visual classification task comprises respectively classifying M first pictures, the auditory classification task comprises respectively classifying N first sounds, the visual-auditory classification task comprises respectively classifying L audiovisual stimulus pairs, each audiovisual stimulus pair comprises a second picture and a second sound corresponding to a target in the second picture, and M is more than or equal to 2,N and more than or equal to 2,L and more than or equal to 2;

the acquisition unit is used for acquiring behavioral response data generated when the perception decision task is completed by a user;

9. Training device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 7.