A feature screening system for assessing the emotional interference effect of impulse sounds

Abstract

The application discloses a kind of feature screening systems for evaluating the emotional interference effect of impulse sound, it is related to cognitive neuroscientific technology field, it solves the problem that there is no effective screening impulse sound characteristics affecting emotion in prior art.The system comprises: sound generation module, for the independent variable of impulse sound is cross designed to form multiple sound stimuli, and multiple subjects are respectively experimented using multiple sound stimuli;Data acquisition module is used to collect experimental data in the process of within-subject experiment;The experimental data include the physiological signal of each subject under each sound stimulus, emotional valence score and arousal score;Data processing module is used to process experimental data in the process of within-subject experiment, and the dependent variable of impulse sound under each sound stimulus is obtained;Data analysis module determines the feature screening result of impulse sound inducing emotion by analyzing the correlation between the independent variable and dependent variable of impulse sound under sound stimulus.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cognitive neuroscience, and particularly relates to a feature screening system for evaluating the emotional interference effect of impulse sound. BACKGROUND

[0002] Noise or unwanted sound is an important source of annoyance and distress. With the continuous advancement of urbanization and industrialization, noise is becoming more and more common in living and working environments, such as factory, traffic, car, air conditioner noise, etc. In particular, many machines commonly used in working environments, such as fans, blowers, compressors, etc., will produce low-frequency noise with a frequency below 100 Hz. It is worth noting that many noises exist in the form of impulse sound. Previous studies have shown that emotions are easily affected by noise, and the higher the noise annoyance, the worse the emotional state. A bad emotional state will negatively affect problem solving, and work performance will also be poorer. If the emotional state is good, the way of looking at problems will be more optimistic, and people in this state are more likely to accept the surrounding environment, and their work performance will also be higher.

[0003] Studies have shown that the valence and arousal dimensions are suitable for subjective description of emotional response to sound. In addition to subjective evaluation, emotions can be expressed through a variety of physiological information, and various information features can be analyzed to evaluate the emotional state of participants. Most studies focus on the analysis of facial expressions or speech, but these types of signals (more or less) are easy to fake. In addition, emotions caused by sound can also be evaluated through physiological signals of the peripheral nervous system (such as heart rate, electromyography, skin resistance) and the central nervous system (such as electroencephalogram). Although subjective indicators for evaluating emotions can distinguish different emotional types, they have the disadvantage that emotions tend to dissipate when introspected, and the intensity of emotions decreases when recalled. Therefore, the accuracy of emotion evaluation based solely on subjective experience is not high. Emotion evaluation based on physiological indicators is not affected by the subjectivity of the measured person, and does not have the problem of intentional or unintentional non-reporting of emotions by the subject, but it is not easy to distinguish various emotions. Therefore, it is more reliable to evaluate emotions by combining subjective evaluation with physiological indicators.

[0004] Although there are many studies on the influence of sound on emotions, there are fewer studies on the influence of impulse sound on emotions, and even fewer studies on what features of impulse sound have a decisive influence on emotions. Therefore, how to effectively screen the features of impulse sound that affect emotions, so as to better use impulse sound to regulate emotions, is a technical problem that needs to be solved at present. SUMMARY

[0005] In view of the above analysis, the embodiments of the present application aim to provide a feature screening system for evaluating the emotional interference effect of impulse sound, to solve the problem of lack of effective screening of impulse sound features that affect emotions in the prior art.

[0006] The application discloses a feature screening system for evaluating emotional interference effect of impulse sound, and the system comprises:

[0007] a sound generating module, which is used for cross-designing independent variables of impulse sound to form multiple sound stimuli, and conducting experiments on multiple subjects respectively by using the multiple sound stimuli; the independent variables comprise sound frequency, repetition frequency and relative loudness;

[0008] a data collecting module, which is used for collecting experimental data in the process of the within-subject experiment; the experimental data comprise physiological signals, emotional valence scores and arousal scores of each subject under each sound stimulus;

[0009] a data processing module, which is used for processing the experimental data in the process of the within-subject experiment to obtain dependent variables of impulse sound under each sound stimulus; the dependent variables comprise average physiological statistical features, emotional valence score mean value and arousal score mean value;

[0010] a data analysis module, which is used for determining feature screening results of impulse sound for inducing emotion by analyzing the correlation between the independent variables and the dependent variables of impulse sound under the sound stimulus.

[0011] On the basis of the above scheme, the application further makes the following improvements:

[0012] Further, the data collecting module comprises a multi-lead physiological recorder, a digital keyboard and a data storage; wherein,

[0013] the multi-lead physiological recorder is used for collecting physiological signals of each subject under each sound stimulus;

[0014] the digital keyboard is used for collecting emotional valence scores and arousal scores of each subject under each sound stimulus;

[0015] the data storage is connected with the multi-lead physiological recorder and the digital keyboard, and is used for storing the physiological signals, the emotional valence scores and the arousal scores of each subject under each sound stimulus.

