Quantitative evaluation method for intervention effect of tuba blowing training on inhibition control ability

Abstract

The application discloses a kind of long horn blowing training intervention effect quantification evaluation method of control ability, steps are as follows: the pulse wave signal of each state of the subject when executing long horn blowing training is collected and denoised, and physiological characteristics are extracted therefrom;Behavior data of the subject when executing blowing training is collected, including blowing airflow signal and put position control data, and behavior characteristics are extracted therefrom;The accuracy and reaction time of the subject executing inhibition control test task are recorded, and the test task score is calculated by probability distribution fitting;With the physiological characteristics and behavior characteristics of the subject as explanatory variables, the test task score as target variable constructs data set, trains regression model and filters optimal feature subset to construct quantification formula, and obtains the quantification evaluation result of intervention effect.The present application relies on long horn blowing training, and provides a kind of quantitative analysis method for the intervention effect research of music therapy on inhibition control ability.

Description

Technical Field

[0001] This invention belongs to the technical field of information technology and smart health, and specifically relates to a quantitative evaluation method for the intervention effect of trombone playing training on inhibitory control ability. Background Technology

[0002] Inhibitory control refers to an individual's ability to suppress dominant or autonomous responses and focus on executing the current cognitive goal when faced with interfering stimuli or conflicting information. It is an important component of the brain's executive function and plays a key role in higher cognitive activities such as cognitive decision-making and behavioral planning.

[0003] Current research on interventions to improve inhibitory control primarily focuses on three areas: cognitive training, exercise intervention, and music therapy. These studies typically employ three evaluation paradigms to assess intervention effectiveness: subjective scales, pre- and post-intervention tests, and intervention-control groups. Subjective scales require participants to express their feelings about various dimensions of the intervention's effectiveness, making this method highly subjective. Pre- and post-intervention tests require participants to perform inhibitory control tasks multiple times, evaluating the intervention's effectiveness by comparing performance before and after the intervention; however, this approach lacks dynamic analysis of the intervention process and has limited data dimensions. The intervention-control group paradigm, building upon pre- and post-intervention tests, further assesses the intervention's effectiveness by comparing performance across different experimental groups, but similarly lacks data analysis of the intervention process. Most of these three evaluation paradigms rely on significance tests to assess intervention effectiveness, resulting in limited evaluation dimensions, a lack of quantitative indicators, and difficulty in constructing a quantitative evaluation model of intervention effectiveness based on "intervention input - performance output," thus hindering dynamic monitoring of the intervention process. Summary of the Invention

[0004] To overcome the shortcomings and deficiencies of current methods for evaluating the intervention effects in inhibitory control ability intervention research, this invention provides a quantitative evaluation method for the intervention effect of trombone playing training on inhibitory control ability. This method relies on trombone playing training music therapy, and models and analyzes the scores of inhibitory control test tasks based on the physiological and behavioral characteristics of each stage of playing training, achieving a multimodal quantitative evaluation of the intervention effect. This has practical significance for intervention research on inhibitory control ability.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A quantitative evaluation method for the intervention effect of trombone playing training on inhibitory control ability, the quantitative evaluation method comprising the following steps: S1. Photoplethysmography (PPG) was used to collect baseline, task-state, and recovery-state pulse wave signals during the subject's AI trombone playing training. The airflow signal during the subject's playing training was collected through the pickup module. The level signal output by the ultrasonic ranging module was recorded. This level signal is the original position control data. The accuracy and reaction time of the inhibition control test task were also recorded.

[0006] The AI ​​trombone is an intelligent musical instrument derived from the modification of a traditional trombone in the laboratory, and it is mainly used for data acquisition in this invention. The baseline pulse wave signal refers to the pulse wave signal collected from the subject in a resting state 5 minutes before performing the playing training; the state of performing the playing training is called the task state, and the pulse wave signal collected in the task state is called the task state pulse wave signal; the state 5 minutes after performing the playing training is called the recovery state, and the pulse wave signal collected in the recovery state is called the task state pulse wave signal. The inhibition control test tasks include the Stroop test task and the variant Flanker test task.

[0007] S2. For the level signal obtained in step S1, a combined variational mode decomposition and wavelet thresholding denoising method is used to filter out high-frequency noise and reduce the influence of motion artifacts. Simultaneously, a Butterworth bandpass filter is used to further improve the pulse wave signal quality. A first-order forward differential-based reverse ergodic method is used to extract the starting point of each pulse wave cycle. Cubic spline interpolation is used to remove baseline drift in the pulse wave signal. After the above processing, denoised baseline state pulse wave signals, denoised task state pulse wave signals, and denoised recovery state pulse wave signals are obtained. For the wind-playing airflow signal, notch filtering and median filtering are used for denoising to obtain a denoised wind-playing airflow signal. The original position control data is converted to obtain position control data represented as a distance sequence.

[0008] S3. Identify the main peak points of the denoised baseline pulse wave signal, the denoised task pulse wave signal, and the denoised recovery pulse wave signal obtained in step S2, and extract pulse rate variability features based on this. For the denoised playing airflow signal, segment it based on the playing training sound score and extract the playing breath stability features. Similarly, segment the position control data and extract the position control accuracy features based on statistical principles.

[0009] This step extracts the physiological and behavioral characteristics of the subjects during the wind instrument training process, and these characteristics will serve as explanatory variables for the quantitative evaluation method of intervention effect.

[0010] S4. Based on histogram analysis and goodness-of-fit test, the probability distribution of the reaction time of the inhibition control test task is fitted. The test task score is obtained by calculating the cumulative probability and combining it with the test task accuracy.

[0011] This step takes into account that the performance on the inhibition control test task indirectly reflects the subject's inhibitory control ability. Therefore, the test task score is calculated using the accuracy and reaction time of the inhibition control test task, and the inhibitory control ability is characterized by the test task score.

