Search

CN-122004794-A - Non-invasive user sleep quality monitoring method and system

CN122004794ACN 122004794 ACN122004794 ACN 122004794ACN-122004794-A

Abstract

The application relates to the technical field of sleep quality monitoring, and provides a non-invasive user sleep quality monitoring method and system, which are characterized in that a feature analysis module, a parameter optimization module, a phase synchronization module and a double-loop feedback module are organically integrated to construct a comprehensive and intelligent sleep quality monitoring system, the modules cooperate with each other and share data, the user sleep quality is analyzed, monitored and interfered from multiple dimensions, the feature analysis module provides abundant and accurate feature data for the follow-up, the parameter optimization module determines the stimulation parameters based on the feature data and the physiological features of the user, the phase synchronization module optimizes the stimulation signals according to the stimulation parameters and the brain phase difference, and the double-loop feedback module monitors the stimulation effect in real time and dynamically adjusts the stimulation effect, so that the system adaptively adjusts the intervention mode according to the real-time sleep state and the individual difference of the user, and the accuracy and the effectiveness of sleep quality monitoring and improvement are improved.

Inventors

  • WANG XIAOFENG
  • Ma Junang

Assignees

  • 脉景（杭州）健康管理有限公司
  • 苏州市脑悦科技有限公司

Dates

Publication Date
20260512
Application Date
20260415

Claims (10)

