Cognitive style test method and device based on VR

Abstract

The invention provides a cognitive style test method and device based on VR, a dynamic three-dimensional VR test environment is constructed, electroencephalogram, eye movement tracking and micro expression recognition technologies are integrated, and multi-dimensional synchronous monitoring of a cognitive state is realized; an intelligent analysis model driven by machine learning is developed, so that the cognitive style detection error rate is reduced to 5% or below; in the testing method, four-dimensional data of electroencephalogram (attention), eye movement (visual strategy), behavior (operation precision) and micro-expression (emotional stability) are combined for the first time, and the limitation of traditional single-dimensional evaluation is broken through; the weight is dynamically adjusted according to the task difficulty, and the method better fits the actual cognitive load change. The environmental interference is increased in the challenge mode, and the eye movement data can better reflect the anti-interference capability of the tested object; a logarithmic function is introduced to compress an extreme value, and robustness is improved; cognitive stability can be obviously influenced by excessive attention fluctuation, and nonlinear punishment is needed.

Description

Technical Field

[0001] The present invention belongs to the cross technical field of cognitive assessment experimental instruments and virtual reality technology, and particularly relates to a method and device for testing cognitive style based on VR. Background Art

[0002] Cognitive Style refers to the characteristics of self-consistency that develop in a suitable manner around inherent personality traits during the information processing process. Cognitive style is one of the important factors affecting students' performance in many subjects. Existing research shows that there is an inseparable relationship between a person's cognitive style and learning style. Cognitive style affects learners' preferences for learning environments and learning methods through learning style as a medium. There is a significant correlation between learners' field cognitive style, learning strategies, and learning motivation. In Rigorous Mathematical Thinking (RMT) for solving math problems, there are differences between students with reflective and impulsive cognitive styles. Junior and senior high school students with different cognitive styles also have significant differences in the learning of geography and physics. In solving reasoning problems, field-independent students have good performance in verbal reasoning tests under low anxiety levels, while field-dependent students have better performance in verbal reasoning tests under high anxiety levels. In addition, there is also research showing that in game-based learning tasks, learners with different cognitive styles have different learning outcomes under different information presentation forms. Thus, it can be seen that the learning effects of different students under different teaching designs are significantly different. Therefore, cognitive style tests are important tools in personalized teaching tests, and obtaining students' cognitive styles has a very important impact on the construction of personalized teaching platforms.

[0003] Cognitive style, also known as cognitive pattern, refers to the individual differences shown when people organize and process information and experience. It is the habitual attitude and style that individuals often adopt in the processes of perception, memory, and thinking. In recent years, many researchers have paid more attention to cognitive style, which has led to many valuable research topics. At the same time, it has also made how to measure cognitive style become a key issue. The cognitive style test is a measurement method commonly used by many scholars, and the instrument used in the cognitive style test is mainly the cognitive style tester. The cognitive style tester is a visual experimental instrument. It is a commonly used and relatively complete instrument in the experiments of field-independent and field-dependent cognitive styles. Its structure includes a bright frame and a bright rod with uniform brightness in a dark visual field background. The rod is inside the frame, and both can be adjusted clockwise or counterclockwise separately, and there is a reading disk that shows the tilt angles of the frame and the rod with pointers at any time. The cognitive style tester can measure the degree to which an inclined frame affects the judgment of the verticality of a rod. The judgment of the test subject is affected by the inclined frame, which is equivalent to the influence of changes in the surrounding environmental conditions. Therefore, the cognitive style tester can measure personality traits and the cognitive styles of different people through the cognitive style of the test subject.

[0004] The main products of cognitive style testers on the market currently are basically composed of the following: 1. An observation tube placed on a platform. 2. A panel perpendicular to the platform, on which there is a frame and a rod. The inclination of the rod and the frame can be adjusted by a knob. 3. There is a semicircular scale on the back of the panel, indicating the inclination of the frame and the rod. The minimum reading is 0.5°. 4. There is a level on the platform. After a large number of usage feedbacks, the following pain points have been summarized for the widely used BD-V-503 model of cognitive style tester: First, in terms of the front-end operation and hardware use of the cognitive style tester, this instrument has disadvantages such as a large appearance, inconvenient operation, and low data accuracy. Second, in terms of the back-end data collection and analysis of the cognitive style tester, there are also major drawbacks: Different from automated instruments, the data measured by the cognitive style tester needs to be manually recorded by the test subject or printed out as a data strip, which is prone to data loss and omission. And when analyzing the data, it needs to be manually input into the computer on a large scale, consuming time and energy. Moreover, for the data obtained from the experiment, this instrument cannot intuitively give the test subject the result analysis, is prone to personal judgment, and the result is not authoritative enough to achieve the experimental purpose well.