[0012] S5. Construct a dataset based on the subjects' physiological and behavioral characteristics. Train a random forest regression model using the test task scores obtained in step S4 as the target variable. Evaluate the performance of different feature subsets through recursive feature elimination and cross-validation, and select the feature subset with the highest performance, i.e., the optimal feature subset. The physiological characteristics are pulse rate variability features extracted from baseline, task-state, and recovery-state pulse wave signals. The behavioral characteristics are breath stability and position control accuracy features extracted from the airflow signal and position control data.

[0013] S6. Perform correlation analysis on the optimal feature subset obtained in step S5. Based on the analysis results, the features are dimensionless, and the orientation of all features is transformed to be positively correlated with the test task score, i.e., positively correlated with inhibitory control ability. Retrain the random forest regression model, construct a quantitative evaluation formula based on feature importance, and obtain a quantitative score for "trombone playing training - inhibitory control ability". This score explains the influence of physiological and behavioral characteristics on inhibitory control ability during playing training. By analyzing the changes in the quantitative score, the quantitative evaluation of the intervention effect of trombone playing training on inhibitory control ability is realized.

[0014] Further, step S1 is as follows: S101. The subject sits in front of a screen displaying the UI interface of the trombone playing training module. A PPG sensor is placed on the desktop where the screen is located. First, the screen displays a 5-minute countdown and prompts the subject to adjust the sensor position and place the left middle finger on the PPG sensor. The subject sits quietly for 5 minutes to collect the baseline pulse wave signal before training. After the countdown ends, the screen prompts the subject to start playing training.

[0015] S102. The screen displays the playing training sound score. During the training, the subject is required to keep his upper body upright, wear headphones, hold the slide of the AI ​​trombone with his right hand, hold the grip of the AI ​​trombone with his left hand, and put a PPG sensor finger sleeve on the middle finger of his left hand. The mouthpiece of the AI ​​trombone is tilted downward at 15 degrees, and the lips are naturally aligned with the mouthpiece of the AI ​​trombone. The PPG sensor finger sleeve is used to collect task-state pulse wave signals.

[0016] This step is to guide the subject to perform the experiment in the correct posture. Using headphones to receive the playing feedback is to prevent the pickup from picking up trombone sounds. Holding the subject's hand still as the pulse wave signal for hand detection is to reduce motion artifacts caused by arm swinging.

[0017] S103. The musical score is played in animated form on the screen. The subject needs to play with stable breath while moving the slide to the correct position according to the notes and position prompts displayed in the musical score. The playing training lasts for a total of [number missing]. In each round, a piece of musical score is played randomly. The pickup module on the AI ​​trombone is used to collect the airflow signal of the subject when performing the playing training, and the level signal output by the ultrasonic ranging module is recorded, which is the original position control data.

[0018] S104. After the playing training, the screen prompts the subject to remove the finger cot and place the left middle finger on the desktop PPG sensor. A 5-minute countdown is displayed, and the subject is prompted to sit quietly for 5 minutes to collect the subject's restorative pulse wave signal.

[0019] S105. The subject sits in front of a display screen showing the UI of the inhibition control test task module. The inhibition control test task consists of a Stroop test task and a variant of the Flanker test task. The screen first prompts the subject to perform the Stroop test task, which is a font color recognition task. A word representing a color is displayed in the center of the screen. When the font color of the word matches the color it represents, it is called a consistent task; when the font color of the word does not match the color it represents, it is called an inconsistent task. The subject needs to select from the color options below the word. After the subject makes a selection, the screen displays the accuracy rate and reaction time for that question, and automatically moves to the next question after 1 second, until all Stroop test tasks are completed. The accuracy rate and reaction time for each test task are recorded.

[0020] S106. After the subject completes all Stroop test tasks, the display prompts the subject to perform a variant Flanker test task. The center of the display shows a 7th-order square matrix composed of directional arrows. Each arrow in the matrix has at most two directions; if two directions exist, they must be opposite. All arrows in the matrix except the middle column point in the same direction; this direction is called the row interference direction. All arrows in the middle column of the matrix except the middle row point in the same direction; this direction is called the column interference direction. The arrow at the center of the diagonal of the matrix is ​​called the target direction arrow. The subject needs to determine the directionality of the row and column interference directions and press the corresponding keyboard arrow key. When the row and column interference directions are the same, the subject needs to press the keyboard arrow key in the same direction as the target direction arrow; this test task is called a consistent task. When the row and column interference directions are opposite, the subject needs to press the keyboard arrow key in the opposite direction to the target direction arrow; this test task is called an inconsistent task. The direction "..." The opposite direction of " is " ",direction" The opposite direction of " is " "After the subject presses the arrow keys on the keyboard, the screen will display the accuracy rate and reaction time for that question, and automatically proceed to the next question after 1 second, until all variations of the Flanker test task are completed. The accuracy rate and reaction time for each test task are recorded;" Furthermore, step S2 is as follows: S201. Preprocess the pulse wave signal acquired in step S1. The specific steps are as follows: S201.1. Variational mode decomposition is used on the baseline pulse wave signal, the task-state pulse wave signal, and the recovery-state pulse wave signal to obtain multiple intrinsic mode function (IMF) components. A fast Fourier transform is performed on each IMF component, and the frequency corresponding to the maximum peak value in the spectrum is calculated. .

[0021] S201.2, Regarding The high-frequency IMF components are denoised using wavelet thresholding to filter out high-frequency noise and reduce motion artifacts; wherein, the This represents the maximum heart rate frequency in the normal population.

[0022] S201.3. The high-frequency IMF component processed in step S201.2 is superimposed with the remaining IMF components to reconstruct the pulse wave signal.

[0023] S201.4. For the reconstructed pulse wave signal, use a Butterworth bandpass filter to further improve the pulse wave signal quality and signal-to-noise ratio.

[0024] S201.5. The first-order forward differential reverse traversal method is used to extract the starting point of each pulse wave cycle in the pulse wave signal to obtain the pulse wave cycle starting point sequence.

[0025] S201.6 Perform cubic spline interpolation on the pulse wave cycle start point sequence to fit the pulse wave baseline, remove the baseline drift of the pulse wave signal, and obtain the denoised baseline state pulse wave signal, the denoised task state pulse wave signal, and the denoised recovery state pulse wave signal.

