Baijiu flavor perception evaluation method based on electroencephalogram and surface myoelectricity

Abstract

The invention belongs to the technical field of baijiu perception analysis, and provides a baijiu flavor perception evaluation method based on electroencephalogram and surface myoelectricity, and the method comprises the following steps: S1, collecting electroencephalogram signals; s2, collecting surface electromyogram signals; s3, physiological electric signal preprocessing and analysis are carried out based on EEGLab and a signal analyzer in MATLAB, an electroencephalogram data analysis method comprises energy spectrum response and power spectrum density analysis of the brain to the Baijiu samples, and a surface myoelectricity analysis method comprises influence and spectrum analysis of different Baijiu samples on muscles. And carrying out Pearson correlation analysis by integrating the electroencephalogram and surface electromyogram signals with an artificial sensory evaluation result. According to the method, the physiological electric signals are fully used for representing the Baijiu sensing process, the Baijiu flavor sensing evaluation method is established from a relatively objective perspective, and the method has high application value and popularization potential.

Description

Technical Field

[0001] The present invention relates to the technical field of liquor perception analysis, and particularly relates to a method for evaluating liquor flavor perception based on electroencephalogram and surface electromyogram. Background Art

[0002] Chinese liquor is an alcoholic beverage made by fermenting, distilling and blending grains. It is divided into twelve fragrance types according to flavor characteristics, including light fragrance, strong fragrance, sauce fragrance, sesame fragrance, etc. Among them, the strong fragrance type liquor, which occupies 70% of the market share, belongs to one of the mainstream Chinese liquors. Flavor perception is one of the key factors determining the quality of liquor. At present, the evaluation methods for the perception of the aroma, taste, texture, etc. of liquor mainly include artificial sensory evaluation and intelligent sensory technology. Compared with the artificial sensory evaluation that consumes manpower and material resources, the intelligent sensory technology has the characteristics of high detection efficiency and objective detection results. Although this technology has been well applied in the differentiation of liquor products, the detection based entirely on sensors still cannot represent the true feelings of the human body towards liquor.

[0003] Physiological electrical signals refer to the potential signals released through the body's own information processing of physiological reactions generated by the human body. Their monitoring and presentation forms usually include electroencephalogram (EEG), electromyogram (EMG), electrocardiogram (ECG), electrodermal activity (EDA), and respiratory signals, etc., which can be used to characterize the true reactions of the human body to different stimuli. Among them, EEG is a method for monitoring physiological electrical signals used to record brain activities, with characteristics such as real-time and non-invasive. The brain generates different electroencephalogram waveforms during the resting state and stimulation. For example, when a person is in a normal waking state, beta waves can be detected in the parietal and frontal lobes, with a frequency range of 13 - 30 Hz. When in a waking and relaxed state, alpha waves can be detected in the occipital lobe, with a frequency range of 8 - 13 Hz. Theta waves can be detected in a light sleep state, with a frequency range of 4 - 8 Hz, and delta waves can be detected when a person is in a deep sleep, with a frequency range of 0.5 - 4 Hz. EEG has been applied to the research of food flavor perception evaluation and the impact on consumers' emotions and behaviors. In addition, EEG is also used to explore the changes in the brain's responses under single taste or olfactory stimuli. For example, some studies have shown that the anterior insula, opercular part, and parietal cortex are basic taste perception regions that can respond to umami. The central area of the brain can respond to sweetness and bitterness, while saltiness has currently been found to have activation in the amygdala. EMG is based on the bioelectric phenomenon in muscle tissue to record the compound action potential of muscle excitation activities, and through appropriate filtering and amplification, the potential changes are presented on an oscilloscope. Surface electromyography is a non-invasive measurement method, which is mainly used in the food field to explore oral behavior differences, distinguish differences between samples, and consumer behavior differences. EEG and EMG can not only truly reflect the human body's feelings and reactions to sample stimuli, but also have strong objectivity, well making up for the deficiencies of artificial sensory evaluation and intelligent sensory technology. However, EEG and EMG have not been applied to the flavor perception evaluation of Baijiu. Therefore, this invention patent collects the physiological electrical signals of the human body during the Baijiu tasting process based on EEG and sEMG and analyzes them to establish a flavor perception evaluation method for Baijiu from a relatively objective perspective. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a method for evaluating the flavor perception of Baijiu based on electroencephalogram and surface electromyogram to solve the problems in the prior art. The technical solution adopted by the present invention is:

