A method for measuring, training and quantitatively evaluating cognitive flexibility of piano sight-reading

Abstract

The application discloses a piano sight-reading cognitive flexibility measurement training quantitative evaluation method, which measures the conversion cost of each stage of piano sight-reading training through a task switching paradigm, synchronously collects baseline and task state pulse wave signals, fuses piano sight-reading performance and physiological characteristics to establish an objective evaluation model of cognitive flexibility and quantifies the training effect of piano sight-reading. The steps are as follows: collecting the behavior performance and pulse wave signals of the subjects in the task switching paradigm test and piano sight-reading through an end-to-network cloud platform; extracting multi-dimensional physiological characteristics after denoising the pulse wave signals; statistically analyzing the conversion cost and using a correlation coefficient to screen features significantly related to the conversion cost; constructing a random forest model based on the screened features to evaluate the cognitive flexibility level; and quantifying the training effect of piano sight-reading on cognitive flexibility by comparing the score changes in each training stage. The application provides an effective method for the measurement training quantitative evaluation of cognitive flexibility.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of cognitive ability evaluation and information technology, and particularly relates to a piano sight-reading cognitive flexibility measurement, training and quantitative evaluation method. BACKGROUND

[0002] Cognitive flexibility is one of the core components of executive function, which refers to the ability of individuals to quickly switch cognitive strategies according to environmental needs. Studies have shown that the level of cognitive flexibility is closely related to learning efficiency, problem solving and creative thinking. High level of cognitive flexibility helps individuals to efficiently switch, adapt to new environments and solve complex problems in a multi-task environment, while defects in cognitive flexibility may lead to behavior solidification and maladaptation. Therefore, the evaluation and training method of cognitive flexibility has important theoretical and practical significance.

[0003] At present, the commonly used cognitive flexibility evaluation methods mainly include subjective evaluation and objective evaluation. Subjective evaluation is carried out through a scale, but it has the problems of strong subjectivity of evaluation results and difficulty in quantification. Objective evaluation mainly includes two ways: one is behavior index-based evaluation, mainly using task switching paradigm to measure the switching cost of subjects when switching different task rules to quantify the level of cognitive flexibility; the other is physiological index-based evaluation, traditional methods mainly use electroencephalogram (EEG), functional magnetic resonance imaging (fMRI) and other technologies to monitor brain activity, but these methods are expensive, complex to operate and difficult to apply in daily training environment. In recent years, it has been found that heart rate variability (HRV) as an index reflecting the activity of the autonomic nervous system has been confirmed to be closely related to cognitive function and executive control. Photoplethysmography (PPG) can more conveniently extract information related to HRV and reflect the changes of autonomic nervous system during cognitive task process. Compared with ECG, PPG sensor has the advantages of portability, non-invasiveness, simple operation and low cost, and is more suitable for real-time monitoring in natural training environment.

[0004] In the training of cognitive flexibility, existing research has confirmed that music training can promote the development of executive functions. Piano sight-reading, in particular, involves the rapid switching of multimodal information, including visual, auditory, and motor information, which aligns closely with the processing characteristics of cognitive flexibility. However, existing research suffers from two main problems: first, it lacks a systematic assessment method that integrates behavioral and physiological indicators; second, it is difficult to quantify the dynamic changes in cognitive flexibility during training, leading to a separation between assessment and training processes and reducing the real-time nature of training effectiveness evaluation. Therefore, establishing a quantitative assessment method for evaluating cognitive flexibility through piano sight-reading has significant scientific and practical value. Summary of the Invention

[0005] The purpose of this invention is to address the problems of strong subjectivity, limited evaluation dimensions, and difficulty in quantifying the training process in existing cognitive flexibility assessment methods. This invention provides a quantitative assessment method for evaluating the impact of piano sight-reading on cognitive flexibility. This method measures the switching costs at each training stage using a task switching paradigm, simultaneously collects pulse wave signals, and integrates behavioral performance with physiological characteristics to establish an objective assessment model. By combining this with a random forest algorithm to analyze the selected features, it achieves dynamic quantification of changes in cognitive flexibility levels throughout the training process. This assessment method provides an objective and systematic evaluation tool for studying the quantitative assessment of the impact of piano sight-reading on cognitive flexibility.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for quantitatively assessing the impact of piano sight-reading on cognitive flexibility, comprising the following steps:

[0008] S1. The pulse wave signals of the subjects were collected and behavioral data were recorded synchronously in the task switching paradigm test task and the piano sight-reading task through the terminal network cloud platform; wherein, the task switching paradigm adopts the cue task switching paradigm, in which the subjects need to perform one of the two tasks according to the cue prompts in each trial, to examine the subjects' cognitive flexibility in switching between different tasks; the piano sight-reading task requires the subjects to play notes according to the time indicator line.