[0005] The new device is designed with three functional modules to measure the cognitive style tendency of the subjects: the cognitive style test module, the embedded figure test module, and the questionnaire module. Among them, the cognitive style test module retains the functions of the original instrument and measures by asking the subjects to adjust the rod according to their own feelings in a vertical and horizontal manner; the embedded figure test module selects and improves on the existing test questions to evaluate the level of the subjects' field-independent cognitive style with scores; the questionnaire module is a supplement and improvement to the above two functional modules, and asks the subjects to answer relevant questions about field independence and field dependence by designing a degree evaluation questionnaire.

[0006] First of all, in terms of the front-end operation and hardware use of the cognitive style tester, the existing instrument has disadvantages such as a large appearance, inconvenient operation, and low data accuracy. Secondly, in terms of data collection and analysis at the back end of the cognitive style tester, the existing instrument also has major drawbacks: different from automated instruments, the data measured by the cognitive style tester needs to be manually recorded by the subjects or printed out as data strips, which is prone to data loss and omission, and a large amount of manual input into the computer is required during data analysis, consuming time and energy. Moreover, for the data obtained from the experiment, the existing instrument cannot intuitively give the subjects the result analysis, which is prone to personal judgment, the results are not authoritative enough, and the experimental purpose cannot be well achieved. Summary of the Invention

[0007] In view of this, the purpose of the present invention is to provide a VR-based cognitive style test method and device.

[0008] A VR-based cognitive style test device includes a VR headset device, a multi-modal physiological acquisition module, and an intelligent analysis module;

[0009] The VR headset device is used to display a virtual cognitive test scene. Specifically, the scene includes a virtual rod and a rectangular frame. The rectangular frame is used to interfere with the tester's perception of adjusting the angle of the virtual rod relative to the vertical direction (i.e., the direction of gravity). The rectangular frame can be tilted at different angles relative to the virtual rod to provide different degrees of interference; the VR headset device is also used to identify the gestures of the tested person and drive the change of the virtual rod angle according to the gestures;

[0010] The multi-modal physiological acquisition module is equipped with an electroencephalograph and an eye tracker, which are used to collect electroencephalogram and eye movement data during the test of the tested person, and at the same time obtain micro-expression data and behavior data;

[0011] The intelligent analysis module uses the random forest algorithm to fuse electroencephalogram data, eye movement data, micro-expression data, and behavior data, and outputs a comprehensive field independence index FII for testing the cognitive style.

[0012] Preferably, the calculation formula of the comprehensive field independence index FII is as follows:

[0013] FII = w1·S EEG + w2·S Eye + w3·S Behavior + w4·S Expression

[0014] Wherein, S EEG , S Eye , S Behavior and S Expression respectively represent electroencephalogram signal score, eye movement tracking score, behavioral data score, and micro-expression score, and ω1, ω2, ω3, ω4 are the weight values corresponding to the above 4 scores respectively.

[0015] Preferably, the electroencephalogram signal score S EEG is as follows:

[0016]

[0017] α energy : The α-wave energy value of the Fp1 / Fp2 channels in the prefrontal region; β energy : The β-wave energy value of the Fp1 / Fp2 channels in the prefrontal region; Attention_Stability: The standard deviation of attention fluctuation throughout the test.

[0018] Preferably, the eye movement tracking score S Eye :

[0019]

[0020] Fixation_Density: The proportion of the high-density area in the fixation point heat map;

[0021] Saccade_Length: The average saccade path length;

[0022] Accuracy_Rate: The correct rate of the embedded figure test.

[0023] Preferably, the behavioral data score S Behavior :

[0024]

[0025] Angle_Error: The average deviation between the adjusted angle of the virtual rod and the vertical direction;

[0026] Max_Error: The preset maximum allowable deviation;

[0027] Time_Used: The task completion time;

[0028] Time_Threshold: The time limit threshold.