[0005] (Modify and supplement the writing)

[0006] The present invention has the following beneficial effects:

[0007] The present invention uses EEG and sEMG to collect the physiological electrical signals of the human body during the evaluation of Chinese liquor. Compared with the artificial sensory evaluation, it can obtain a more objective perception result of the flavor of Chinese liquor, providing new ideas for the establishment of the flavor quality evaluation system of Chinese liquor and scientific data for the development of related products of Chinese liquor. Description of the Drawings

[0008] Figure 1 is the experimental paradigm and the EEG electrode distribution map of the present invention;

[0009] Figure 2 is the response energy spectrum map of the brain of a random subject of the present invention to the Chinese liquor samples. (A) The brain response energy spectrum map after smelling the samples, (B) The brain response energy spectrum map after tasting the samples;

[0010] Figure 3 is the PSD trend of different frequency bands in the same brain region after smelling water and three samples of the present invention. (A) Whole brain, (B) Frontal lobe, (C) Left temporal lobe, (D) Right temporal lobe, (E) Central, (F) Parietal lobe and occipital lobe;

[0011] Figure 4 is the PSD trend between different brain regions in the same frequency band after smelling water and three samples of the present invention. (A) δ frequency band, (B) θ frequency band, (C) α frequency band, (D) β frequency band, (E) γ frequency band;

[0012] Figure 5 is the PSD trend of different frequency bands in the same brain region after tasting water and three samples of the present invention. (A) Whole brain, (B) Frontal lobe, (C) Left temporal lobe, (D) Right temporal lobe, (E) Central, (F) Parietal lobe and occipital lobe;

[0013] Figure 6 is the PSD trend between different brain regions in the same frequency band after tasting water and three samples of the present invention. (A) δ frequency band, (B) θ frequency band, (C) α frequency band, (D) β frequency band, (E) γ frequency band;

[0014] Figure 7 is the activation times (A - C) and activation degrees (D - F) of different muscles after tasting different Chinese liquors. (A, D) Corrugator supercilii muscle (B, E) Zygomatic major muscle (C, F) Mylohyoid muscle;

[0015] Figure 8It is the Pearson correlation analysis result between the PSD value of the EEG signal of the present invention and the scores of 16 artificial sensory evaluation words. (A) The correlation between the PSD values of different frequency bands in the F brain region and the scores of sensory description words in the sniffing experiment. (B) The correlation between the PSD values of different brain regions in the δ frequency band and the scores of sensory description words in the sniffing experiment. (C) The correlation between the PSD values of different frequency bands in the F brain region and the scores of sensory description words in the tasting experiment. (D) The correlation between the PSD values of different brain regions in the δ frequency band and the scores of sensory description words in the tasting experiment. (E) The correlation between the EMG signal and the scores of sensory description words. Detailed implementation manners

[0016] Next, in combination with the Figures 1-8 examples of the present invention, the technical solutions in the examples of the present invention will be clearly and completely described. Obviously, the described examples are only a part of the examples of the present invention, rather than all the examples. If not specifically specified, the technical means used in the examples are conventional means well known to those skilled in the art.