[0026] S202. The blowing airflow signal acquired in step S1 is processed for noise reduction using notch filter and median filter.

[0027] The design principle of a notch filter is as follows: The transfer function of the Laplace transform notch filter can be expressed as:

[0028]

[0029] in, The center frequency of the notch filter. , Notch factor; Notch filters are designed based on three main parameters: (1) Notch filter center angular frequency, ; (2) Notch depth, i.e., the attenuation depth at the center frequency; (3) Notch bandwidth , and This is the cutoff angular frequency.

[0030] The notch factor can be obtained by solving the Z-transform:

[0031] in, ,Will , Substituting into formula (7) will complete the notch filter design.

[0032] This step takes into account that the airflow signal is essentially a non-stationary wideband signal, so there is no need to extract specific frequency components. However, the signal may be subject to power frequency noise interference and pulse interference. Therefore, notch filter and median filter are selected for noise reduction.

[0033] S203. Based on the original shift control data obtained from S1, according to the speed of sound... Duration of high level The level signal is converted into distance to obtain the position control data represented by a distance sequence.

[0034] The distance conversion formula is as follows:

[0035] in, This is the distance between the ultrasonic transmitter and receiver.

[0036] Furthermore, step S3 is as follows: S301. Extract pulse rate variability features from the denoised baseline pulse wave signal, the denoised task pulse wave signal, and the denoised recovery pulse wave signal obtained in step S2.

[0037] S301.1 Use the first-order forward differential ergodic method to identify the peak point of the main wave of the pulse wave signal.

[0038] S301.2. Perform a difference operation on the main peak point sequence to obtain the time interval sequence between two adjacent pulse wave peaks, i.e., the PP interval sequence.

[0039] S301.3. Estimate the frequency domain power spectrum of the PP interval sequence using an autoregressive model (AR model).

[0040] The expression for the AR model is as follows:

[0041] in, yes The sequence value at time 10:00. For constant terms, Let be the autoregressive coefficient, representing the th . The impact of each lag term on the current sequence value It is white noise. .

[0042] The AR model calculates the autoregressive coefficients using a parameter estimation method based on least squares. The transfer function of the AR model is obtained as follows:

[0043] Furthermore, there are PP spacer sequences. Spectral estimation results of the AR model:

[0044] The noise variance can be calculated using the following formula:

[0045] in, The length of the PP interval sequence. Indicates AR model for sequence The predicted value at any given time.

[0046] This step takes into account that extracting the frequency domain features of pulse rate variability requires high frequency resolution. Reconstructing the PP interval sequence using cubic spline interpolation is easily limited by the interpolation frequency, resulting in low frequency resolution. Therefore, [the following option is chosen]. An autoregressive model is used to estimate the power spectral density of the PP interval sequence by calculating the autoregressive coefficients.

[0047] S301.4 Extracting the time-domain, frequency-domain, and nonlinear features of Pulse Rate Variability (PRV) from the PP interval sequence. Extracting PRV features from the baseline pulse wave signal to obtain baseline PRV features, and constructing a baseline PRV feature vector from these features. Extract PRV features from the task-state pulse wave signal to obtain the task-state PRV features. Subtract the baseline PRV features from the task-state PRV features to obtain the baseline-reduced task-state PRV features. Construct the baseline-reduced task-state PRV feature vector from the baseline-reduced task-state PRV features. Extracting PRV features from the recovering pulse wave signal yields the recovering PRV features. Subtracting the baseline PRV features from the recovering PRV features results in the baseline-debasemented recovering PRV features. These baseline-debasemented recovering PRV features are then used to construct the baseline-debasemented recovering PRV feature vector. .

[0048] The pulse rate variability features extracted in this step include six time-domain features: mean pulse interval (PRV_MEANPP), standard deviation of pulse interval (PRV_SDPP), root mean square of the difference between adjacent pulse intervals (PRV_RMSSD), the proportion of the number of times the difference between adjacent pulse intervals is greater than 20ms and 50ms to the total number of pulse intervals (PRV_PNN20, PRV_PNN50), and triangle index (PRV_TI); three frequency-domain features: low-frequency power (PRV_LF), high-frequency power (PRV_HF), and the ratio of low-frequency power to high-frequency power (PRV_LF / PRV_HF); and four non-linear features: sample entropy (PRV_SampEn), short-term changes in the Poincare plot (PRV_SD1), long-term changes in the Poincare plot (PRV_SD2), and PRV_SD1 / PRV_SD2. The specific steps for calculating sample entropy are as follows: 1) Assume the length of the PP interval sequence is... For a given pattern dimension A set can be constructed 3D vector , usually take Then we have:

[0049] 2) In Set similarity tolerance threshold at the endpoint Assuming , The intervals represent and The tolerance range, if If the corresponding endpoints are all within the tolerance range, then it is considered that... and At similar tolerance threshold Similarity, define vector and The ratio of the approximate number to the total number of both is This allows us to obtain the average value of the ratio of the approximate quantity to the total quantity. :

[0050] 3) Order Repeat the above steps to obtain Finally, the sample entropy of the sequence is obtained:

[0051] S302. The preprocessed airflow signal obtained in step S2 is segmented based on the measures of the airflow training score. Each segment contains the same number of notes. The sliding window algorithm is used to divide each segment into frames. The short-time energy (STE) of each frame is calculated to obtain the short-time energy sequence. The coefficient of variation (CV) and sample entropy of each short-time energy sequence are calculated. Finally, the mean is taken to obtain the average coefficient of variation (STE_MEANCV) and average sample entropy (MEANSampEn) of the short-time energy sequence. At the same time, the standard deviation (STE_MEAN_SD) of the mean of each short-time energy sequence is calculated.

[0052] This step takes into account that the airflow signal is inherently a non-stationary broadband signal, therefore it uses frame-by-frame calculation of short-time energy, and characterizes the stability of the breath during playing by evaluating the fluctuation of short-time energy. Among them, STE_MEANCV and MEANSampEn reflect the breath stability of the subject's playing for each note, and STE_MEAN_SD reflects the consistency of the subject's breath stability for each note, that is, the overall breath stability of playing.