[0029] Preferably, the expression score S Expression :

[0030]

[0031] The proportion of neutral expression frames; Stress_Intensity: the intensity of stress expression, with a value range of 0-10 points.

[0032] Preferably, the test is divided into three difficulty levels: beginner mode, advanced mode, and challenge mode. The weight values in the three modes are as follows:

[0033]

[0034] Preferably, the beginner mode, advanced mode, and challenge mode are respectively:

[0035] In the beginner mode, the tilt angle of the rectangular frame controlled by the VR headset device ≤ 15°, with a time limit of 60 seconds;

[0036] In the advanced mode, the tilt angle of the rectangular frame controlled by the VR headset device changes between 15° and 30°, with a time limit of 45 seconds;

[0037] In the challenge mode, while the tilt angle of the rectangular frame controlled by the VR headset device changes, environmental interference is superimposed, with a time limit of 30 seconds.

[0038] Preferably, the method for testing cognitive style includes:

[0039] FII ≥ 80 is field-independent;

[0040] 60 ≤ FII < 80 is transitional;

[0041] FII < 60 is field-dependent.

[0042] A method for testing cognitive style, including:

[0043] (1) Device initialization and calibration

[0044] After starting the device, detect the connection status of the VR headset device, physiological sensor, and touch screen, and complete three-dimensional attitude calibration; after the user wears the device, the pupil tracking camera assists in aligning the human eye visual axis, and the electroencephalograph and eye tracker are calibrated respectively;

[0045] (2) Cognitive style test process

[0046] The test subject wears a VR headset, aligns their eyes with the VR endoscope tube scene to start the experiment; during the experiment, the test subject needs to adjust the virtual rod in the scene to the vertical state through their own perception; after multiple tests, the multi-modal physiological data of the test subject during each test is statistically analyzed, as well as the angle between the virtual rod and the vertical direction after the experiment. The test is carried out in a three-level difficulty mode.

[0047] (3) Data analysis and report generation

[0048] After the test is completed, the system calls a random forest model to integrate the multi-modal data and outputs the FII index and the cognitive style classification result.

[0049] The present invention has the following beneficial effects:

[0050] The present invention provides a VR-based cognitive style test method and device, constructs a dynamic three-dimensional VR test environment, integrates electroencephalogram, eye movement tracking and micro-expression recognition technologies, and realizes multi-dimensional synchronous monitoring of cognitive states; develops an intelligent analysis model driven by machine learning to reduce the cognitive style detection error rate to less than 5%; this device has the characteristics of small size (only requires a deployment space of 5m), high test efficiency (the completion time is shortened by 20%), and high degree of result visualization, and can be widely applied to the fields of educational evaluation, vocational selection and clinical diagnosis. In the test method, for the first time, four-dimensional data of electroencephalogram (attention), eye movement (visual strategy), behavior (operation accuracy), and micro-expression (emotional stability) are combined to break through the limitations of traditional single-dimensional evaluation. Cognitive style is a complex psychological trait that requires multi-angle cross-validation (for example, high field-independent individuals are usually accompanied by efficient visual search and stable emotions); the weights are dynamically adjusted according to the task difficulty, which is more in line with the actual cognitive load changes. Under the challenge mode, the environmental interference increases, and the eye movement data can better reflect the anti-interference ability of the test subject; the logarithmic function is introduced to compress extreme values to improve the robustness. Excessive attention fluctuations will significantly affect cognitive stability and require non-linear punishment. 2 Description of the drawings Description of the drawings

[0051] Figure 1 It is the external overall structure diagram of the intelligent cognitive style test instrument of the present invention;

[0052] Figure 2 It is the internal UI view of the intelligent cognitive style test instrument platform of the present invention. Detailed implementation manners

[0053] The following are specific embodiments in conjunction with the drawings to describe the present invention in detail.

[0054] This device adopts a modular design concept, integrates virtual reality technology, a multi-modal physiological signal acquisition system and an intelligent analysis algorithm, and breaks through the functional limitations of traditional instruments. Its core architecture consists of the following four parts:

[0055] 1. VR Interaction Module

[0056] The VR interaction module serves as the direct interface between the user and the system. As shown Figure 1 below, it mainly includes a VR headset device, which supports immersive three-dimensional scene interaction. Among them, the lens assembly of the VR headset device integrates a 32-step adjustable liquid crystal phase retarder (adjustment accuracy of λ / 100), which can dynamically compensate for ambient light interference. A pupil tracking camera (sampling rate of 120Hz) is also provided, which can detect the visual axis deviation of the human eye (accuracy of ±0.05°) in real time through an 850nm infrared fill light module.