[0016] Further, the data processing module comprises a data preprocessing unit and a feature extraction unit; wherein,

[0017] the data preprocessing unit is used for sequentially preprocessing the physiological signals of each subject under each sound stimulus respectively;

[0018] the feature extraction unit is used for extracting features from the preprocessed physiological signals to obtain physiological statistical features of the corresponding subject under the corresponding sound stimulus.

[0019] Further, the physiological signals include electrocardiogram signals and corrugator electromyogram signals; wherein, in the data preprocessing unit:

[0020] The preprocessing of the electrocardiogram signals includes, in sequence, removing baseline drift, removing power frequency noise, and removing irrelevant electromyogram noise.

[0021] The preprocessing of the corrugator electromyogram signals includes, in sequence, filtering, removing power frequency noise, rectifying, smoothing, and normalizing.

[0022] The rectifying is configured to change all negative half waves of the corrugator electromyogram signal after removing power frequency noise into positive half waves with unchanged amplitudes.

[0023] Further, in the process of conducting experiments on multiple subjects respectively, N test times are set for each sound stimulus, each test time is taken as a condition, N is a positive integer, and all conditions are randomly presented.

[0024] In the process of playing the sound stimulus in each test time, the physiological signals of the subject in the corresponding test time are collected.

[0025] After the playing of the sound stimulus in each test time ends, the emotional valence score and the arousal degree score of the corresponding test time are collected.

[0026] Further, the physiological statistical features include heart rate features, heart rate variability features, and corrugator electromyogram features; and the average physiological statistical features include average heart rate features, average heart rate variability features, and average corrugator electromyogram features.

[0027] In the feature extraction unit, the following is performed:

[0028] The heart rate and the heart rate variability are extracted from the preprocessed electrocardiogram signals; the heart rate and the heart rate variability of each subject under each sound stimulus in all test times are averaged respectively to obtain the heart rate features and the heart rate variability features of the corresponding subject under the corresponding sound stimulus.

[0029] The average corrugator electromyogram amplitude is extracted from the preprocessed corrugator electromyogram signals; the average corrugator electromyogram amplitude of each subject under each corresponding sound stimulus in all test times is averaged to obtain the corrugator electromyogram features of the corresponding subject under the corresponding sound stimulus.

[0030] The heart rate features, the heart rate variability features, and the corrugator electromyogram features of all subjects under each sound stimulus are averaged respectively to obtain the average heart rate features, the average heart rate variability features, and the average corrugator electromyogram features under the corresponding sound stimulus.

[0031] Further, the data processing module further includes a score processing unit.

[0032] The score processing unit is used for averaging the emotional valence scores and arousal scores of all subjects under each sound stimulus respectively to obtain the average emotional valence score and the average arousal score under the corresponding sound stimulus.

[0033] Further, in the data analysis module, the feature screening result of the pulse sound inducing emotion is determined by using repeated measurement variance analysis method, or a combination of multivariate linear regression analysis method and Durbin-Watson test method.

[0034] Further, when the data analysis module determines the feature screening result of the pulse sound inducing emotion by using the repeated measurement variance analysis method, the following is performed:

[0035] The relationship between the independent variables and the dependent variables of the pulse sound under various sound stimuli is subjected to repeated measurement variance analysis;

[0036] The repeated measurement variance analysis is subjected to spherical test, and if the spherical test fails, the degrees of freedom and the statistical value in the result of the repeated measurement variance analysis are corrected to update the result of the repeated measurement variance analysis;

[0037] According to the result of the repeated measurement variance analysis, the correlation between the independent variables and the dependent variables of the pulse sound is evaluated, and the independent variable with the greatest impact on emotion in the pulse signal is screened as the feature screening result of the pulse sound inducing emotion.

[0038] Further, when the data analysis module determines the feature screening result of the pulse sound inducing emotion by using the combination of the multivariate linear regression analysis method and the Durbin-Watson test method, the following is performed:

[0039] A hierarchical regression model is established by using the multivariate linear regression analysis method, and the independent variables of the pulse sound are added one by one to investigate the relationship between the independent variables and the dependent variables of the pulse sound under various sound stimuli;

[0040] The R 2 The explanatory rate of the independent variables of the pulse sound under different sound stimuli on the dependent variables of the corresponding pulse sound is investigated by taking R

[0041] Compared with the prior art, the present application can achieve at least one of the following beneficial effects:

[0042] The feature screening system for evaluating the emotional interference effect of the pulse sound provided by the present application can effectively screen the main pulse sound features affecting emotion, so that the pulse sound can be better used to regulate emotion, and the problem of lacking effective screening of pulse sound features affecting emotion in the prior art is well solved.