[0017] Step S1: Process of collecting electroencephalogram (EEG) signals: After the EEG signal collection device is connected, the subject can adjust the sitting posture to a comfortable state, press the EEG cap tightly against the subject's scalp according to the international 10 / 20 electrode distribution. After wearing it, apply conductive paste at the electrodes to reduce the impedance. The experiment requires that the impedance of each electrode be kept below 15 kΩ, and the sampling frequency is 1000 Hz. After the collected signal is stable, the formal experiment starts. If abnormal waveforms appear during the experiment, the experiment will be stopped. After troubleshooting, the experiment will continue. The EEG experiment is divided into two parts: sniffing and tasting, which are carried out separately. During the experiment, the subject closes his eyes throughout the process, and is informed of the entire experimental process, and tries to keep still during the collection process. Three samples and water are randomly delivered by a peristaltic pump and enter the subject's mouth through a food-grade rubber tube. The pumping volume each time is 0.75 mL, and each sample is repeated 20 times. During each repetition, the subject needs to gargle and have sufficient rest. During the sniffing experiment, the samples are randomly delivered to the front of the subject's nose by a sniffing diffusion device. The distance between the sniffing port of the device and the front of the subject's nose is 5 cm, and the ventilation system is turned on during the experiment to ensure the air circulation in the room.

[0018] Step S2. Process of surface electromyogram (sEMG) signal acquisition: The muscle parts for sEMG acquisition include the corrugator supercilii (right side), the zygomaticus major (right side), and the mylohyoid (right side), and the grounding electrode is at the tip of the nose. The face is cleaned before acquisition. During the signal acquisition process, the electrode patches are first attached to the corresponding muscle surfaces, and then the BioNomadix sensors are connected to the electrode patches. The sampling frequency is 2000 Hz. Each sensor module can simultaneously acquire signals from two channels and is matched with each corresponding MP160 system module. Subsequently, the MP160 system is connected to the computer through the network cable port, and the corresponding number of muscle channels is designed. The software AcqKnowledge is used to display and record the signals. The sEMG signals change periodically during the monitoring process and are completely recorded by the software system. The signals can clearly reflect the contraction changes of the muscles. During the experiment, the subjects independently perform sniffing and tasting according to the video prompts to restore real drinking actions, and gargle and take sufficient rest between each repeated experiment. The experiments with samples and water are both repeated 7 times. Each time of drinking uses a 50 mL tasting cup, and the intake is 2 mL.

[0019] Step S3. Data processing: Matlab and EEGLab are used to analyze the physiological electrical signal data. First, the data is preprocessed. Then, the methods for electroencephalogram data analysis include the energy spectrum response of the brain to the liquor samples and the power spectral density (PSD) analysis. The methods for surface electromyogram analysis include the effects of different liquor samples on the muscles and the spectrum analysis, and the Pearson correlation analysis is performed on the comprehensive electroencephalogram and surface electromyogram signals and the results of the artificial sensory evaluation.

[0020] Preferably, in Steps S1 and S2, the liquor samples used are YMR (35.8% vol), YHDQLC (42% vol), and ZGMJ (52% vol), all collected from Jiangsu Province, China.

[0021] Preferably, in Steps S1 and S2, the subjects are members of the quantitative descriptive analysis (QDA) evaluation group in the previous study, a total of 12 (6 males, 6 females, with an average age of 25 years). All the subjects are in good health, without psychological or physiological diseases, and are all right-handed. All the subjects voluntarily participate in the experiment. They do not take foods and drugs that can affect cerebral blood circulation such as coffee 24 hours before the experiment, ensure that the subjects understand the basic content and process of the experiment, and sign the informed consent form.

[0022] Preferably, the device used in Step S1 is an Eego 64-channel electroencephalogram system, which includes an electrode cap, an amplifier, a sensor, lead wires, etc. The device used in Step S2 is an MP160 physiological polygraph. This system includes electrodes, lead wires, and amplifiers, and is designed as a modular structure. Each module contains signal acquisition electrodes and signal processing amplifiers.

[0023] Preferably, in step S1, the brain region channels concerned by EEG are 64 channels, and the brain region positions include FP1, FP2, F7, F3, Fz, F4, F8, FT7, FC5, FC3, FC1, FCz, FC2, FC4, FC6, FT8, T7, C5, C3, C1, Cz, C2, C4, C6, T8, TP7, CP5, CP3, CP1, CPz, CP2, CP4, CP6, TP8, P7, P5, P3, P1, Pz, P2, P4, P6, P8, PO7, PO3, Poz, PO4, PO8, O1, Oz, O2, M1. The above represent the positions where the electrode patches are attached to the scalp of the brain. For details, see Figure 1 . The partitioning method commonly used in research. In step S2, the surface muscles concerned by sEMG include the corrugator supercilii (right), the zygomaticus major (right), and the mylohyoid (right).