[0057] The VR headset device is used to display a virtual cognitive test scene. Specifically, the scene includes a virtual rod and a rectangular frame. As shown Figure 2 below, the rectangular frame is used to interfere with the tester's perception of adjusting the angle of the virtual rod relative to the vertical direction (i.e., the direction of gravity). The rectangular frame can be tilted at different angles relative to the virtual rod to provide different degrees of interference.

[0058] The VR headset device is also used to recognize the gestures of the tested person and drive the change of the virtual rod angle according to the gestures. The task objective of the tested person is to control the angle of the virtual rod through gestures and make it as parallel to the vertical direction as possible.

[0059] 2. The touch interactive screen (6-inch capacitive screen) displays the deviation angle of the virtual rod (accuracy of ±0.1°), the environmental interference index, and the physiological index heat map in real time. Among them, the tested person can also drive the change of the virtual rod angle by controlling the touch interactive screen. The test scene developed based on the Unity engine in the present invention includes a dynamic feedback mechanism: when the deviation angle of the virtual rod > 5°, the interface displays a red gradient warning; when the deviation < 2°, it turns into a green prompt, accompanied by a score floating animation, which is used to provide an intuitive display for the tester and enhance the user interaction experience.

[0060] 3. Multimodal Physiological Acquisition Module

[0061] This module is deployed with an electroencephalograph (EEG) and an eye tracker, which are used to collect EEG and eye movement data respectively. At the same time, micro-expression and behavior data are integrated to achieve multi-dimensional synchronous monitoring of cognitive states. The EEG signal acquisition uses the NeuroSky MindWave EEG device (sampling rate ≥ 128 Hz), and focuses on extracting the energy ratio of alpha waves (8 - 12 Hz) and beta waves (12 - 30 Hz) in the Fp1 / Fp2 channels of the prefrontal lobe area to characterize the attention level. The eye movement tracking system (Tobii Pro, accuracy ≤ 0.5°) records the fixation heat map, saccade path length, and the correct rate of the embedded figure test to reflect the visual strategy efficiency. The OpenFace micro-expression recognition system analyzes facial feature points to quantify the proportion of neutral expression frames and the intensity of stress expressions (such as frowning) to evaluate emotional stability. The behavior data records operation accuracy indicators such as the angle adjustment error of the virtual stick and the task completion time through the touch screen.

[0062] 4. Intelligent Analysis Module

[0063] The intelligent analysis module uses the random forest algorithm to fuse EEG, eye movement, micro-expression, and behavior data, and outputs the Field Independence Index (FII). The algorithm design includes a dynamic weight adjustment mechanism: in the primary mode, it focuses on behavior data (weight 40%), in the advanced mode, it increases the weight of eye movement data (35%), and in the challenge mode, it combines EEG and micro-expression data (weight 30%) to enhance the anti-interference evaluation. By introducing a logarithmic function to compress extreme values, it suppresses the influence of excessive attention fluctuations on the score and improves the robustness of the model. Finally, the FII is divided into field-independent type (FII ≥ 80), transitional type (60 ≤ FII < 80), and field-dependent type (FII < 60), and a visual report is generated through the touch screen, including sub-item scores, heat maps, and teaching suggestions.

[0064] The test process of the cognitive style test device of the present invention is as follows:

[0065] (1) Device Initialization and Calibration

[0066] After starting the device, the system automatically detects the connection status of the VR headset, physiological sensors, and touch screen, and completes three-dimensional attitude calibration (residual error < 0.01° within 25 seconds) through a three-redundant IMU sensor (range ±16 g). After the user wears the device, the pupil tracking camera assists in aligning the human eye visual axis, and the EEG device and eye tracker are each calibrated for 10 seconds (fixating on a 5-point calibration pattern).

[0067] (2) Cognitive Style Test Process

[0068] The test subject wears a VR headset and aligns their eyes with the VR endoscope scene to start the experiment. During the experiment, the test subject needs to adjust the virtual rod in the scene to the vertical state through their own perception. After multiple tests, the multimodal physiological data of the test subject at each test, as well as the angle between the virtual rod and the vertical direction after the test, are statistically analyzed. The test is divided into three difficulty levels:

[0069] Beginner mode: The frame tilt angle ≤ 15°, time limit is 60 seconds. The user adjusts the virtual rod to the vertical state through the touch screen or gestures, and the system records the angle deviation and operation time.