[0009] S2. Preprocess the collected behavioral and physiological data, including handling outliers in the test task data, calculating reaction time conversion costs, calculating the playing time fit of the piano sight-reading task data, and denoising the raw pulse wave signal; wherein, the reaction time conversion cost reflects the cognitive cost borne by the subject during task conversion, and the playing time fit reflects the subject's temporal characteristics in the sight-reading task.

[0010] S3. Feature point identification and multidimensional feature extraction are performed on the denoised pulse wave signal, specifically including the construction of the main wave peak sequence and the extraction of multidimensional physiological features; wherein, the main wave peak sequence is the basis for analyzing heart rate variability features; the multidimensional physiological features include pulse wave morphological features, heart rate variability features and blood oxygen saturation features, which are used to reflect the regulation of the autonomic nervous system during cognitive tasks;

[0011] S4. Based on the conversion cost and multidimensional features, correlation analysis and feature selection are performed. First, the distribution characteristics of the conversion cost are analyzed using statistical methods to assess the differences in task conversion ability among the subject group. Then, key physiological characteristics related to cognitive flexibility level are screened using correlation analysis and the variance inflation factor method. Finally, a dataset for objectively assessing cognitive flexibility level is constructed. The distribution characteristic analysis reveals individual differences in cognitive flexibility level; the variance inflation factor is a statistical indicator used to quantify the degree of multicollinearity among features.

[0012] S5. Combining the random forest algorithm with behavioral and physiological features, an objective assessment of cognitive flexibility is achieved through feature selection and model construction. Specifically, a random forest regression model is built based on the dataset established in S4. This model takes features as input and outputs a quantitative cognitive flexibility score. Then, the K-fold cross-validation method is used to evaluate the model performance, calculating indicators such as the coefficient of determination (R²) and root mean square error (RMSE) to ensure the model's stability and generalization ability.

[0013] S6. Based on the evaluation results of different piano training stages, quantify the training effect of piano sight-reading on cognitive flexibility. First, verify the significance of the difference in conversion cost before and after training using a paired t-test statistical method to assess the statistical significance of the training effect. Second, based on the cognitive flexibility scoring model constructed in S5, calculate the changes in cognitive flexibility levels at different training stages to quantify the improvement effect brought about by training. Finally, visually display the changing trend of cognitive flexibility scores of the subject group through a line graph, reflecting the continuous impact of piano sight-reading training on cognitive flexibility. The paired t-test is a statistical method used to compare the differences in measurement values ​​of the same group of subjects at different time points; the line graph visualization method can intuitively show the dynamic changes in cognitive flexibility during the training process.

[0014] Further, step S1 is as follows:

[0015] S101, build an experimental environment through an end-to-end cloud platform, including a display, a key device, a PPG sensor, and a MIDI piano keyboard, fix the PPG sensor at the tip of the subject's left index finger, and transmit data to a cloud server in real time through a network for storage. The experiment is carried out under suitable environmental conditions. The purpose of this step is to build a stable and reliable data acquisition system to ensure the synchronous acquisition and transmission of behavioral data and physiological data;

[0016] S102, in the task switching paradigm test task stage, each trial first presents a central fixation point to guide the attention of the subject. Then a digital-letter stimulus pair composed of random combinations of letters and numbers is presented in the center of the screen, where the left and right positions of the letters and numbers are randomly determined. The color of the stimulus pair serves as the task cue of the current trial, prompting the subject to perform the type of task required. When the task cue is orange, the subject needs to judge whether the number is odd; when the task cue is blue, the subject needs to judge whether the letter is a vowel. The subject needs to use the right index finger and middle finger to make a choice by pressing the keys. When the subject makes a response and presses the keys, the target stimulus disappears immediately, and then a blank screen with a random duration is presented to reduce the influence of expectation effect on reaction time. Record the reaction time of each trial, the type of task (task repetition or switching) and the accuracy of the task during the test. This step induces the psychological processing process related to cognitive flexibility through the task switching paradigm, requiring the subject to flexibly switch between two judgment rules, thereby obtaining objective behavioral indicators reflecting cognitive flexibility;

[0017] S103, in the piano sight-reading training stage, a score based on the solfege notation is presented in the center of the display, where the numbers 1-7 represent different pitches. The playing speed is controlled by moving a red vertical indicator line from left to right on the score, and when the indicator line touches the left edge of the note, it indicates that the note should be played. The score design contains multiple rhythm change points, which require the subject to quickly adjust the established rhythm expectation by adjusting the note length and the number of notes per unit time, thereby effectively exercising the subject's cognitive flexibility. During the training, the subject's key time sequence and standard note playing time sequence are recorded, and the playing time fit degree is calculated, with a training time of 15 minutes each time. This step designs a sight-reading task with rhythm change mode to simulate the switching process in cognitive tasks and systematically train the subject's cognitive flexibility.