[0043] The technical solutions in the present application can be combined with each other to realize more preferred combination solutions. Other features and advantages of the present application will be described in the following description, and some advantages will become apparent from the description, or will be understood by those skilled in the art through implementation of the present application. The objects and other advantages of the present application can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF DRAWINGS

[0044] The accompanying drawings are included to provide a further understanding of the present application, and are incorporated herein and constitute a part of the detailed description. The same reference numbers in different drawings refer to the same elements throughout the drawings.

[0045] Figure 1 A structural schematic diagram of a feature screening system for evaluating the emotional interference effect of impulse sound provided by an embodiment of the present application;

[0046] Figure 2 A flowchart of an indoor experiment provided by an embodiment of the present application;

[0047] Figure 3 A flowchart of an experimental trial provided by an embodiment of the present application;

[0048] Figure 4 A comparison schematic diagram of valence of a subject under different sound stimuli provided by an embodiment of the present application;

[0049] Figure 5 A comparison schematic diagram of arousal of a subject under different sound stimuli provided by an embodiment of the present application;

[0050] Figure 6 A comparison schematic diagram of corrugator muscle electrical amplitude (RMS) of a subject under different sound stimuli provided by an embodiment of the present application;

[0051] Figure 7 A comparison schematic diagram of heart rate (BMP) of a subject under different sound stimuli provided by an embodiment of the present application. DETAILED DESCRIPTION

[0052] The preferred embodiments of the present application will be described in detail below with reference to the accompanying drawings, which form a part of this application. The accompanying drawings illustrate the principles of the present application and, together with the description, serve to explain the present application.

[0053] One specific embodiment of the present application discloses a feature screening system for evaluating the emotional interference effect of impulse sound, a structural schematic diagram of which is shown in FIG. 1. Figure 1 The system comprises:

[0054] a sound generating module, configured to cross design independent variables of the impulse sound to form a plurality of sound stimuli, and conduct experiments on a plurality of subjects respectively by using the plurality of sound stimuli; the independent variables include sound frequency, repetition frequency and relative loudness;

[0055] a data collecting module, configured to collect experimental data in an intra-subject experiment; in the intra-subject experiment, cross design independent variables of the impulse sound to form a plurality of sound stimuli, and conduct experiments on a plurality of subjects respectively by using the plurality of sound stimuli; the independent variables include sound frequency, repetition frequency and relative loudness; the experimental data include physiological signals, emotional valence scores and arousal scores of each subject under each sound stimulus;

[0056] a data processing module, configured to process the experimental data in the intra-subject experiment to obtain dependent variables of the impulse sound under each sound stimulus; the dependent variables include average physiological statistical characteristics, emotional valence score mean value and arousal score mean value;

[0057] a data analysis module, configured to determine a characteristic screening result of the impulse sound inducing emotion by analyzing a correlation between the independent variables and the dependent variables of the impulse sound under the sound stimulus.

[0058] It should be noted that in the embodiment, under the intra-subject experiment condition, a plurality of impulse sounds are set according to a scene inducing emotion; different sound frequencies, repetition frequencies and relative loudnesses are cross designed to form a plurality of sound stimuli. In the process of conducting experiments on a plurality of subjects respectively, N test times are set for each sound stimulus, each test time is regarded as a condition, N is a positive integer; all conditions are randomly presented; in the process of playing the sound stimulus in each test time, physiological signals of the subject in the corresponding test time are collected; after the playing of the sound stimulus in each test time is completed, emotional valence scores and arousal scores of the corresponding test time are collected.

[0059] Next, each module in the embodiment is described as follows:

[0060] (1) Sound generating module

[0061] In the specific implementation process, the sound generating module can be implemented by using an audio device, and the corresponding sound stimulus is formed by adjusting the sound frequency, repetition frequency and relative loudness of the audio device.

[0062] (2) Data collecting module

[0063] The data collecting module includes a multi-lead physiological recorder, a digital keyboard and a data storage device; wherein,

[0064] The multi-lead physiological recorder is configured to collect physiological signals of each subject under each sound stimulus;

[0065] a digital keyboard, used for collecting the emotional valence score and the arousal score of each subject under each sound stimulus;

[0066] a data storage, connected with the multi-lead physiological recorder and the digital keyboard, used for storing the physiological signal, the emotional valence score and the arousal score of each subject under each sound stimulus.

[0067] It should be noted that, considering that the experiment process of each subject under each sound stimulus also includes multiple trials, therefore, each component in the data collection module can collect the above information with the trial as the minimum object. That is, the multi-lead physiological recorder is used for collecting the physiological signal of each trial of each subject under each sound stimulus; the digital keyboard is used for collecting the emotional valence score and the arousal score of each trial of each subject under each sound stimulus; the data storage, connected with the multi-lead physiological recorder and the digital keyboard, is used for storing the physiological signal, the emotional valence score and the arousal score of each trial of each subject under each sound stimulus.