[0070] Intermediate mode: Dynamically adjust the frame tilt angle (15° - 30°), time limit is 45 seconds, testing the user's ability to adapt to a dynamic environment.

[0071] Challenge mode: Adjust the frame tilt angle while superimposing environmental interference (such as particle effects), time limit is 30 seconds, evaluating the user's stress resistance performance under high interference.

[0072] (3) Data analysis and report generation

[0073] After the test is completed, the system calls a random forest model to integrate the multimodal data, and outputs the FII index and the cognitive style classification result. The visualization report shows the attention distribution, emotional fluctuations, and operation accuracy through a heat map, and generates a PDF document to provide personalized teaching suggestions (such as field-dependent learners need to strengthen environmental cue training).

[0074] Note: This device needs to be used in an environment with stable light and low electromagnetic interference. The operator needs to receive professional training to ensure the reliability of the data.

[0075] Among them, the specific evaluation algorithm is as follows:

[0076] 1. Calculate the comprehensive field independence index (FII)

[0077] FII = w1·S EEG + w2·S Eye + w3·S Behavior + w4·S Expression

[0078] Among them, S EEG 、S Eye 、S Behavior and S Expression represent the electroencephalogram signal score, eye movement tracking score, behavioral data score, and micro-expression score respectively. ω1, ω2, ω3, ω4 are the weight values corresponding to the above 4 scores respectively.

[0079] 2. Calculation formulas for each item:

[0080] 1). Electroencephalogram (EEG) signal score SEEG :

[0081]

[0082] α energy : The energy value of the α wave (8 - 12 Hz) in the Fp1 / Fp2 channels of the prefrontal region.

[0083] β energy : The energy value of the β wave (12 - 30 Hz) in the Fp1 / Fp2 channels of the prefrontal region.

[0084] Attention_Stability: The standard deviation of attention fluctuations (normalized) throughout the test.

[0085] 2), Eye movement tracking score S Eye :

[0086]

[0087] Fixation_Density: The proportion of the high - density area in the fixation point heat map (reflecting the information integration ability).

[0088] Saccade_Length: The average saccade path length (reflecting visual search efficiency).

[0089] Accuracy_Rate: The correct rate of the embedded figure test.

[0090] 3), Behavioral data score S Behavior :

[0091]

[0092] Angle_Error: The average deviation between the adjusted angle of the virtual rod and the vertical direction.

[0093] Max_Error: The preset maximum allowable deviation (such as 30°).

[0094] Time_Used: The task completion time.

[0095] Time_Threshold: The time limit threshold.

[0096] 4), Micro - expression score S Expression :

[0097]

[0098] The proportion of the number of neutral expression frames (reflecting emotional stability). Stress_Intensity: The intensity of stress expressions (such as frowning) (0 - 10 points).

[0099] 3. Weight Allocation and Standardization

[0100] Dynamic Weight Adjustment: Dynamically adjust the weights according to the test stages (primary / advanced / challenge mode):

[0101]

[0102] Reason: In the challenge mode, the environmental interference increases, and the weight of eye movement data is increased to reflect the ability to optimize visual strategies.

[0103] Data Standardization: Adopt Z-score normalization to ensure that the mean value of each sub-item score is 0 and the standard deviation is 1, avoiding the influence of dimensional differences.

[0104] Evaluation Hierarchy System (divide cognitive style types according to FII scores):

[0105] Field-independent type (FII≥80): High autonomous information processing ability, less affected by environmental interference.

[0106] Transitional type (60≤FII<80): Medium adaptability, partially dependent on environmental cues.

[0107] Field-dependent type (FII<60): Dependent on external references, easily affected by environmental interference.

[0108] Empirical Verification: The comparative experiment with the traditional BD-V-503 instrument shows that the error rate of the method of the present invention is reduced to 4.7% (12.3% for traditional equipment), and the classification consistency is increased by 28%.