[0018] Further, the step S2 process is as follows:

[0019] S201, remove outliers from the trial data in the test task, including removing trials with reaction times below a set threshold and extreme value trials exceeding multiple standard deviations. For each subject, the average reaction time of the repeated task trials and the switching task trials is calculated respectively. ), and calculate the conversion cost during the reaction. The playing time fit is calculated from the sight-reading reaction time data in the piano sight-reading task. This step, through the processing of behavioral data, obtains objective behavioral indicators reflecting the level of cognitive flexibility and behavioral indicators reflecting piano sight-reading performance;

[0020] S202. The original pulse wave signal is decomposed using the Improved Complete Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) algorithm to obtain multiple Intrinsic Mode Functions (IMFs), hereinafter referred to as IMFs. The ICEEMDAN algorithm is developed based on Empirical Mode Decomposition (EMD). By introducing adaptive noise and ensemble averaging, it reduces mode aliasing and improves the stability and accuracy of the decomposition. This step extracts intrinsic mode components with different frequency characteristics from the complex pulse wave signal through an efficient signal decomposition method.

[0021] S203, Targeting the former Each IMF component is processed using wavelet thresholding denoising. An appropriate wavelet basis is selected for multi-level decomposition, and a hard thresholding function is used to remove high-frequency noise. This step removes high-frequency noise from the pulse wave signal using wavelet denoising technology, improving the signal-to-noise ratio.

[0022] S204, the front after noise reduction The denoised IMF component is superimposed with the remaining IMF components to obtain the reconstructed signal. This step combines the denoised signal with other important frequency band information to reconstruct a more accurate and clear pulse wave signal for subsequent analysis.

[0023] S205. Baseline drift is removed using a combination of differential and cubic spline interpolation. This step removes baseline drift from the pulse wave signal, providing a more reliable signal for subsequent physiological analysis.

[0024] S206. Anomaly detection and removal are performed using a dynamic thresholding method based on the correlation coefficient between adjacent periods. This step reduces the impact of motion artifact noise introduced by changes in body position and improves the waveform quality of the pulse wave signal.

[0025] Furthermore, step S3 is as follows:

[0026] S301, based on the pulse wave signal obtained in step S2, the first-order difference method is used to identify the peak points of each cycle, and the interval sequence between adjacent peak points is extracted as the basic data for heart rate variability analysis. This step provides a standardized data basis for subsequent multi-dimensional feature extraction through sequence reconstruction;

[0027] S302, analyze the pulse waveform and extract morphological features reflecting autonomic nervous activity; perform time domain analysis on the time interval sequence obtained in S301 to extract statistical features reflecting heart rate variability; when performing frequency spectrum analysis, first perform cubic spline interpolation resampling on the time interval sequence, set the sampling frequency Hz to ensure the equal interval of the data. Then use the Welch method (window length , overlap rate %) to extract the energy distribution characteristics of multiple frequency bands from the data to represent autonomic nervous regulation; construct a Poincaré map for nonlinear analysis to obtain nonlinear features reflecting heart rate dynamics; at the same time, calculate the blood oxygen saturation features based on the photoelectric characteristics of the pulse wave signal. This step obtains physiological features that comprehensively reflect the state of autonomic nervous regulation through multi-dimensional analysis.

[0028] Further, the step S4 process is as follows:

[0029] S401, calculate the mean, standard deviation, maximum, minimum, skewness and kurtosis of the conversion cost, and visualize the group differences through box plots. This step provides a basis for subsequent analysis by analyzing the statistical properties of cognitive flexibility;

[0030] S402, evaluate the correlation strength between each feature and the conversion cost through Spearman correlation analysis, and retain the features with absolute correlation coefficient and significance level ; use the variance inflation factor VIF analysis to eliminate features with VIF to eliminate collinearity and ensure independence between features. 、 、 The correlation coefficient comparison threshold, significance level comparison threshold, and inflation factor comparison threshold are respectively, this step selects behavior features and physiological features that are significantly correlated with cognitive flexibility through statistical analysis methods;

[0031] S403, construct a data set for objectively evaluating cognitive flexibility based on the behavior features and physiological features selected in step S402 and the conversion cost, and randomly divide the data set into a training set and a test set to provide a standardized data basis for subsequent model training and verification.