[0068] (3) data processing module

[0069] The data processing module includes a data preprocessing unit and a feature extraction unit; wherein,

[0070] The data preprocessing unit is used for sequentially preprocessing the physiological signal of each subject under each sound stimulus. The physiological signal includes an electrocardiogram signal and a frown electromyogram signal. In the data preprocessing unit, the preprocessing of the electrocardiogram signal sequentially includes removing baseline drift, removing power frequency noise and removing irrelevant electromyogram noise; the preprocessing of the frown electromyogram signal sequentially includes filtering, removing power frequency noise, rectification, smoothing and normalization; the rectification is used for changing all negative half waves of the frown electromyogram signal after removing power frequency noise into positive half waves with unchanged amplitude.

[0071] The feature extraction unit is used for extracting features from the preprocessed physiological signal to obtain the physiological statistical features of the corresponding subject under the corresponding sound stimulus. The physiological statistical features include heart rate features, heart rate variability features and frown electromyogram features; the average physiological statistical features include average heart rate features, average heart rate variability features and average frown electromyogram features. In the feature extraction unit, the following is performed:

[0072] extracting the heart rate and the heart rate variability from the preprocessed electrocardiogram signal; averaging the heart rate and the heart rate variability of each subject under each sound stimulus in all trials to obtain the heart rate features and the heart rate variability features of the corresponding subject under the corresponding sound stimulus;

[0073] The mean value of the corrugator muscle electrical amplitude is extracted from the preprocessed corrugator muscle electrical signal; the mean values of the corrugator muscle electrical amplitude of each subject under each corresponding sound stimulus in all test times are averaged to obtain the corrugator muscle electrical characteristics of the corresponding subject under the corresponding sound stimulus;

[0074] The heart rate characteristics, the heart rate variability characteristics, and the corrugator muscle electrical characteristics of all subjects under each sound stimulus are averaged respectively to obtain the average heart rate characteristics, the average heart rate variability characteristics, and the average corrugator muscle electrical characteristics under the corresponding sound stimulus.

[0075] In addition, the data processing module further comprises a score processing unit; the score processing unit is configured to average the emotional valence scores and the arousal scores of all subjects under each sound stimulus respectively to obtain the mean values of the emotional valence scores and the arousal scores under the corresponding sound stimulus.

[0076] (4) Data analysis module

[0077] In the data analysis module, the feature screening result of the pulse sound inducing emotion is determined by using repeated measurement variance analysis method, or a combination of multivariate linear regression analysis method and Debin-Watson test method.

[0078] Preferably, when the data analysis module determines the feature screening result of the pulse sound inducing emotion by using the repeated measurement variance analysis method, the following is performed:

[0079] The relationship between the independent variables and the dependent variables of the pulse sound under various sound stimuli is subjected to repeated measurement variance analysis.

[0080] Whether the repeated measurement variance analysis passes is tested by using spherical test method; if the spherical test does not pass, the degrees of freedom and the statistical values in the result of the repeated measurement variance analysis are corrected to update the result of the repeated measurement variance analysis.

[0081] According to the result of the repeated measurement variance analysis, the correlation between the independent variables and the dependent variables of the pulse sound is evaluated, and the independent variable with the greatest impact on emotion in the pulse signal is screened out as the feature screening result of the pulse sound inducing emotion.

[0082] Preferably, when the data analysis module determines the feature screening result of the pulse sound inducing emotion by using the combination of the multivariate linear regression analysis method and the Debin-Watson test method, the following is performed:

[0083] A hierarchical regression model is established by using the multivariate linear regression analysis method, and the independent variables of the pulse sound are added one by one to investigate the relationship between the independent variables and the dependent variables of the pulse sound under various sound stimuli.

[0084] The R 2Using indicators as a reference, we examined the explanatory power of the independent variables of impulse sounds under different sound stimuli on the dependent variables of the corresponding impulse sounds (the higher the explanatory power, the more important the independent variable); combined with the Durbin-Watson test, we selected the independent variables in the impulse signals that had the greatest impact on emotions as the feature selection results of impulse sounds that induce emotions.

[0085] To facilitate better implementation of this solution by those skilled in the art, the present invention also provides a specific implementation process for pulse sound feature selection using a feature selection system for evaluating the effect of pulse sound on emotional interference, as shown in the flowchart below. Figure 2 As shown, the explanation is as follows.

[0086] (1) Experimental task design

[0087] This experiment employed a within-subjects experimental design, a true experimental design where each or every group of subjects receives experimental treatment at all levels of the independent variable; also known as a repeated measures design or within-subjects design. In this embodiment, the following settings were made based on the emotionally evoking scenario: sound frequencies were set to 5000Hz, 10000Hz, and 15000Hz; repetition frequencies were set to 1Hz, 100Hz, and 400Hz; and relative loudness was set to -12dB and -24dB. The pulse width was controlled at 100μs. A cross-design was used for sound frequency, repetition frequency, and relative loudness, resulting in a total of 18 sound stimuli. The dependent variable for the impulse noise was the physiological signal of the subjects hearing the sound stimuli, as well as their emotional valence and arousal scores (i.e., the scores given by the subjects for assessing emotional valence and arousal). In the specific implementation process, a 9-point subscale can be used to quantify the emotional valence score and arousal score. For example: valence is a 9-point subscale, ranging from -4 to 4, where -4 represents very unpleasant and 4 represents very pleasant; arousal is a 9-point subscale, ranging from -4 to 4, where -4 represents very calm and 4 represents very excited.