[0024] Preferably, in step S3, the EEG signal preprocessing includes electrode positioning, rereferencing, filtering, segmentation, and baseline correction, rejecting bad segments and interpolating bad leads, and independent component analysis (ICA); the sEMG signal preprocessing includes band-pass filtering from 20 - 450 Hz to eliminate noise and artifacts, and notch filtering at 50 Hz and 100 Hz to remove power frequency interference.

[0025] The energy spectrum response results of the brain to the baijiu sample: Use the spectrum analysis function built in EEGLab to analyze the preprocessed data, and randomly select the spectrogram and the corresponding brain topographic map of a subject for presentation, as Figure 2 shown. The abscissa in the figure represents frequency, and the ordinate represents the power spectral density (PSD). The overall change trend of the spectrogram of the subject shown as an example is consistent with that of other subjects. For the olfactory stimulation of baijiu received, the spectrogram is as Figure 2 (A) shown, showing a first decline, then a rise to reach the first PSD peak, and the frequency range where the peak appears is 9 - 10 Hz. Then the trend gradually flattens, and a second peak appears at about 22 Hz. This peak is smaller than the first peak. After the change flattens, it drops again at about 44 Hz in frequency, and finally the PSD value is basically parallel to the abscissa. For the brain spectrum change after receiving the gustatory stimulation of baijiu, it is as Figure 2 (B) shown, and the overall trend is not very different from the change trend caused by olfaction.

[0026] EEG power spectral density analysis results: Further calculate the average value of the PSD of all subjects, and statistically analyze the PSD trend differences in different frequency bands in the same brain region. Among them, the results of the sniffing experiment are as Figure 3As shown. It can be seen from the figure that the PSD values after smelling the Chinese liquor samples are lower than those after smelling water in the δ and α bands of each brain region. There is not much difference in the responses of the brain to olfactory stimuli between water and samples in other frequency bands.

[0027] Further statistical analysis of the PSD differences between different brain regions in the same frequency band during the smelling experiment shows the results as Figure 4 shown. Generally speaking, when comparing the smelling of water and Chinese liquor, the brain regions with significant differences in brain activities include the frontal lobe, central region, parietal lobe, and occipital lobe. The results of this study show that after smelling the Chinese liquor samples, the activities of the frontal lobe of the brain decrease within 0.5 - 13 Hz, and the activities of the occipital lobe and temporal lobe of the brain increase within 13 - 50 Hz. It can be seen that compared with water, the smell of Chinese liquor makes the subjects think it is an unpleasant smell, which may be an instinctive reaction after the alcohol in Chinese liquor stimulates the human body. Combining the analysis of the previous artificial sensory evaluation results, although the subjects subjectively think that Chinese liquor has various aromas, from a hidden perspective, the human brain will still instinctively defend or inhibit strong stimuli. ANOVA is used to analyze the response changes of the brain after smelling three Chinese liquor samples. The results show that the differences between different brain regions and different frequency bands are not significant. The reason for the analysis is that the smell of ethanol still dominates in the three samples, and for similar smells, the brain will produce similar electroencephalogram signal changes, especially in the orbitofrontal cortex in the θ and γ frequency bands.