[0032] Further, the step S5 process is as follows:

[0033] S501. Based on the dataset constructed in step S4, a recursive feature elimination method based on random forest is adopted. In each round, the feature with the lowest feature importance score is removed. A random forest regression model is constructed and trained. The random forest regression model is used to evaluate the cognitive flexibility level of the subjects. The parameters include: number of decision trees, maximum depth, number of features when splitting at each node, and minimum number of leaf node samples. The input of the random forest regression model is the filtered features, and the output is a cognitive flexibility score of 0-100 points. The higher the score, the higher the cognitive flexibility level.

[0034] S502, adopts Cross-validation is used to evaluate the performance of the random forest regression model, calculating metrics such as the coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE) to ensure the model has good generalization ability. This step, through rigorous validation methods, guarantees the reliability of the constructed cognitive flexibility assessment model.

[0035] Furthermore, step S6 is as follows:

[0036] S601. Perform a paired t-test (or Wilcoxon signed-rank test if the data is non-normally distributed) on the conversion cost data of the subjects before and after training to calculate the p-value and assess the statistical significance of the training effect; simultaneously calculate Cohen's d effect size to quantify the magnitude of the training effect. Cohen's d effect size is a standardized difference index used to measure the degree of difference between the experimental group and the control group, or between the same group before and after training. The larger the effect size, the more significant the difference between them. This step verifies the effect of piano sight-reading training on cognitive flexibility based on direct comparison of behavioral data.

[0037] S602. Using the random forest regression model for cognitive flexibility level constructed in step S5, the subject's cognitive flexibility score is calculated at different training stages, and the relative percentage improvement is calculated based on the score changes. This step achieves quantitative evaluation of training effectiveness through a standardized scoring system.

[0038] S603. Display the group scoring trend using a line graph. This step uses visualization to visually present the changes in cognitive flexibility scores during piano sight-reading training, and analyzes the overall trend of training effectiveness.

[0039] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0040] (1) The application proposes a cognitive flexibility evaluation method fusing behavior data and physiological data, which improves the limitations of traditional methods relying only on subjective questionnaires or single behavior data. Traditional methods have problems such as single evaluation dimension, lack of objective physiological basis, and difficulty in quantifying the training process, while the application constructs a more comprehensive cognitive flexibility evaluation system through multi-dimensional index analysis, combines physiological indicators reflecting autonomic nervous system activity in pulse wave signals, and improves the objectivity and accuracy of evaluation.

[0041] (2) The application combines the ICEEMDAN algorithm and the wavelet threshold denoising technology to preprocess the pulse wave signal. By adaptively decomposing the pulse wave signal, the intrinsic mode components (IMF) in the signal are effectively extracted, and then the wavelet denoising technology is used to remove high-frequency noise, finally realizing multi-level signal optimization. This method significantly improves the signal-to-noise ratio of the pulse wave signal, solves the technical problems of excessive signal interference and noise in traditional denoising methods, and effectively improves the accuracy and stability of physiological index extraction.

[0042] (3) The application uses the piano sight-reading as a training method and evaluation tool for cognitive flexibility, and establishes a cognitive flexibility level evaluation model based on behavior data and physiological data using the random forest algorithm. The model realizes the objective evaluation of the cognitive flexibility level of the subjects through feature selection and cross-validation technology, and can accurately quantify the effect of piano sight-reading training by comparing the evaluation results of different training stages. The application breaks through the limitations of traditional cognitive training and evaluation separation, and provides a complete technical solution for cognitive flexibility training effect evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0044] Figure 1 The flowchart of the piano sight-reading cognitive flexibility training effect quantitative evaluation method disclosed in the embodiment 1 of the present application;

[0045] Figure 2 The task switching paradigm experiment flowchart in the embodiment 1 of the present application;

[0046] Figure 3 The processing flowchart of the original pulse wave signal in the embodiment 1 of the present application;

[0047] Figure 4This is a flowchart of physiological feature extraction in Embodiment 1 of the present invention;

[0048] Figure 5 This is a diagram showing the decomposition effect of the ICEEMDAN algorithm on the pulse wave in Embodiment 2 of the present invention.

[0049] Figure 6 This is a pulse wave signal image after ICEEMDAN combined with wavelet threshold denoising in Embodiment 2 of the present invention;

[0050] Figure 7 This is a pulse wave signal diagram after baseline removal in Embodiment 2 of the present invention;

[0051] Figure 8 This is a diagram showing the detection effect of the main wave peak of the pulse wave in Embodiment 2 of the present invention. Detailed Implementation

[0052] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative effort are within the scope of protection of the present application.