[0053] The formula for calculating short-time energy is as follows:

[0054] in, Indicates the first The short-time energy of a frame, For frame length, For the first The sampled values ​​of each sampling point.

[0055] CV is the standard deviation of the signal. and mean The ratio is given by the following formula:

[0056] S303. The slide position control data obtained in step S2 is divided into segments based on the measures of the playing training sound score. Each segment contains the same number of notes, and each note has its corresponding target position. The mean absolute error (SP_MAE) and root mean square error (SP_RMSE) of the distance between the actual position and the target position are calculated, and the position control accuracy (SP_ACC) is further calculated.

[0057] The AI ​​trombone has 7 positions, the first... Position The distance to the ranging module is Then there is a positional distance sequence. When the actual position meets the position determination criteria, the AI ​​trombone will determine the actual position as the target position. The position determination criteria are as follows:

[0058] in, To actually position oneself, This indicates that the actual position was determined to be the first. Position, To determine the actual distance from the positioning point to the ranging module, This is the distance from the position of maximum length of the sliding tube to the ranging module.

[0059] Based on the position judgment standard, for a duration of If the duration of the note segment, where the actual position is judged as the target position, is less than... If the note is not played correctly, then the SP_ACC is calculated as follows:

[0060] Furthermore, step S4 is as follows: S401. For each inhibition control test task, plot a reaction time histogram to analyze the data distribution pattern.

[0061] S402. Through the Anderson-Darling (AD) goodness-of-fit test, the target probability distribution model accepting the null hypothesis is selected as the probability distribution for reaction time, and the cumulative probability of reaction time is obtained. The null hypothesis is that the reaction time follows a target probability distribution.

[0062] This step is to verify the reasonableness of fitting the target probability distribution during reaction time.

[0063] S403. Regarding the accuracy of the test tasks, since the accuracy is already a normalized metric, the accuracy of each test task is calculated as the sum of the accuracy and the reaction time of that task. Multiply the scores to get the score for each test task, and add the scores of each test task together to get the score of the test subject for this test task.

[0064] Furthermore, step S5 is as follows: S501. Construct a dataset based on the pulse rate variability features, breath stability features, and position control accuracy features extracted in step S3.

[0065] S502. Using the test task score obtained in step S4 as the target variable and physiological and behavioral characteristics as explanatory variables, train a random forest regression model.

[0066] S503. Using the random forest regression model obtained in step S502 as the base model, recursive feature elimination is used to progressively remove the features with the lowest importance. Cross-validation is then used to calculate the RMSE and coefficient of determination (R²) of the feature subset. 2 ), evaluate the performance of the feature subset at each recursion, and finally select the R subset. 2 The feature subset with the highest R² and lowest RMSE is selected. If there is a conflict between R² and RMSE, the feature subset with the highest R² is preferred. Here, RMSE and R² are... 2 It is the most commonly used metric for evaluating the performance of regression models in cross-validation, and its calculation formula is as follows:

[0067]

[0068] in, For the sample size, For the first The true value of each sample represents the test task score. For the first The predicted value for each sample, This is the average of the true values ​​in the sample.

[0069] Furthermore, step S6 is as follows: S601. Perform Spearman correlation analysis on the optimal feature subset obtained in step S5 to obtain the positive and negative correlation between each feature in the subset and the test task score.

[0070] The principle of Spearman correlation analysis is to convert each feature value of the sample into rank data, that is, the ranking of the feature value with respect to the value of that feature in all samples. By calculating the correlation between the rank data, the Spearman rank correlation coefficient is obtained. The range is , This indicates that the relationship between the two variables is positively correlated. This indicates that the relationship between the two variables is negatively correlated. This indicates that there is no correlation between the two variables. The calculation formula is as follows:

[0071] in, It is two variables. The rank difference of each observation This represents the number of samples.

[0072] This step takes into account that the physiological and behavioral characteristics may not have a simple linear correlation with the target variable. Therefore, Spearman correlation analysis is chosen to measure the monotonic relationship between the characteristics and the target variable based on the correlation between the data ranks.

[0073] S602. Based on the positive and negative correlations obtained in step S601, perform Min-Max standardization on each feature.

[0074] For positively correlated features, the Min-Max standardized expression is as follows:

[0075] in, Indicates the first Features in each sample The original data, This represents the standardized data. , These are the minimum and maximum values ​​of this feature in the dataset, respectively.

[0076] For negatively correlated features, the Min-Max inverse normalization expression is as follows:

[0077] This step is to unify the feature orientation of all features, eliminate differences in feature correlation, and facilitate the construction of subsequent quantitative evaluation formulas.

[0078] S603. Retrain the regression model and construct a quantitative evaluation formula based on feature importance to obtain a quantitative score for "trombone playing training - inhibitory control ability". This score explains the influence of physiological and behavioral characteristics on inhibitory control ability during playing training. By analyzing the changes in the quantitative score, the quantitative evaluation of the intervention effect of trombone playing training on inhibitory control ability can be achieved.

[0079] Compared with the prior art, the present invention has the following advantages and beneficial effects: 1) This invention collects pulse wave, exhalation, and position distance data during trombone playing training using low-cost and easy-to-use data acquisition equipment, and extracts physiological and behavioral characteristics from them to establish a predictive model for inhibitory control test scores. It constructs a multimodal quantitative evaluation technology scheme for the intervention effect of trombone playing training on inhibitory control ability, realizes the objective and dynamic evaluation of the intervention effect, and overcomes the problems of strong subjectivity and single evaluation index dimension of traditional methods.

[0080] 2) This invention proposes a statistical method for calculating the score of the inhibition control test task. By fitting the probability distribution of the reaction time of the inhibition control test task through histogram analysis and goodness-of-fit test, the cumulative probability of each reaction time is calculated. By weighted summing of the cumulative probability and the corresponding test task accuracy, an objective and scientific score of the inhibition control test task is obtained, and a reliable standard for quantitatively evaluating the intervention effect is established.