[0088] (2) Experimental Procedure

[0089] ① First, present the experimental instructions to the subjects, explaining the experimental content and reaction methods.

[0090] ② Secondly, participants were asked to complete a practice experiment, totaling 10 trials. The procedure for each trial was as follows: a fixation point was presented in the center of the screen for 1000ms, followed by a 1000ms blank screen. This format is commonly used in psychology and cognitive science research; presenting the fixation point helps participants focus their attention, while presenting the blank screen reduces interference from other factors, facilitating better focus on the subsequent auditory stimulus. The auditory stimulus was then played for 5 seconds. After the auditory stimulus finished playing, participants were asked to rate the valence and arousal of the emotion displayed on the screen on a 9-point scale, based on the emotion they felt upon hearing the auditory stimulus, by moving the mouse. Then, the next trial began.

[0091] ③Finally, after each subject is familiar with the requirements of the practice experiment, enter the formal experiment. Process with practice test, 20 test times per condition. A total of 360 test times. Each condition is randomly presented. Once in the middle of the formal experiment, about 3-5 minutes.

[0092] The flow chart of the experimental test is shown in Figure 3

[0093] (3) Data acquisition

[0094] This process corresponds to the data acquisition module. The MP150 multi-channel physiological recorder of the physiological signal BIOPAC collects the physiological data of the subjects, with a sampling rate of 1000Hz, using a high-pass filter of 0.5Hz and a low-pass filter of 35Hz. Record the physiological signals of the subjects during the sound stimulus playing process, including the corrugator electromyogram, electrocardiogram signal (heart rate, heart rate variability obtained after processing), etc. The data collected by the multi-channel physiological recorder does not include the data of the evaluation stage (i.e. when the subjects evaluate their emotional feelings when hearing the sound), so as not to cause confusion.

[0095] (4) Data processing

[0096] This process corresponds to the data processing module.

[0097] 1) Electrocardiogram signal preprocessing:

[0098] ① Remove baseline drift: First, design two median filters, i.e. pass the electrocardiogram signal through a 200ms median filter to remove QRS complex and P wave; then pass it through a 600ms median filter to remove T wave, and obtain the drifting baseline; finally, subtract the drifting baseline from the electrocardiogram signal to obtain the signal after filtering the baseline drift.

[0099] ② Remove power frequency noise: Power frequency interference has a fixed frequency, and a 50Hz notch filter is used to notch process the signal after filtering the baseline drift, filtering out power frequency interference.

[0100] ③ Remove irrelevant electromyogram noise: Irrelevant electromyogram noise (such as some electromyogram noise produced by movement, blinking or sweating) interference belongs to high frequency interference, while the frequency of electrocardiogram signal is mainly concentrated in 5-20Hz, so a 5-20Hz band-pass filter is used to filter the electrocardiogram signal after filtering the power frequency interference, to reduce the influence of irrelevant electromyogram noise.

[0101] 2) Corrugator electromyogram signal preprocessing:

[0102] ① Filtering: Since the frequency spectrum of the corrugator electromyogram signal is mainly distributed between 28-500Hz, use a 28-500Hz band-pass filter to filter the corrugator electromyogram signal;

[0103] ​② Remove power frequency noise: use 50Hz notch filter to filter the filtered frown muscle electrical signal, filter out power frequency interference;

[0104] ③ Rectification: all negative half waves in the signal are changed to positive half waves, and the amplitude remains unchanged (i.e. retain its amplitude).

[0105] ④ Smooth: use moving average method to smooth the rectified signal to remove high frequency noise.

[0106] ⑤ Normalization: normalize the smoothed signal to make its amplitude range between 0 and 1.

[0107] Subsequently, the heart rate and heart rate variability can be extracted from the preprocessed electrocardio signal by using the peak detection algorithm (see the existing method, which will not be described here), and the frown muscle electrical amplitude mean value can be extracted from the preprocessed frown muscle electrical signal (i.e. all frown muscle electrical amplitudes in the preprocessed frown muscle electrical signal are averaged to obtain the frown muscle electrical amplitude mean value). The heart rate and heart rate variability of each subject under each sound stimulus are averaged in all trials to obtain the heart rate characteristics and heart rate variability characteristics of each subject under each sound stimulus. The frown muscle electrical amplitude mean value of each subject under each sound stimulus is averaged in all trials to obtain the frown muscle electrical characteristics of each subject under each sound stimulus. Then, the heart rate characteristics, heart rate variability characteristics and frown muscle electrical characteristics of all subjects under each sound stimulus are averaged to obtain the average heart rate characteristics, average heart rate variability characteristics and average frown muscle electrical characteristics under each sound stimulus. In addition, the emotional valence scores and arousal scores of all subjects under each sound stimulus are directly averaged to obtain the emotional valence score mean value and arousal score mean value under each sound stimulus.