[0053] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0054] Example 1

[0055] This embodiment discloses a quantitative assessment method for evaluating the impact of piano sight-reading on cognitive flexibility, such as... Figure 1 As shown, the specific steps are as follows:

[0056] S1. Collect pulse wave signals and record behavioral data of subjects in the task switching paradigm test task and piano sight-reading training task through the terminal network cloud platform.

[0057] Figure 2 This is a schematic diagram of the experimental process for the above task transformation paradigm.

[0058] S101, build an experimental environment through an end-to-end cloud platform, including a display, a button device, a PPG sensor (sampling frequency 100 Hz), and a MIDI piano keyboard, fix the PPG sensor on the tip of the subject's left index finger, and transmit the data to the cloud server in real time through the network, and the experiment is carried out in a quiet laboratory with appropriate light intensity and temperature of 20-26℃;

[0059] S102, in the task switching paradigm test phase, each trial first presents a central fixation point ("+") for 150 ms to guide the subject's attention. Then a number-letter stimulus pair composed of random combinations of letters and numbers (such as "A4", "B7", etc.) is presented in the center of the screen, where the left and right positions of the letters and numbers are randomly determined. The color of the stimulus pair serves as the task cue for the current trial, indicating the type of task the subject needs to perform. When the task cue is orange, the subject needs to determine whether the number is odd; when the task cue is blue, the subject needs to determine whether the letter is a vowel. The subject needs to use the right index finger and middle finger to make a choice by pressing the button, where button "1" represents "yes" and button "2" represents "no". When the subject makes a response and presses the button, the target stimulus disappears immediately, and then a blank screen with a random duration of 1.0 to 1.8 seconds is presented to reduce the influence of expectation effect on reaction time. Record the reaction time of each trial, the task type (task repetition or switching) of each trial, and the task accuracy, and complete a total of 49 trials;

[0060] wherein the reaction time is calculated as: ;

[0061] is the reaction time of each trial, is the time when the subject makes a response, is the time when the target stimulus is presented.

[0062] S103, in the piano sight-reading training phase, a score based on the solfege notation is presented in the center of the display, where the numbers 1-7 represent different pitches. The speed of playing is controlled by moving a red vertical indicator line from left to right on the score, and when the indicator line touches the left edge of the note, it indicates that the note should be played. The score design contains multiple rhythm change points, which require the subject to quickly adjust the established rhythm expectation and avoid playing by inertia, thereby effectively exercising the subject's cognitive flexibility. Record the subject's key time sequence and standard note playing time sequence during training, and calculate the playing time fit degree, with each training lasting 15 minutes.

[0063] S2. Preprocess the collected behavioral and physiological data, including handling outliers in the test task data, calculating reaction time conversion costs, calculating the playing time fit in the piano sight-reading task, and denoising the raw pulse wave signal.

[0064] Figure 3 The flowchart above shows the denoising process for the original pulse wave signal.

[0065] S201. Outlier removal is performed on the trial data in the test task, including removing trials with reaction times less than 200 ms and extreme values ​​exceeding 3 standard deviations. For each subject, the average reaction time is calculated for both repeated task trials and switched task trials. ), and calculate the conversion cost during the reaction. .

[0066] The formula for the conversion cost during the reaction is: ;

[0067] This reflects the additional cognitive cost required for task switching. The average reaction time for repeated tasks. The average reaction time for the trial of the task switching.

[0068] The average playing time fit is calculated based on the sight-reading time sequence in the piano sight-reading task and the standard note playing time sequence.

[0069] Among them, the average playing time fit The formula is as follows:

[0070]

[0071] For the first The standard playing time for each note For the first The sight-reading reaction time for each note, that is, the playing time between adjacent notes. This represents the total number of notes in a round of sight-reading training.

[0072] S202. The original pulse wave signal is decomposed using the ICEEMDAN algorithm to obtain multiple intrinsic mode components, which are referred to as IMFs below.

[0073] S203. For the first three IMF components, wavelet thresholding denoising is used. In wavelet thresholding denoising, the wavelet basis function is selected as "db8" wavelet, the wavelet decomposition level is five, and a hard thresholding function is used to remove high-frequency noise.

[0074] The threshold used in the threshold function is a global threshold, as shown in the following formula:

[0075]

[0076]

[0077] wherein, is the noise standard deviation estimate, is the signal length, is the median absolute deviation, 0.6745 is the Gaussian distribution correction factor, and is a fixed value.