[0081] 3) This invention calculates the importance of physiological and behavioral features by combining feature selection and cross-validation. Based on the optimal number of features obtained from cross-validation, the least important features are eliminated one by one through a recursive feature elimination algorithm, thereby removing redundant and irrelevant features and improving computational efficiency and accuracy.

[0082] 4) This invention eliminates the dimensions of physiological and behavioral characteristics through dimensionless elimination, and obtains a quantitative score of the intervention effect by weighted summation based on the importance weights of each characteristic. This quantitative score helps people establish an accurate and intuitive understanding of the intervention effect, and helps them develop more scientific training programs. Furthermore, by adapting to the specific behavioral characteristics of playing different musical instruments, the assessment technique proposed in this invention can be applied to intervention studies of playing different musical instruments. Attached Figure Description

[0083] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0084] Figure 1 This is a flowchart of a method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability, as disclosed in Embodiment 1 of the present invention. Figure 2 This is a structural diagram of the experimental system modules in Embodiment 1 of the present invention; Figure 3 This is a flowchart of the physiological and behavioral data preprocessing in Embodiment 1 of the present invention; Figure 4 This is a flowchart of physiological and behavioral feature extraction in Embodiment 1 of the present invention; Figure 5 This is a flowchart of the quantitative evaluation model for intervention effects based on physiological and behavioral characteristics in Embodiment 1 of the present invention; Figure 6 This is a waveform diagram of the original pulse wave signal in Embodiment 2 of the present invention; Figure 7 This is a waveform diagram of the airflow signal during blowing in Embodiment 2 of the present invention; Figure 8 This is a schematic diagram of the position control data sequence in Embodiment 2 of the present invention; Figure 9 This is a diagram showing the result of feature point recognition of the denoised pulse wave signal in Embodiment 2 of the present invention; Figure 10 This is a distribution curve of inconsistent task response time in Embodiment 2 of the present invention; Figure 11 This refers to the RMSE and R of different feature subsets during recursive feature elimination in Embodiment 2 of the present invention. 2 Schematic diagram. Detailed Implementation

[0085] 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.

[0086] 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.

[0087] Example 1 Figure 1 This is a flowchart of a method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability, provided by an embodiment of the present invention. The specific steps are as follows: S1. Collect baseline, task, and recovery pulse wave signals of the subject during the AI ​​trombone playing training process. Collect the airflow signal of the subject during the playing training through the pickup module. Record the level signal output by the ultrasonic ranging module, i.e. the original position control data. Record the accuracy and reaction time of the inhibition control test task.

[0088] The AI ​​trombone is an intelligent musical instrument obtained by modifying a traditional trombone in the laboratory, and it is mainly used for data acquisition in this invention. The state 5 minutes before the start of playing training is called the baseline state, and the pulse wave signal collected in this state is called the baseline pulse wave signal; the state of playing training is called the task state, and the pulse wave signal collected in the task state is called the task state pulse wave signal; the state 5 minutes after playing training is called the recovery state, and the pulse wave signal collected in the recovery state is called the task state pulse wave signal.

[0089] The subjects performed trombone playing training and inhibitory control testing tasks in a stable, quiet, and comfortable indoor environment. They completed the training and testing tasks on the experimental system according to the operator's instructions. The experimental system's modular structure diagram is shown below. Figure 2 As shown.

[0090] S2. Preprocess the collected physiological and behavioral data.

[0091] In this embodiment, the acquired pulse wave signal is subject to noise interference such as baseline drift, electromyographic noise, and power frequency noise; the blown airflow signal is subject to noise interference such as power frequency noise and impulse noise. The flowchart for physiological and behavioral signal preprocessing is as follows: Figure 2 As shown, the specific steps are as follows: S201. Preprocess the acquired pulse wave signal. The specific steps are as follows: S201.1. Variational mode decomposition is used on the baseline pulse wave signal, the task-state pulse wave signal, and the recovery-state pulse wave signal to obtain 10 IMF components. Fast Fourier transform is performed on each IMF component, and the frequency corresponding to the maximum peak value in the spectrum is calculated. .

[0092] S201.2. Considering that the human heart rate ranges from 45 to 180 bpm, corresponding to a frequency of 0.75 to 3 Hz, for... The IMF components are denoised using wavelet thresholding to filter out high-frequency noise and reduce motion artifacts.

[0093] The wavelet thresholding method used in this embodiment has the following parameters: "db9" wavelet basis function, 5 decomposition levels, global threshold and hard threshold function. The global threshold formula is as follows: ,in, The standard deviation of noise. This is the signal length.

[0094] S201.3. The high-frequency IMF component after noise reduction is superimposed with the remaining IMF component to reconstruct the pulse wave signal.

[0095] S201.4. For the reconstructed pulse wave signal, a fourth-order Butterworth bandpass filter with a passband frequency range of 0.5~3.25Hz is used to further improve the quality and signal-to-noise ratio of the pulse wave signal.

[0096] S201.5. The first-order forward differential reverse traversal method is used to extract the starting point of each pulse wave cycle in the pulse wave signal.

[0097] The execution steps of the first-order forward difference inversion traversal method are as follows: 1) Perform a first-order forward difference operation on the pulse wave signal to obtain the first-order difference sequence of the pulse wave signal, as shown in the following formula:

[0098] 2) Extract the first-order difference sequence with a spacing greater than [missing information]. The maximum point, i.e., the rising peak point of the pulse wave, is where, The sampling rate of the pulse wave signal. The estimated frequency is obtained by performing a Fast Fourier Transform on the pulse wave signal. In this embodiment... .

[0099] 3) Based on the rising peak point, search for the differential zero point in the differential sequence to obtain the starting point of the pulse wave period.

[0100] S201.6 Perform cubic spline interpolation on the pulse wave cycle start point sequence to fit the pulse wave baseline, remove the baseline drift of the baseline state pulse wave signal, the task state pulse wave signal and the recovery state pulse wave signal to obtain the denoised baseline state pulse wave signal, the denoised task state pulse wave signal and the denoised recovery state pulse wave signal.

[0101] S202. The collected airflow signal is processed using... , A notch filter with a bandwidth of 5Hz is used to filter out power frequency noise interference, and a median filter with a window size of 21 is used to filter out impulse noise interference.