[0028] After tasting water and the three samples, the PSD trends in different frequency bands in the same brain region are as Figure 5 shown. Figure 5(A) shows that across the whole brain, taste stimulation leads to stronger brain responses in the δ, α, and β bands. The brain activity intensity after tasting Chinese liquor is greater than that after tasting water, which is also very obvious in the γ band. Except that the PSD value in the α band is the highest in the occipital lobe, the PSD value is the highest in the δ band in other brain regions. The trends of differences among the three samples are basically the same in the frequency bands corresponding to each brain region. In the δ band of each brain region, the PSD value after tasting YMR (38.5% vol) is the highest, followed by ZGMJ (52% vol), and the smallest is YHDQLC (42% vol). In the α band of each brain region, the corresponding PSD values from small to large are for tasting YMR, YHDQLC, and ZGMJ in sequence. In the β band of each brain region, the corresponding PSD values from small to large are for tasting ZGMJ, YHDQLC, and YMR in sequence. There are no obvious differences in the PSD values after tasting the three samples in other bands. The PSD value does not show a single increase or decrease with the increase in alcohol content, indicating that among the three samples in this study, the alcohol content is not the main factor leading to the differences in brain response changes. Further ANOVA analysis shows that there are extremely significant differences in the brain activity changes after tasting water and Chinese liquor in the δ band of all brain regions (p < 0.001). Among them, there are also significant differences in the brain frontal lobe activities caused by water and the samples in the β band (p < 0.05), while there are no significant differences in the brain activity differences caused by tasting the three Chinese liquors.

[0029] The PSD trends of different brain regions in the same band after tasting water and the three samples are as Figure 6 shown. In the δ and θ bands, the activity of the brain frontal lobe is the strongest, but as the frequency increases, the activity intensity of the brain frontal lobe gradually decreases. At the same time, the activities of the brain occipital lobe and the right temporal lobe gradually increase. Through ANOVA analysis, it is found that there are extremely significant differences in the brain activity intensities of the frontal lobe, left temporal lobe, right temporal lobe, and central region in the δ band after tasting water and Chinese liquor (p < 0.001), but there are no significant differences in the brain activity intensities among different brain regions after tasting the three samples.

[0030] The effects of different Chinese liquor samples on muscles: First, calculate the activation times of the muscles in three parts, mainly based on the activation events in the EMG signals, that is, the signals with amplitudes exceeding the preset threshold. Here, the threshold is set to 3 times the standard deviation. Whenever the amplitude of the signal exceeds this threshold, this point is considered an activation event of the muscle. Count the activity times of each muscle after tasting the three Chinese liquor samples respectively, as Figure 7As shown in (A-C). It can be seen from the figure that the activation times of the corrugator supercilii muscle of the subjects from less to more are YMR, YHDQLC, ZGMJ. The activation times of the zygomaticus major muscle from more to less are YHDQLC, ZGMJ, YMR. The activation times of the mylohyoid muscle from more to less are ZGMJ, YMR, YHDQLC. After independent-sample t-tests, it was found that there were significant differences in the activation times of the corrugator supercilii muscle after tasting YMR and ZGMJ (p < 0.05). After tasting YMR and YHDQLC, and YMR and ZGMJ respectively, there were significant differences in the activation times of the zygomaticus major muscle (p < 0.05). However, there were no significant differences in the mylohyoid muscle after tasting different samples.

[0031] Furthermore, the activation degree of different muscles was obtained by quantifying the amplitude of the EMG signal. The methods for quantifying the EMG signal include calculating the root mean square value (RMS), average absolute value, integral, etc. Among them, RMS is relatively more stable. Therefore, in this study, calculating RMS was used to characterize the activation degree of the muscle. The results are as Figure 7 shown in (D-F). There were no significant differences in the activation degrees of the muscles at the three sites. However, it can also be seen that the activation degrees of the corrugator supercilii muscle and the zygomaticus major muscle were the greatest after tasting ZGMJ, and the activation degree of the mylohyoid muscle was the greatest after tasting YMR. Most studies have shown that facial muscles can be used to characterize the emotional changes of people. For example, the corrugator supercilii muscle is usually proportional to negative emotions, the zygomaticus major muscle is proportional to positive emotions, and the mylohyoid muscle is a muscle involved in the swallowing process. Generally, a larger number of activations of this muscle indicates a larger number of swallowing times. Generally speaking, among the three kinds of white spirits, the subjects had stronger positive emotions after tasting YHDQLC, and the artificial sensory evaluation results showed that the highest comprehensive score of the three samples was also YHDQLC. In addition, from the activation times and activation degrees of the mylohyoid muscle, ZGMJ caused more swallowing times. The alcohol content of this sample was the highest, which promoted the swallowing action of the subjects to a certain extent.