[0078] S204, superimposing the first three IMF components after noise reduction and the remaining IMF components to obtain a reconstructed signal after noise reduction;

[0079] S205, removing baseline drift by using differential method combined with cubic spline interpolation method, the specific steps are as follows:

[0080] 1) calculating the first derivative of the signal, finding the periodic starting point as the baseline drift point;

[0081] 2) using cubic spline interpolation method to fit all periodic starting points to obtain the baseline signal;

[0082] 3) subtracting the baseline signal from the denoised signal to obtain the final signal after baseline removal.

[0083] S206, using dynamic threshold method based on adjacent cycle Pearson correlation coefficient to remove abnormal cycle detection and removal.

[0084] ICEEMDAN combined with wavelet threshold denoising and cubic spline interpolation method for baseline removal is used to remove high-frequency noise and low-frequency baseline drift in the signal, so as to improve the signal-to-noise ratio of the signal; then, in order to further filter out the influence of part of the motion artifact noise introduced by the subject due to body motion change, the pulse wave signal is subjected to abnormal cycle detection and removal, so as to improve the waveform quality of the pulse wave signal.

[0085] S3, identifying and extracting multi-dimensional features of the pulse wave signal obtained in step S2, specifically including main wave peak point sequence construction and multi-dimensional physiological feature extraction;

[0086] Figure 4 is the flow chart of the above pulse wave signal physiological feature extraction.

[0087] S301, based on the pulse wave signal obtained in step S2, using first-order difference method to identify each cycle main wave peak point, so as to extract the interval sequence between adjacent peak points as the basic data for heart rate variability analysis.

[0088] S302, analyze the pulse waveform, extract the morphological features reflecting the autonomic nervous activity; time domain analysis is performed on the time interval sequence obtained in S301 to extract statistical features reflecting heart rate variability; when performing frequency spectrum analysis, first perform cubic spline interpolation resampling on the time interval sequence, set the sampling frequency to 4Hz to ensure the equal interval of the data. Then use the Welch method (window length 256, overlap rate 50%) to extract the energy distribution features of multiple frequency bands from it, which are normalized low-frequency components that can reflect sympathetic and parasympathetic activities , and the ratio of low frequency to high frequency which can reflect the balance of sympathetic and parasympathetic nerves. Among them, and are defined as follows:

[0089]

[0090]

[0091] Poincaré graph is constructed for nonlinear analysis to obtain nonlinear features reflecting heart rate dynamics; at the same time, based on the photoelectric characteristics of the pulse wave signal, the blood oxygen saturation features are calculated.

[0092] S4, correlation analysis and feature selection based on conversion cost and multi-dimensional features, first analyze the distribution characteristics of the conversion cost through statistical methods to evaluate the differences in the task switching ability of the subject group; then, through correlation analysis and variance inflation factor method, the key features related to the level of cognitive flexibility are selected; finally, the data set for objectively evaluating the level of cognitive flexibility is constructed;

[0093] S401, calculate the statistics including the mean, standard deviation, maximum, minimum, skewness and kurtosis of the conversion cost, and visualize the group differences through box plots;

[0094] S402, evaluate the correlation strength between each feature and the conversion cost through Spearman correlation analysis, and retain the features with correlation coefficient absolute value and significance level ; use variance inflation factor VIF analysis to exclude features with VIF to eliminate collinearity and ensure the independence between features;

[0095] S403, construct a data set for objectively evaluating the level of cognitive flexibility by combining behavioral features and physiological features, and randomly divide the data set into a training set (80%) and a test set (20%) for subsequent model training and verification to provide a standardized data basis.

[0096] S5, combine the behavior indicators and physiological characteristics with the random forest algorithm, realize the objective evaluation of the cognitive flexibility level through feature selection and model construction. Specifically, based on the data set established in S4, a random forest regression model is constructed, which takes the features as input and outputs a quantitative cognitive flexibility score; then the 10-fold cross-validation method is used to evaluate the model performance, and the determination coefficient and root mean square error are calculated to ensure the stability and generalization ability of the model;

[0097] S501, based on the data set constructed in step S4, the recursive feature elimination method based on random forest is used, and the feature with the lowest feature importance score is removed in each round, a random forest regression model is constructed and trained, the random forest regression model is used to evaluate the cognitive flexibility level of the subject, the parameters include: the number of decision trees, the maximum depth, the number of features when splitting each node, and the minimum leaf node sample size, the random forest regression model input is the filtered features, and the output is a cognitive flexibility score of 0-100 points, the higher the score, the higher the cognitive flexibility level;

[0098] S502, the 10-fold cross-validation method is used to evaluate the performance of the random forest regression model, and the determination coefficient and root mean square error are calculated to ensure that the model has good generalization ability.