[0102] S203. Based on the collected raw position control data, according to the speed of sound... Duration of high level The level signal is converted into distance to obtain the position control data represented by a distance sequence.

[0103] S3. Identify the main peak points of the denoised baseline pulse wave signal, the denoised task pulse wave signal, and the denoised recovery pulse wave signal. Based on this, extract the pulse rate variability features. For the denoised blowing airflow signal, segment it based on the blowing training sound score and extract the blowing breath stability features. Similarly, segment the position control data and extract the position control accuracy features based on statistical principles.

[0104] This embodiment extracts the time-domain, frequency-domain, and nonlinear characteristics of pulse rate variability. Simultaneously, it references commonly used statistical indicators to extract statistical features from the airflow signal and position control data. The flowchart for physiological and behavioral feature extraction is as follows: Figure 4 As shown, the feature extraction process is as follows: S301. Extract pulse rate variability features from the denoised baseline pulse wave signal, the denoised task pulse wave signal, and the denoised recovery pulse wave signal.

[0105] S301.1 Use the first-order forward differential ergodic method to identify the peak point of the main wave of the pulse wave signal.

[0106] S301.2. Perform a difference operation on the main peak point sequence to obtain the time interval sequence between two adjacent pulse wave peaks, i.e., the PP interval sequence.

[0107] S301.3. Use a 21st-order AR model to estimate the frequency domain power spectrum of the PP interval sequence.

[0108] S301.4 Extract the time-domain, frequency-domain, and nonlinear features of pulse rate variability (PRV) from the sequence to obtain the baseline pulse rate variability feature vector. Baseline-free task-state pulse rate variability eigenvector and baseline-recovery state pulse rate variability eigenvector .

[0109] Pulse rate variability features include six time-domain features: PRV_MEANPP, PRV_SDPP, PRV_RMSSD, PRV_PNN20, PRV_PNN50, and PRV_TI; three frequency-domain features: PRV_LF, PRV_HF, and PRV_LF / PRV_HF; and four nonlinear features: PRV_SampEn, PRV_SD1, PRV_SD2, and PRV_SD1 / PRV_SD2.

[0110] Therefore, a total of [number] signals were extracted from the pulse wave signal. , and There are three sets of pulse rate variability feature vectors. These three sets of vectors are distinguished by different prefixes. Taking PRV_MEANPP as an example, BL_PRV_MEANPP is... Vector features, TB_PRV_MEANPP is Vector features, RB_PRV_MEANPP are Vector features.

[0111] S302. For the noise-reduced airflow signal, the signal is segmented according to the notes of the playing training sound score. The short-time energy of each segment is calculated frame by frame. Based on statistical principles, three playing breath stability features, namely STE_MEANCV, MEANSampEn and STE_MEAN_SD, are extracted from the short-time energy sequence.

[0112] S303. For position control data, the data is segmented according to the notes in the playing training sound score. Based on statistical principles, three position control accuracy features, SP_MAE, SP_RMSE and SP_ACC, are extracted.

[0113] S4. Based on histogram analysis and goodness-of-fit test, a probability distribution is fitted to the reaction time of the inhibition control test task. The test task score is obtained by calculating the cumulative probability and combining it with the test task accuracy. The specific steps are as follows: S401. For each inhibition control test task, a reaction time histogram is plotted to analyze the data distribution. Based on current research on reaction time distribution models, this embodiment selects a log-normal distribution to fit the reaction time histogram.

[0114] The probability density function of the log-normal distribution is as follows:

[0115] in, As expected, The standard deviation is denoted as .

[0116] The cumulative probability function of the log-normal distribution is as follows:

[0117] in, This is the cumulative distribution function of the standard normal distribution.

[0118] S402. After passing the AD goodness-of-fit test, accept the null hypothesis, that is, assume that the reaction time of the test task follows a log-normal distribution, and calculate the cumulative probability of the reaction time. .

[0119] S403. Regarding the accuracy of the test tasks, since the accuracy is already a normalized metric, the accuracy of each test task is calculated as the sum of the accuracy and the reaction time of that task. Multiply the scores to get the score for each test task, and add the scores of each test task together to get the score of the test subject for this test task.

[0120] The expression for calculating the test task score is as follows:

[0121] in, , These represent the subjects' first and second subjects, respectively. The accuracy and reaction time of the test task. The range of values ​​is .

[0122] S5. Construct a dataset based on the physiological and behavioral characteristics of the subjects, train a random forest regression model with the test task scores obtained in step S4 as the target variable, evaluate the performance of different feature subsets through recursive feature elimination and cross-validation, and select the feature subset with the highest performance.

[0123] The flowchart for establishing a quantitative evaluation model for intervention effectiveness based on physiological and behavioral characteristics is shown below. Figure 5 As shown.

[0124] S501. Based on the 42 physiological features of the 3 sets of pulse rate variability feature vectors extracted in step S3 and the 6 behavioral features of playing breath stability and position control accuracy, a dataset is constructed.

[0125] S502. Using the test task score obtained in step S4 as the target variable and physiological and behavioral characteristics as explanatory variables, train a random forest regression model.

[0126] This step takes into account that there may be both linear and nonlinear relationships between physiological and behavioral characteristics and the target variable. Directly using linear regression models such as ridge regression may lead to the loss of feature information. Therefore, this embodiment selects the random forest regression model to obtain the predictive importance of each feature to the target variable, providing a reliable basis for subsequent feature selection.

[0127] S503. Using the random forest regression model obtained in step S502 as the base model, recursive feature elimination is used to gradually remove the features with the lowest importance. The RMSE and R-value of the feature subset are calculated through 10-fold cross-validation. 2 The performance of the feature subset is evaluated at each recursion, and finally R is selected. 2 The feature subset with the highest R² and lowest RMSE is selected. If there is a conflict between R² and RMSE, the feature subset with the highest R² is selected first.