[0032] sEMG spectrum analysis results: After performing spectrum analysis on the EMG signals, it was found that the activity frequencies of the three muscles were all lower than 500 Hz, and the power decreased significantly after exceeding 500 Hz, indicating that the changes in these three muscles during the process of tasting Chinese liquor mainly occurred in the low-frequency range. On the one hand, the reason is that the activities of facial muscles are mostly concentrated in the low-frequency range. Expressions such as frowning and smiling require delicate and continuous muscle control, and most facial muscles are slow-twitch muscle fibers, which are mainly used for maintaining low-intensity muscle activities for a long time. On the other hand, because during the actual experiment, the facial expressions of the subjects after tasting Chinese liquor were not very obvious, it also shows that EMG can capture some more subtle facial changes that are easily overlooked by the naked eye. Further calculating the power spectral density corresponding to the three muscles and performing a significant difference analysis, the results showed that there were significant differences in the changes of the three muscles after tasting YMR and YHDQLC (p < 0.05). After tasting YMR and ZGMJ, there were significant differences in the changes of the zygomatic major muscle and the mylohyoid muscle (p < 0.05), mainly manifested as lower PSD values of the zygomatic major muscle and the mylohyoid muscle after tasting YMR.

[0033] Correlation analysis of EEG and EMG signals with the results of artificial sensory evaluation: To illustrate the correlation between physiological electrical signals and the results of artificial sensory evaluation, in this study, Pearson correlation analysis was performed between the PSD values of EEG and EMG signals and the results of artificial sensory evaluation respectively. The previous results of artificial sensory evaluation used quantitative descriptive analysis (QDA). A total of 16 flavor descriptive words were formed for Chinese liquor samples including the samples in this experiment. These 12 subjects still used the 9-point method to evaluate different Chinese liquor samples. Based on the above analysis, the PSD values of different frequency bands in the frontal region and the PSD values of different brain regions in the δ band in the EEG data obtained from the sniffing experiment and the tasting experiment were respectively correlated with the artificial sensory scores, and the results are as Figure 8 (A-D) shown.

[0034] Figure 8 (A) and (B) are the results of the sniffing experiment. It can be seen that in the frontal lobe region, mainly the δ band and the α band have a strong correlation with the aroma and taste vocabulary in the sensory descriptive words, while in the δ band, mainly the right temporal lobe region of the brain has a strong correlation with the sensory descriptive words. The results of the tasting experiment are as Figure 8(C) and (D) show that in the frontal lobe region of the brain, the response in the δ band is strongly correlated with artificial senses. Within the δ band, the frontal lobe and the left temporal lobe are mainly strongly correlated with sensory descriptors. Currently, it is generally believed that the right brain is stronger than the left brain in the brain's perception of taste, the left brain is more adept at memory and emotion control, and the frontal lobe is good at thinking and judgment. The results of this study show that the frontal lobe and the temporal lobe are mainly mobilized during the whole process of tasting baijiu, indicating that the subjects will also involuntarily tend to experience and memory during perception, and emotions will also affect the final perception and evaluation of the samples. The same correlation analysis was also performed on the PSD values of the EMG signals, and the results are as Figure 8 (E) shows. It can be seen that the PSD value of the mylohyoid muscle has a significant correlation with sensory descriptors, followed by the corrugator supercilii muscle (p < 0.05). The activity of the corrugator supercilii muscle is generally related to negative emotions, but in this study, the corrugator supercilii muscle is negatively correlated with cellar aroma, grain aroma and bitterness. Therefore, the sensory evaluation made by the subjects on baijiu is not the result of a single emotion. In addition, the mylohyoid muscle is negatively correlated with softness. The fewer the number of activities of the mylohyoid muscle indicates the fewer the number of swallows, which to a certain extent can also indicate that the sample has a weaker irritation. Therefore, it can also be used as one of the characterization methods for the softness degree of baijiu.