[0099] S6, according to the evaluation results of different piano training stages, the training effect of piano sight-reading on cognitive flexibility is quantified. First, the paired T-test statistical method is used to verify the significant difference of the transition cost before and after training, and the statistical significance of the training effect is evaluated; secondly, based on the cognitive flexibility score model constructed in S5, the change of cognitive flexibility level in different training stages is calculated, and the improvement effect brought by training is quantified; finally, the change trend of cognitive flexibility score of the subject group is intuitively displayed through the broken line chart, reflecting the continuous influence of piano sight-reading training on cognitive flexibility.

[0100] S601, paired T-test (Wilcoxon signed rank test for non-normal distribution data) is performed on the transition cost data of the subjects before and after training, and the p value is calculated to evaluate the statistical significance of the training effect; at the same time, Cohen's d effect size is calculated to quantify the size of the training effect. This step is based on the direct comparison of behavior data to verify the influence of piano sight-reading training on cognitive flexibility;

[0101] S602, based on the random forest regression model, the cognitive flexibility scores of the subjects in different training stages are calculated, and the relative improvement percentage is calculated based on the score change. This step realizes the quantitative evaluation of the training effect through the standardized scoring system;

[0102] S603, display the group score trend through the line chart. This step intuitively presents the change of the cognitive flexibility score in the piano sight-reading training process through a visual method, and analyzes the overall trend of the training effect.

[0103] Embodiment 2

[0104] This embodiment further discloses a piano sight-reading cognitive flexibility measurement and quantitative evaluation method, as shown in Figure 1 The specific steps are as follows:

[0105] S1, refer to the corresponding steps in Embodiment 1, which will not be repeated here;

[0106] S2, refer to the corresponding steps in Embodiment 1, which will not be repeated here. The decomposition effect of ICEEMDAN algorithm on pulse wave signal is as shown in Figure 5 The pulse wave signal after using ICEEMDAN combined with wavelet threshold denoising is as shown in Figure 6 The pulse wave signal after baseline removal is as shown in Figure 7

[0107] S3, refer to the corresponding steps in Embodiment 1, which will not be repeated here. The main wave peak point detection effect is as shown in Figure 8

[0108] S4, refer to the corresponding steps in Embodiment 1, which will not be repeated here. The distribution characteristics of the transition cost are shown in Table 1; the top five physiological characteristics significantly related to the transition cost are shown in Table 2;

[0109] Table 1. Summary of transition cost statistical characteristics of all subjects

[0110]

[0111] Table 2. Top five physiological important characteristics and correlation analysis of transition cost

[0112]

[0113] S5, refer to the corresponding steps in Embodiment 1, which will not be repeated here. The cognitive flexibility evaluation model indicators are shown in Table 3;

[0114] Table 3. Cognitive flexibility evaluation model indicators

[0115]

[0116] S6, refer to the corresponding steps in Embodiment 1, which will not be repeated here. The transition cost change of piano sight-reading training is shown in the table.

[0117] Table 4. Transition cost change before and after piano sight-reading training ​​

[0118]

[0119] In summary, the five physiological characteristics shown in Table 2 mainly reflect the regulatory ability of the autonomic nervous system, especially the parasympathetic nervous system, which plays a key role in cognitive resource allocation and is closely related to cognitive flexibility; combined with the good fitting effect of 0.68 achieved by the random forest regression model in Table 3 on the test set, and the trend of significant decrease in conversion cost after training shown in Table 4, the feasibility and effectiveness of the present application in the quantitative evaluation of cognitive flexibility are fully verified.

[0120] The technical features of the above embodiments can be combined in any manner. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist contradictions, they should be considered as within the scope of the present disclosure.

[0121] The above embodiments are the preferred embodiments of the present application, but the embodiments of the present application are not limited to the above embodiments, and any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present application shall be equivalent replacement modes and shall be within the scope of protection of the present application.