[0108] (5) Data statistical analysis

[0109] The process corresponds to the data analysis module. The changes of emotional scores (emotional valence score mean value and arousal score mean value) and physiological indicators (average heart rate characteristics, average heart rate variability characteristics and average frown muscle electrical characteristics) induced by different sound stimuli are analyzed by repeated measurement variance analysis. When the sphericity test fails, the degrees of freedom F and the statistical value p are corrected by using the greenhouse-geisser method. The sphericity test is used to test whether the variances of the differences between different measurements are equal. If the sphericity test p value is less than 0.05, it fails, and the p value needs to be corrected when the sphericity test fails.

[0110] Secondly, the hierarchical regression model can also be established using multiple linear regression analysis algorithm, and the independent variables of pulse sound are gradually included to investigate the relationship between different sound stimuli, physiological indicators and emotional scores (i.e., the relationship between the independent variables and dependent variables of pulse sound under various sound stimuli). Meanwhile, R 2 indicators are used as references to investigate the explanatory rate of each sound stimulus on physiological indicators and emotional scores. The Durbin-Watson test is used to test whether the residual term has autocorrelation phenomenon, and the collinearity diagnosis is selected to judge whether the multiple variables are related to each other and the degree of correlation. After testing, the data of the present study meet all the assumptions of linear regression.

[0111] (6) Research results

[0112] ① Analysis results of emotional valence

[0113] Firstly, the 3(frequency: 5000hz, 10000hz, 15000hz) x 3(repetition frequency: 1hz, 100hz, 400hz) x 2(loudness: -12db, -24db) repeated measurement analysis of variance of emotional valence data of all subjects is performed.

[0114] The results show that the main effect of frequency (i.e., sound frequency) is significant, with a degree of freedom F(2, 38) = 30.16, a statistical value p < 0.001, and a significant value effect η 2 p = 0.61. After comparing the emotional valence of subjects under three frequency conditions, it is found that the emotional valence level of subjects is the lowest when 5000hz sound is presented, and the emotional valence level of subjects is the highest when 15000hz sound is presented, i.e., the lower the frequency of presented sound, the more negative emotions the subjects show. The main effect of repetition frequency (i.e., repetition frequency) is significant, F(2, 38) = 53.70, p < 0.001, η 2 p = 0.74. After comparing the two pairs of repetition frequency conditions, it is found that the subjects will produce positive emotions when the repetition frequency of pulse sound is 1hz, but the subjects will produce negative emotions when the repetition frequency of pulse sound is 100hz and 400hz, and the emotional valence level is the lowest when the repetition frequency is 100hz. The main effect of loudness (i.e., relative loudness) is significant, F(1, 39) = 94.29, p < 0.001, η 2 p = 0.71. After comparing the emotional valence of subjects under two relative loudness conditions, it is found that the subjects will produce more negative emotions when -12db loudness sound is presented. The results of the main effect of each variable are shown in Figure 4

[0115] ​The results of multiple regression analysis showed that the degree of variation in emotional valence explained by the loudness index was 6%, the degree of variation in emotional valence explained by the frequency variable was 6%, and the degree of variation in emotional valence explained by the repeated frequency variable was 44%, which increased the prediction of emotional valence by 38%. Therefore, the repeated frequency had a more significant impact on emotional valence.

[0116] ② Analysis results of emotional arousal

[0117] The results of repeated measures ANOVA showed that the main effect of frequency was significant, F(2, 38) = 18.32, p < 0.001, η 2 p = 0.49. After further pairwise comparison of the three conditions, it was found that the arousal level induced by the 5000hz sound was greater than that induced by the 10000hz sound, and the arousal level induced by the 10000hz sound was greater than that induced by the 5000hz sound. The main effect of repeated frequency was significant, F(2, 38) = 35.82, p < 0.001, η 2 p = 0.65. The significant difference of repeated frequency was that the arousal level of the subjects was the lowest when the repeated frequency of the presented sound was 1hz, and the arousal level of the subjects was the highest when the repeated frequency of the presented sound was 100hz. The main effect of loudness was significant, F(1, 39) = 63.37, p < 0.001, η 2 p = 0.62. The specific performance of the subjects under different loudness conditions was that the arousal level of the subjects when presented with -12db loudness sound was greater than that when presented with -24db loudness sound. For the results of the main effects of each variable, see Figure 5 .