[0035] The embodiments described above are only descriptions of the preferred modes of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations, variations, modifications, and substitutions made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for evaluating the flavor perception of Chinese liquor based on electroencephalogram and surface electromyogram, characterized in that, It includes the following steps: Step S1, the process of electroencephalogram (EEG) signal acquisition: After the EEG signal acquisition device is connected, the subject wears an EEG cap, and then olfactory and tasting experiments are carried out; Step S2, the process of surface electromyogram (sEMG) signal acquisition: The muscle parts for sEMG acquisition include the corrugator supercilii muscle, the zygomaticus major muscle, and the mylohyoid muscle. The ground electrode is at the tip of the nose. The electrode patches are attached to the corresponding muscle surfaces, and then the BioNomadix sensor is connected to the electrode patches. The sampling frequency is 1500 - 2000 Hz. The software AcqKnowledge is used to display and record the signals. During the experiment, the subject autonomously performs olfactory and tasting according to the video prompts to restore real drinking actions; Step S3, data processing: Matlab and EEGLab are used to analyze the physiological electrical signal data. First, the data is pre - processed. Then, the EEG data analysis methods include the energy spectrum response of the brain to baijiu samples and power spectral density analysis. The sEMG analysis methods include the effects of different baijiu samples on muscles and spectral analysis, and Pearson correlation analysis is performed on the comprehensive EEG and sEMG signals and the results of artificial sensory evaluation.

2. The method for evaluating the flavor perception of white liquor based on electroencephalogram and surface electromyogram according to claim 1, wherein In step S1, the EEG cap is closely attached to the subject's scalp. After wearing it, conductive paste is injected at the electrodes to reduce impedance. The impedance of each electrode is kept below 15 kΩ. The sampling frequency is 1000 Hz. If abnormal waveforms appear during the experiment, the experiment will be stopped. After troubleshooting, the experiment will continue.

3. The method for evaluating the flavor perception of white liquor based on electroencephalogram and surface electromyogram according to claim 1, characterized in that, In step S1, during the tasting experiment, three samples and water are randomly delivered by a peristaltic pump and enter the subject's mouth through a food - grade rubber tube. The amount pumped in each time is 0.5 - 1.0 mL, and each sample is repeated 15 - 20 times.

4. The method for evaluating the flavor perception of liquor based on electroencephalogram and surface electromyogram according to claim 1, characterized in that, In step S1, during the olfactory experiment, the samples are randomly delivered to the front of the subject's nose by an olfactory diffusion device. The distance between the olfactory port of the device and the front of the subject's nose is 4 - 6 cm.

5. The method for evaluating the flavor perception of Chinese liquor based on electroencephalogram and surface electromyogram according to claim 1, wherein, In step S1, the EEG signal acquisition device is a 64 - channel EEG system; In step S2, a physiological polygraph is used for sEMG signal acquisition.

6. The method for evaluating the flavor perception of Baijiu based on electroencephalogram and surface electromyogram according to claim 1, characterized in that, In step S1, the brain regions include FP1, FP2, F7, F3, Fz, F4, F8, FT7, FC5, FC3, FC1, FCz, FC2, FC4, FC6, FT8, T7, C5, C3, C1, Cz, C2, C4, C6, T8, TP7, CP5, CP3, CP1, CPz, CP2, CP4, CP6, TP8, P7, P5, P3, P1, Pz, P2, P4, P6, P8, PO7, PO3, Poz, PO4, PO8, O1, Oz, O2, M1.

7. The method for evaluating the flavor perception of liquor based on electroencephalogram and surface electromyogram according to claim 1, characterized in that, In step S2, during the sEMG signal acquisition process, the experiments of samples and water are both repeated 7 times. Each time of drinking uses a 50 - mL tasting cup, and the intake amount is 2 mL.

8. The method for evaluating the flavor perception of white liquor based on electroencephalogram and surface electromyogram according to claim 1, wherein In step S3, the data pre - processing includes EEG signal pre - processing and sEMG signal pre - processing, where: EEG signal pre - processing includes electrode localization, rereferencing, filtering, segmentation, baseline correction, rejection of bad segments and interpolation of bad channels, and independent component analysis; The surface electromyogram signal preprocessing includes band-pass filtering from 20 to 450 Hz to eliminate noise and artifacts, and notch filtering at 50 Hz and 100 Hz to remove power frequency interference.

Images