Claims

1. A method for measuring, training and quantitatively evaluating the cognitive flexibility of piano sight-reading, characterized in that, The measurement and quantitative evaluation method comprises the following steps: S1, collecting the pulse wave signals of the subjects in the task switching paradigm test task and the piano sight-reading task through the end-to-network cloud platform and recording the behavior data; S2, calculate the switching cost and the playing time fitness after pre-processing the behavior data, and denoise the original pulse wave signal; wherein, for each subject, calculate the average reaction time of the repeated task trial and the switching task trial , and calculate the reaction time switching cost , the formula of the reaction time switching cost is: , reflects the additional cognitive cost required for task switching, is the average reaction time of the repeated task trial, is the average reaction time of the switching task trial; According to the sight-reading time sequence in the piano sight-reading task and the standard note playing time sequence, the average playing time fitting degree is calculated; Average playing time fit The formula is as follows: , is the standard playing time of the th note, is the sight-reading reaction time of the th note, i.e. the playing time between adjacent notes, is the total number of notes in a round of sight-reading training; S3, feature extraction is performed on the denoised pulse wave signal to obtain pulse wave morphological features, heart rate variability features and blood oxygen saturation features; S4, correlation analysis is performed based on the switching cost and multi-dimensional features, and feature screening is performed to construct a dataset containing behavior features and physiological features; S5, combining the random forest algorithm to fuse the behavior features and the physiological features, realizing the objective evaluation of the cognitive flexibility level through feature selection and model construction; S6, according to the evaluation results of different piano training stages, quantifying the training effect of piano sight-reading on cognitive flexibility.

2. The piano sight-reading cognitive flexibility measurement and quantitative evaluation method according to claim 1, characterized in that, The step S1 process is as follows: S101, building an experimental environment through the end-to-network cloud platform, the subjects make responses through the keys, and wear pulse wave sensors to record pulse wave signals; S102, in the task switching paradigm test task stage, presenting the target stimulus of letter and number combination on the display, and performing different judgment tasks according to the task clues; S103, in the piano sight-reading training stage, presenting the music score based on the tablature on the display, and playing notes according to the time indication line.

3. The method for measuring and quantifying the cognitive flexibility of a piano sight-reading according to claim 1, wherein, The step S2 process is as follows: S201, preprocessing the behavior data, calculating the reaction time switching cost of the test task and the playing time fitting degree of the piano sight-reading task; S202, using the improved adaptive noise ensemble empirical mode decomposition to decompose the original pulse wave signal to obtain multiple intrinsic mode components, which are referred to as IMFs below; S203, for the first Q IMFs, using the wavelet threshold denoising method for processing; S204, superimposing the first Q denoised IMFs and the remaining IMFs to obtain the denoised reconstructed signal; S205, using a cubic spline interpolation fitting curve to remove the baseline drift; S206, detecting and removing abnormal periods of the pulse wave signal.

4. The method for measuring and quantifying the cognitive flexibility of a piano sight-reading according to claim 1, wherein, The step S3 process is as follows: S301, performing feature point recognition to extract the main wave peak point sequence of the pulse wave signal; S302, analyzing the pulse wave signal and the main wave peak point sequence to extract multi-dimensional physiological features including pulse wave morphological features, heart rate variability features and blood oxygen saturation features.

5. The method for measuring and quantifying the cognitive flexibility of a piano sight-reading according to claim 1, wherein, The step S4 process is as follows: S401, analyzing the distribution characteristics and group differences of the switching cost; S402, screening out the behavior features and physiological features through correlation analysis and variance inflation factor analysis; S401, calculating the statistics including the mean, standard deviation, maximum value, minimum value, skewness and kurtosis of the switching cost, and visualizing the group differences through the box plot; S402、Through Spearman correlation analysis, evaluate the correlation strength between each feature and the conversion cost, and retain the absolute value of the correlation coefficient and the significance level features; The variance inflation factor VIF analysis is used to eliminate the characteristics with VIF > 10 to eliminate collinearity, , , respectively, the correlation coefficient comparison threshold, the significance level comparison threshold, and the inflation factor comparison threshold, and the behavior characteristics and physiological characteristics significantly related to cognitive flexibility are screened out. S403, constructing a dataset for objectively evaluating the cognitive flexibility level through the behavior features and the physiological features.

6. The method for measuring the cognitive flexibility of a piano sight-reading according to claim 1, wherein, The step S5 process is as follows: S501、Based on the data set constructed in step S4, the random forest-based recursive feature elimination method is used to remove the feature with the lowest feature importance score in each round, and a random forest regression model is constructed and trained. The random forest regression model is used to evaluate the cognitive flexibility level of the subject. The parameters include: the number of decision trees, the maximum depth, the number of features when splitting each node, and the minimum leaf node sample size. The random forest regression model input is the filtered features, and the output is a cognitive flexibility score of 0-100 points. The higher the score, the higher the cognitive flexibility level. S502、Use cross-validation to evaluate the performance of the random forest regression model.

7. The method of claim 1, wherein the method is a piano sight-reading cognitive flexibility assessment method. The step S6 process is as follows: S601、Verify the significant difference in conversion cost before and after training by paired T-test; S602、Based on the random forest regression model, calculate the cognitive flexibility score of the subject at different training stages, and calculate the relative improvement percentage based on the score change; S603、Show the group score trend through a line chart.

Images