[0128] S6. Perform Spearman correlation analysis on the optimal feature subset obtained in S5. Based on the analysis results, perform Min-Max standardization on the features and transform the orientation of all features to be positively correlated with the score of the inhibition test task, that is, positively correlated with the inhibition control ability. Retrain the regression model and construct a quantitative evaluation formula based on feature importance to obtain the quantitative score of "trombone playing training - inhibition control ability".

[0129] The specific steps are as follows: S601. Perform Spearman correlation analysis on the optimal feature subset obtained in step S5.

[0130] S602. Based on the correlation between each feature in the optimal feature subset and the test task score, perform Min-Max standardization on each feature.

[0131] S603. Retrain the regression model and construct a quantitative evaluation formula based on feature importance to obtain a quantitative score for "trombone playing training - inhibitory control ability".

[0132] Example 2 This embodiment further discloses a method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability, including the following steps: S1. Referring to step S1 in Example 1, collect physiological and behavioral data of a subject during the wind playing training process, namely, pulse wave signals at baseline, task state, and recovery state, wind playing airflow signals, and position control data. Simultaneously, record the subject's accuracy and reaction time during the inhibitory control test task. The original pulse wave signal waveform is shown below. Figure 6 As shown, the waveform of the airflow signal during blowing is as follows: Figure 7 As shown.

[0133] S2. Referring to step S2 in Example 1, preprocessed physiological and behavioral data are obtained. A schematic diagram of the position control data sequence in the behavioral data is shown below. Figure 8 As shown.

[0134] S3. Referring to step S3 in Example 1, extract the physiological and behavioral characteristics of the subject. The result of denoising pulse wave signal feature point recognition is shown in the figure below. Figure 9 As shown.

[0135] S4. Referring to step S4 in Example 1, fit the probability distribution during the reaction time, and calculate the score of the inhibition control test task based on the accuracy. The reaction time distribution curve for inconsistent tasks is shown in the figure below. Figure 10 As shown.

[0136] S5. Referring to step S5 in Example 1, recursive feature elimination with cross-validation is used to obtain the optimal feature subset with 4 features. The features of the optimal feature subset include PRV_SDPP, PRV_HF, STE_MEAN_SD, and SP_MAE. The RMSE and R of different feature subsets during recursive feature elimination are... 2 Schematic diagram as follows Figure 11 As shown.

[0137] S6. Based on Example 1, the dimensionless methods for each feature are shown in Table 1: Table 1. Dimensionless Methods

[0138] After dimensionless transformation of each feature in the optimal feature subset, the random forest regression model is retrained to obtain the importance weights of each feature. The feature importance weights are shown in Table 2. Table 2. Feature Importance Weight Table

[0139] In summary, a quantitative score for "trombone playing training - inhibitory control ability" was obtained by weighted summation of feature importance weights. This score explains the influence of physiological and behavioral characteristics on inhibitory control ability during playing training. By analyzing the changes in the quantitative score of each playing training session, a quantitative assessment of the intervention effect of trombone playing training on inhibitory control ability can be achieved.

[0140] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0141] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability, characterized in that, The quantitative evaluation method includes the following steps: S1. Using photoplethysmography (PPG), baseline, task-state, and recovery-state pulse wave signals were collected from the subject during training with an AI trombone. Airflow signals during training were also collected via a pickup module. The output level signal from the ultrasonic ranging module was recorded; this level signal represents the original position control data. The accuracy and reaction time of the inhibition control test task were also recorded. The AI ​​trombone is a smart instrument derived from a modified traditional trombone and used for data acquisition. The state 5 minutes before the start of training is called the baseline state, and the pulse wave signal collected in this state is called the baseline pulse wave signal. The state during training is called the task state, and the pulse wave signal collected in the task state is called the task state pulse wave signal. The state 5 minutes after the end of training is called the recovery state, and the pulse wave signal collected in the recovery state is called the recovery state pulse wave signal. S2. For the baseline, task, and recovery pulse wave signals, a variational mode decomposition and wavelet thresholding combined denoising method is used to filter out high-frequency noise and reduce motion artifacts. Cubic spline interpolation is used to remove baseline drift of the pulse wave signal, resulting in denoised baseline, task, and recovery pulse wave signals. For the wind-playing airflow signal, notch filtering and median filtering are used for denoising, resulting in denoised wind-playing airflow signals. The original position control data is converted to obtain position control data represented by a distance sequence. S3. Identify the main peak points of the denoised baseline pulse wave signal, the denoised task pulse wave signal, and the denoised recovery pulse wave signal, and extract pulse rate variability features based on this. For the denoised blowing airflow signal, segment it based on the blowing training sound score and extract the blowing breath stability features. Similarly, segment the position control data and extract the position control accuracy features based on statistical principles. S4. Based on histogram analysis and goodness-of-fit test, the probability distribution of the reaction time of the inhibition control test task is fitted. The test task score is obtained by calculating the cumulative probability and combining it with the test task accuracy. S5. Construct a dataset based on the physiological and behavioral characteristics of the subjects. Train a random forest regression model using the test task score obtained in step S4 as the target variable. Evaluate the performance of different feature subsets through recursive feature elimination and cross-validation, and select the feature subset with the highest performance, i.e., the optimal feature subset. The physiological characteristics are pulse rate variability features extracted from the denoised baseline pulse wave signal, the denoised task pulse wave signal, and the denoised recovery pulse wave signal. The behavioral characteristics are the breath stability and position control accuracy features extracted from the airflow signal and position control data. S6. Perform correlation analysis on the optimal feature subset obtained in S5, and dimensionlessly transform the features according to the analysis results. Retrain the random forest regression model, construct a quantitative evaluation formula based on feature importance, and obtain a quantitative score for "trombone playing training - inhibitory control ability". Analyze the quantitative score to achieve a quantitative evaluation of the intervention effect of trombone playing training on inhibitory control ability.