[0109] In summary, the above are only the preferred embodiments of the present invention, and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A VR-based cognitive style test device, characterized in that It includes a VR headset device, a multi-modal physiological acquisition module, and an intelligent analysis module; The VR headset device is used to display a virtual cognitive test scenario. Specifically, the scenario includes a virtual rod and a rectangular frame. The rectangular frame is used to interfere with the tester's perception of adjusting the angle of the virtual rod relative to the vertical direction. The rectangular frame can be tilted at different angles relative to the virtual rod to provide different degrees of interference. The VR headset device is also used to recognize the gestures of the tested person and drive the change of the virtual rod angle according to the gestures; The multi-modal physiological acquisition module is equipped with an electroencephalograph and an eye tracker, which are used to collect electroencephalogram and eye movement data during the test of the tested person, and at the same time obtain micro-expression data and behavior data; The intelligent analysis module uses the random forest algorithm to fuse electroencephalogram data, eye movement data, micro-expression data, and behavior data, and outputs a comprehensive field independence index FII for testing cognitive styles.

2. The cognitive style test device based on VR according to claim 1, characterized in that, The calculation formula of the comprehensive field independence index FII is as follows: FII = w1·S EEG + w2·S Eye + w3·S Behavior + w4·S Expression Among them, S EEG , S Eye , S Behavior and S Expression respectively represent the electroencephalogram signal score, eye movement tracking score, behavioral data score, and micro-expression score, and ω1, ω2, ω3, and ω4 are the weight values corresponding to the above four scores respectively.

3. The cognitive style test device based on VR according to claim 2, characterized in that, EEG signal score S EEG is as follows: α energy : The α-wave energy value of the Fp1 / Fp2 channels in the prefrontal region; β energy : The β-wave energy value of the Fp1 / Fp2 channels in the prefrontal region; Attention_Stability: The standard deviation of attention fluctuations throughout the test.

4. The cognitive style testing device based on VR according to claim 3, wherein Eye movement tracking score S Eye : Fixation_Density: The proportion of the high-density area in the fixation point heat map; Saccade_Length: The average saccade path length; Accuracy_Rate: The correct rate of the embedded figure test.

5. The cognitive style test device based on VR according to claim 4, characterized in that, Behavior data score S Behavior : Angle_Error: The average deviation between the adjusted angle of the virtual rod and the vertical direction; Max_Error: The preset maximum allowable deviation; Time_Used: The task completion time; Time_Threshold: The time limit threshold.

6. The cognitive style test device based on VR according to claim 5, characterized in that Facial expression score S Expression : Proportion of neutral expression frames; Stress_Intensity: Intensity of stress expression, with a value range of 0 - 10 points.

7. The cognitive style test device based on VR according to claim 6, wherein, The test is divided into three difficulty levels: the primary mode, the advanced mode, and the challenge mode. The weight values in the three modes are as follows:

8. The cognitive style test device based on VR according to claim 7, characterized in that, The primary mode, the advanced mode, and the challenge mode are respectively: In the primary mode, the VR headset device controls the tilt angle of the rectangular frame ≤ 15°, and the time limit is 60 seconds; In the advanced mode, the VR headset device controls the tilt angle of the rectangular frame to change between 15° and 30°, and the time limit is 45 seconds; In the challenge mode, while the VR headset device controls the tilt angle change of the rectangular frame, environmental interference is superimposed, and the time limit is 30 seconds.

9. The cognitive style test device based on VR according to claim 8, wherein The method for testing cognitive styles includes: FII≥80 is the field-independent type; 60≤FII<80 is the transitional type; FII<60 is the field-dependent type.

10. A test method for the cognitive style test device described in claim 9, characterized in that, It includes: (1) Device initialization and calibration After starting the device, detect the connection status of the VR headset device, the physiological sensor, and the touch screen, and complete the three-dimensional attitude calibration; After the user wears the device, the pupil tracking camera assists in aligning the human eye visual axis, and the electroencephalograph and the eye tracker are calibrated respectively; (2) Cognitive style test process The tested person wears the VR headset device, and starts the experiment with the eyes aligned with the VR endoscope tube scene. During the experiment, the tested person needs to adjust the virtual rod in the scene to the vertical state through their own perception. After multiple tests, the multi-modal physiological data of the tested person during each test, as well as the angle between the virtual rod and the vertical direction after the end, are statistically analyzed. The test is carried out according to the three difficulty levels; (3) Data analysis and report generation After the test is completed, the system calls the random forest model to integrate the multi-modal data and outputs the FII index and the cognitive style classification result.

Images