[0118] The results of multiple regression analysis showed that the degree of variation in emotional arousal explained by the loudness index was 4%, the degree of variation in emotional arousal explained by the frequency variable was still 4%, and the degree of variation in emotional arousal explained by the repeated frequency variable was 29%, which increased the prediction degree by 25%. Therefore, the emotional arousal was most affected by the repeated frequency variable.

[0119] ③ Analysis results of corrugator muscle electromyography (EMG) amplitude

[0120] The results of repeated measures ANOVA showed that the main effect of repeated frequency was significant, F(2, 34) = 4.41, p < 0.05, η 2 p = 0.21. After pairwise comparison, it was found that the corrugator muscle electromyography amplitude of the subjects under the condition of 400hz repeated frequency sound was greater than that under the condition of 1hz repeated frequency sound, and the corrugator muscle electromyography amplitude of the subjects was the smallest when the repeated frequency sound was 1hz. The main effect of loudness was also significant, F(1, 35) = 7.06, p < 0.05, η2 p = 0.17. After analyzing the EMG data under different loudness conditions, it was found that the amplitude of the corrugator muscle of the subjects was greater when the sound was presented at -24 db than at -12 db. The frequency main effect was not significant, F(2, 34) = 1.43, p = 0.25, indicating that the changes in the corrugator muscle of the subjects were not significantly different under different frequency sounds. This shows that the repetition frequency and loudness of the sound can independently affect the changes in the corrugator muscle, but the effect of frequency is not obvious. The results of the main effect of each variable are shown in Table 2. Figure 6 .

[0121] The results of the multiple regression analysis showed that loudness could explain 1% of the changes in the corrugator muscle, and after including the frequency variable in the model, the degree of explanation of the changes in the corrugator muscle remained at 1%. After adding the repetition frequency variable, it was found that it could predict the changes in the corrugator muscle by 2%, an increase of 1%. Therefore, loudness and repetition frequency can be used as influencing factors to predict the changes in the corrugator muscle.

[0122] (4) Analysis results of heart rate (BPM)

[0123] The results of the repeated measures ANOVA showed that the main effect of repetition frequency was significant, F(2, 31) = 16.26, p < 0.001, η 2 p = 0.51. Further analysis showed that the heart rate of the subjects was the fastest under the condition of 100 Hz, and the slowest under the condition of 1 Hz. The main effect of frequency was marginally significant, F(2, 31) = 16.26, p = 0.09, η 2 p = 0.15, the heart rate of the subjects was faster when they heard 15000 Hz sound than when they heard 5000 Hz sound, and the heart rate was faster when they heard 5000 Hz sound than when they heard 10000 Hz sound. However, the main effect of loudness was not significant, F(1, 32) = 2.13, p = 0.16, indicating that there was no significant difference in the heart rate of the subjects when they heard sounds of different loudness. The results of BPM under different experimental conditions are shown in Table 3. Figure 7 .

[0124] The model established by multiple regression analysis showed that the independent effects of loudness, frequency, and repetition frequency on heart rate changes were not obvious.

[0125] Combining the results of the subjective evaluation scale and the objective physiological indicators, it can be seen that although multiple characteristics of the pulse sound can affect emotions, the effect of repetition frequency on emotions is the most obvious, so in the follow-up, repetition frequency can be mainly selected as the main research factor to regulate emotions.

[0126] To sum up, the feature screening system for evaluating the emotional interference effect of the pulse sound can effectively screen the pulse sound features affecting the emotion, so that the pulse sound is better used to regulate the emotion.

[0127] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer readable storage medium. The computer readable storage medium is a disk, an optical disk, a read-only memory, a random access memory, etc.

[0128] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited to this. Any changes or replacements within the technical range disclosed by the present application can be easily thought of by those skilled in the art, and should be covered within the protection scope of the present application.

Claims

1. A feature screening system for assessing the emotional interference effect of impulse sound, characterized by, The system comprises: a sound generating module for cross-designing independent variables of the pulsed sound to form a plurality of sound stimuli, and conducting experiments on a plurality of subjects respectively by using the plurality of sound stimuli; the independent variables include sound frequency, repetition frequency and relative loudness; a data acquisition module for acquiring experimental data in the process of the within-subject experiment; the experimental data include physiological signals, emotional valence scores and arousal scores of each subject under each sound stimulus; a data processing module for processing the experimental data in the process of the within-subject experiment to obtain dependent variables of the pulsed sound under each sound stimulus; the dependent variables include average physiological statistical characteristics, emotional valence score average and arousal score average; a data analysis module for screening the independent variable of the pulsed signal that has the greatest impact on emotion as a characteristic screening result of the pulsed sound for inducing emotion by analyzing the correlation between the independent variables and the dependent variables of the pulsed sound under the sound stimulus.