2. The method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability according to claim 1, characterized in that, The process of step S1 is as follows: S101. The subject sits in front of a screen, which displays the UI interface of the trombone playing training module. A PPG sensor is placed on the desktop where the screen is located. First, the screen displays a 5-minute countdown and prompts the subject to adjust the sensor position and place the left middle finger on the PPG sensor. The subject sits quietly for 5 minutes to collect the baseline pulse wave signal before training. After the countdown ends, the screen prompts the subject to start playing training. S102. The screen displays the playing training sound score. During the training, the subject is required to keep his upper body upright, wear headphones, hold the slide of the AI ​​trombone with his right hand, hold the grip of the AI ​​trombone with his left hand, and wear a PPG sensor finger sleeve on the middle finger of his left hand. The mouthpiece of the AI ​​trombone is tilted downward at 15 degrees, and the lips are naturally aligned with the mouthpiece of the AI ​​trombone. The PPG sensor finger sleeve is used to collect task-state pulse wave signals. S103. The screen plays a musical score in the form of an animation. The subject needs to play with a stable breath and move the slide to the correct position according to the notes and position prompts displayed in the musical score. The playing training is conducted in 10 rounds. In each round, a musical score is played randomly. The pickup module on the AI ​​trombone is used to collect the airflow signal of the subject when performing the playing training. The level signal output by the ultrasonic ranging module is recorded, which is the original position control data. S104. After the playing training, the screen prompts the subject to remove the finger cot and place the left middle finger on the desktop PPG sensor. A 5-minute countdown is displayed, and the subject is prompted to sit quietly for 5 minutes to collect the subject's restorative pulse wave signal. S105. The subject sits in front of a display screen showing the UI of the inhibition control test task module. The inhibition control test task consists of the Stroop test task and the variant Flanker test task. The screen first prompts the subject to perform the Stroop test task and records the accuracy and reaction time of each test task during the execution. S106. After the subject has completed all the Stroop test tasks, the display screen prompts the subject to perform the variant Flanker test task, and records the accuracy and reaction time of each test task during the execution.

3. The method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability according to claim 1, characterized in that, The process of step S2 is as follows: S201. Preprocess the pulse wave signal acquired in step S1. The specific steps are as follows: S201.

1. Variational mode decomposition is used on the baseline pulse wave signal, the task-state pulse wave signal, and the recovery-state pulse wave signal to obtain multiple intrinsic mode function components (IMFs). For each IMF component, a fast Fourier transform is performed, and the frequency corresponding to the maximum peak value in the spectrum is calculated. ; S201.2, after The high-frequency IMF components are denoised using wavelet thresholding to filter out high-frequency noise and reduce the effects of motion artifacts. S201.

3. The high-frequency IMF component processed in step S201.2 is superimposed with the remaining IMF components to reconstruct the pulse wave signal. S201.

4. For the reconstructed pulse wave signal, use a Butterworth bandpass filter to further improve the pulse wave signal quality and signal-to-noise ratio; S201.5 Extract the starting point of each pulse wave cycle from the pulse wave signal to obtain the pulse wave cycle starting point sequence; S201.6 Perform cubic spline interpolation on the pulse wave cycle start point sequence, fit the pulse wave baseline, remove the baseline drift of the pulse wave signal, and obtain the denoised baseline state pulse wave signal, the denoised task state pulse wave signal, and the denoised recovery state pulse wave signal. S202. The blowing airflow signal collected in step S1 is processed for noise reduction using notch filter and median filter. S203. The original position control data obtained in step S1 is converted into distance based on the sound speed and the duration of the high level, resulting in position control data represented by a distance sequence.

4. The method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability according to claim 1, characterized in that, The process of step S3 is as follows: S301. For the denoised baseline pulse wave signal, the denoised task pulse wave signal and the denoised recovery pulse wave signal, the first-order forward differential ergodic method is used to identify the main peak point of the pulse wave signal. Based on the main peak point sequence, the time interval sequence between two adjacent pulse wave peaks is obtained, and the pulse rate variability feature is extracted from the time interval sequence. S302. The noise-reduced airflow signal obtained in step S2 is segmented based on the measures of the blowing training sound score. The sliding window algorithm is used to divide each signal into frames, and the short-time energy sequence is calculated. The blowing breath stability feature is then extracted from it. S303. The position control data obtained in step S2 is segmented based on the measures of the playing training score, and position control accuracy features are extracted based on statistical principles.

5. The method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability according to claim 4, characterized in that, The process of step S4 is as follows: S401. For each inhibition control test task, plot a reaction time histogram to analyze the data distribution pattern. S402. Through the Anderson-Darling goodness-of-fit test, the target probability distribution model accepting the null hypothesis is selected as the probability distribution for reaction time, and the cumulative probability of reaction time is obtained. The null hypothesis is that the reaction time follows a target probability distribution. S403. Regarding the accuracy of the test tasks, since the accuracy is already a normalized metric, the accuracy of each test task is calculated as the sum of the accuracy and the reaction time of that task. Multiply the scores to get the score for each test task, and add the scores of each test task together to get the score of the test subject for this test task.

6. The method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability according to claim 5, characterized in that, The process of step S5 is as follows: S501. Construct a dataset based on the pulse rate variability features, playing breath stability features, and position control accuracy features extracted in step S3. S502. Using the test task score obtained in step S4 as the target variable and physiological and behavioral characteristics as explanatory variables, train a random forest regression model. S503. Using the random forest regression model obtained in step S502 as the base model, recursive feature elimination is used to progressively remove the features with the lowest importance. The root mean square error (RMSE) and coefficient of determination (R²) of the feature subset are calculated through cross-validation. 2 The performance of the feature subset is evaluated at each recursion, and finally R is selected. 2 The feature subset with the highest R² and lowest RMSE is selected. If there is a conflict between R² and RMSE, the feature subset with the highest R² is selected first.

7. The method for quantitatively evaluating the intervention effect of trombone playing training on inhibitory control ability according to claim 6, characterized in that, The process of step S6 is as follows: S601. Perform Spearman correlation analysis on the optimal feature subset obtained in step S5 to obtain the positive and negative correlation between each feature in the subset and the test task score. S602. Based on the positive and negative correlations obtained in step S601, perform Min-Max standardization on each feature; S603. Retrain the random forest regression model, construct a quantitative evaluation formula based on feature importance, and obtain a quantitative score for "trombone playing training - inhibitory control ability".

Images