2. The trait screening system of claim 1, wherein, The data acquisition module comprises a multi-lead physiological recorder, a digital keyboard and a data storage; wherein, the multi-lead physiological recorder is used for acquiring physiological signals of each subject under each sound stimulus; the digital keyboard is used for acquiring emotional valence scores and arousal scores of each subject under each sound stimulus; the data storage is connected with the multi-lead physiological recorder and the digital keyboard, and is used for storing the physiological signals, emotional valence scores and arousal scores of each subject under each sound stimulus.

3. The trait screening system of claim 2, wherein, The data processing module comprises a data preprocessing unit and a feature extraction unit; wherein, the data preprocessing unit is used for sequentially preprocessing the physiological signals of each subject under each sound stimulus respectively; the feature extraction unit is used for extracting features from the preprocessed physiological signals to obtain physiological statistical characteristics of the corresponding subject under the corresponding sound stimulus.

4. The trait screening system of claim 3, wherein, The physiological signals include electrocardiogram signals and frown electromyogram signals; wherein, in the data preprocessing unit: the preprocessing of the electrocardiogram signals sequentially includes removing baseline drift, removing power frequency noise and removing irrelevant electromyogram noise; the preprocessing of the frown electromyogram signals sequentially includes filtering, removing power frequency noise, rectification, smoothing and normalization; the rectification is used for changing all negative half waves of the frown electromyogram signal after removing power frequency noise into positive half waves with unchanged amplitude.

5. The trait screening system of claim 4, wherein, In the process of conducting experiments on the plurality of subjects respectively, N test times are set for each sound stimulus, each test time is regarded as a condition, N is a positive integer; all conditions are randomly presented; in the process of playing the sound stimulus in each test time, the physiological signals of the subject in the corresponding test time are acquired; after the playing of the sound stimulus in each test time is completed, the emotional valence scores and arousal scores of the corresponding test time are acquired.

6. The trait screening system of claim 5, wherein, The physiological statistical characteristics include heart rate characteristics, heart rate variability characteristics and frown electromyogram characteristics; the average physiological statistical characteristics include average heart rate characteristics, average heart rate variability characteristics and average frown electromyogram characteristics; in the feature extraction unit, the following is executed: Extract the heart rate and heart rate variability from the pretreated electrocardiogram signal; average the heart rate and heart rate variability of each subject under each sound stimulus in all test times, and obtain the heart rate feature and heart rate variability feature of the corresponding subject under the corresponding sound stimulus; Extract the corrugator muscle electrical amplitude mean value from the pretreated corrugator muscle electrical signal; average the corrugator muscle electrical amplitude mean value of each subject under each corresponding sound stimulus in all test times, and obtain the corrugator muscle electrical feature of the corresponding subject under the corresponding sound stimulus; Average the heart rate feature, heart rate variability feature and corrugator muscle electrical feature of all subjects under each sound stimulus, and correspondingly obtain the average heart rate feature, average heart rate variability feature and average corrugator muscle electrical feature under the corresponding sound stimulus.

7. The trait screening system of claim 6, wherein, The data processing module further comprises a score processing unit; The score processing unit is configured to average the emotional valence score and arousal score of all subjects under each sound stimulus, and obtain the emotional valence score mean value and arousal score mean value under the corresponding sound stimulus.

8. The trait screening system of claims 1-7, wherein, In the data analysis module, the feature screening result of the pulse sound inducing emotion is determined by using repeated measurement variance analysis, or a combination of multivariate linear regression analysis method and Debin-Watson test method.

9. The trait screening system of claim 8, wherein, When the data analysis module determines the feature screening result of the pulse sound inducing emotion by using repeated measurement variance analysis, the following steps are performed: Perform repeated measurement variance analysis on the relationship between the independent variables and dependent variables of the pulse sound under various sound stimuli; Use spherical test to test whether the repeated measurement variance analysis passes, if the spherical test does not pass, correct the degrees of freedom and statistical values in the repeated measurement variance analysis result to update the repeated measurement variance analysis result; According to the repeated measurement variance analysis result, the correlation between the independent variables and dependent variables of the pulse sound is evaluated, and the independent variable with the greatest impact on emotion in the pulse signal is screened out as the feature screening result of the pulse sound inducing emotion.

10. The trait screening system of claim 8, wherein, When the data analysis module determines the feature screening result of the pulse sound inducing emotion by using a combination of multivariate linear regression analysis method and Debin-Watson test method, the following steps are performed: Use multivariate linear regression analysis method to establish a hierarchical regression model, and gradually add the independent variables of the pulse sound to investigate the relationship between the independent variables and dependent variables of the pulse sound under various sound stimuli; R 2 With the indicators as the reference, the explanatory rate of the independent variables of the pulse sound under different sound stimuli on the dependent variables of the corresponding pulse sound was investigated; combined with the Debin-Watson test method, the independent variable with the greatest impact on emotion in the pulse signal was screened out as the characteristic screening result of the pulse sound for inducing emotion.

Images