  1. 1. The non-invasive user sleep quality monitoring system is characterized by comprising a characteristic analysis module, a parameter optimization module, a phase synchronization module and a double-loop feedback module; the feature analysis module is used for acquiring the EEG signals, extracting the space-time features of the EEG signals, and generating a feature vector group based on the space-time features of the EEG signals connected by the graph neural network; The parameter optimization module is used for calculating the similarity between the feature vector set and the transcranial alternating current stimulation parameters, generating a stimulation parameter set, and adjusting an individual stimulation threshold value and the stimulation parameter set according to the metabolic rate and the heart rate variability of the user; the phase synchronization module is used for determining a stimulation signal waveform according to the stimulation parameter set, receiving phase coherence across a brain region to interfere the stimulation signal waveform, determining the relation between a sleep stage and a stimulation mode based on the feature vector set and an individual stimulation threshold value, and performing multi-round transcranial alternating current stimulation on a user according to the stimulation mode; After each round of transcranial alternating current stimulation, extracting the variation of the feature vector group, and judging whether to reset the generated stimulation parameter group based on the variation of the feature vector group and the variation of the metabolic rate of the user; the double-loop feedback module is used for judging whether stimulus tolerance occurs according to the variation of the feature vector group, the variation amplitude of the metabolic rate of the user and the stimulus effect so as to adjust the stimulus mode and the individual stimulus threshold.
  2. 2. A non-invasive user sleep quality monitoring system according to claim 1, wherein said feature vector set generation logic comprises: Acquiring EEG signals and preprocessing the EEG signals; Extracting spatiotemporal features of EEG signals, the spatiotemporal features of EEG signals comprising Spatial-temporal distribution of wave power, Wave power ratio and phase coherence across brain regions; Constructing a graph structure, determining nodes and edges of the graph structure, inputting the graph structure into a graph neural network, updating the nodes through convolution operation, outputting feature vectors of each node, and combining the feature vectors of all the nodes to generate a feature vector group.
  3. 3. A non-invasive user sleep quality monitoring system according to claim 2, wherein the logic for generating the set of stimulation parameters comprises: Encoding the transcranial alternating current stimulation parameters and converting the encoded transcranial alternating current stimulation parameters into stimulation feature vectors; Calculating the distance between the feature vector group and the stimulation feature vector by the Markov distance; normalizing the distance between the feature vector group and the stimulation feature vector, converting the distance into the similarity between the feature vector group and the stimulation feature vector, configuring a similarity threshold, and if the similarity between the feature vector group and the stimulation feature vector is larger than the similarity threshold, incorporating the transcranial alternating current stimulation parameters into the stimulation parameter group; and if the similarity between the feature vector group and the stimulation feature vector is not greater than the similarity threshold, screening and removing the transcranial alternating current stimulation parameters, wherein the transcranial alternating current stimulation parameters comprise stimulation frequency, stimulation intensity and stimulation duration.
  4. 4. A non-invasive user sleep quality monitoring system according to claim 3, wherein said adjustment logic for said set of stimulation parameters comprises: acquiring the metabolic rate and the heart rate variability of the user, and calculating a dynamic safety factor in a correlation manner by the metabolic rate and the heart rate variability of the user so as to determine the safety range of the stimulus intensity; Adjusting individual stimulation thresholds based on the user metabolic rate, heart rate variability, and stimulation intensity; iteratively adjusting a set of stimulation parameters targeting a stimulation effect, wherein the stimulation effect comprises Rate of change of wave power and number of user wakeups.
  5. 5. A non-invasive user sleep quality monitoring system according to claim 4, wherein the intervention logic of the stimulus signal waveform comprises: Receiving phase coherence of a cross brain region, and analyzing phase differences of each brain region in real time; Adjusting the compensation speed based on the metabolic rate of the user, and compensating the phase of the stimulation signal according to the compensation speed and the phase difference of each brain region; The stimulation signal phase is adjusted in conjunction with the individual stimulation threshold.
  6. 6. A non-invasive user sleep quality monitoring system according to claim 5, wherein the sleep stage versus stimulus pattern relationship determination logic comprises: fusing the feature vector group with an individual stimulation threshold to generate sleep stage features, and performing cluster analysis on the sleep stage features to obtain a clustering result of the sleep stage; Determining a stimulation pattern from the stimulation parameter set and the stimulation signal waveform combination; Simulating a stimulation mode in different sleep stages, and monitoring the stimulation effect to evaluate the effect of the stimulation mode; Based on the clustering result of the sleep stages and the effect of the stimulation pattern, the relation between the sleep stages and the stimulation pattern is determined.
  7. 7. The non-invasive user sleep quality monitoring system according to claim 6, wherein the decision logic for resetting the set of generated stimulation parameters comprises: after each round of transcranial alternating current stimulation, the Euclidean distance of the feature vector group before and after stimulation is calculated to extract the variation of the feature vector group; Synchronously monitoring the change amplitude of the user metabolism rate, configuring a change amount threshold and an amplitude threshold, and triggering reset to generate a stimulation parameter set if the change amount of the feature vector set is larger than the change amount threshold and the change amplitude of the user metabolism rate is larger than the amplitude threshold; A reset signal is generated and transmitted to the parameter optimization module to reset the set of generated stimulation parameters.
  8. 8. A non-invasive user sleep quality monitoring system according to claim 7, wherein said stimulus tolerance determination logic comprises: The variation of the feature vector group and the variation amplitude of the user metabolism rate are weighted to obtain the stimulus tolerance; Will stimulate the effect in The change rate of wave power and the awakening times of the user are calculated through weighting to obtain the score of the stimulation effect; and (3) configuring a tolerance threshold, comparing the stimulus tolerance degree with the tolerance threshold, and comprehensively judging the stimulus tolerance by combining the scores of the stimulus effects.
  9. 9. A non-invasive user sleep quality monitoring system according to claim 2, wherein nodes in the graph structure represent spatiotemporal features of EEG signals of different brain regions, and edges in the graph structure represent functional relationships between brain regions.
  10. 10. A non-invasive user sleep quality monitoring method implemented based on the non-invasive user sleep quality monitoring system of any of claims 1-9, comprising: acquiring an EEG signal, extracting space-time characteristics of the EEG signal, and generating a characteristic vector group by connecting the space-time characteristics of the EEG signal based on a graph neural network; calculating the similarity between the feature vector set and the transcranial alternating current stimulation parameters to generate a stimulation parameter set; adjusting individual stimulation thresholds according to the metabolic rate and heart rate variability of the user, and adjusting a stimulation parameter set based on the individual stimulation thresholds; determining a stimulation signal waveform according to the set of stimulation parameters, and receiving phase coherence across the brain region to interfere with the stimulation signal waveform; Determining the relation between the sleep stage and the stimulation mode based on the feature vector group and the individual stimulation threshold value, and performing multi-round transcranial alternating current stimulation on the user according to the stimulation mode; After each round of transcranial alternating current stimulation, extracting the variation of the feature vector group, and judging whether to reset the generated stimulation parameter group based on the variation of the feature vector group and the variation of the metabolic rate of the user; And comprehensively judging whether stimulus tolerance occurs according to the variation of the feature vector group, the variation amplitude of the metabolism rate of the user and the stimulus effect, and if the stimulus tolerance occurs, adjusting the stimulus mode and the individual stimulus threshold.

Description

Non-invasive user sleep quality monitoring method and system Technical Field The application relates to the technical field of sleep quality monitoring, in particular to a non-invasive user sleep quality monitoring method and system. Background At present, sleep quality monitoring technology is widely researched and applied, but in the aspect of signal analysis, most systems only simply analyze basic characteristics of EEG signals, but fail to deeply mine abundant space-time characteristics in the EEG signals, neglect in the prior art leads to insufficient and accurate understanding of sleep states of users, meanwhile, in the aspect of determining stimulation parameters, the traditional method lacks sufficient consideration of individual physiological characteristics of the users, transcranial alternating current stimulation is used as a means for improving sleep quality, the setting of the parameters adopts fixed or universal standards, and the individual physiological indexes are not combined for dynamic adjustment, so that the stimulation effect varies from person to person, individual requirements of different users are difficult to meet, and discomfort or adverse effects are caused to the users even due to improper stimulation parameters. The prior art can not effectively intervene in the waveform of the stimulation signal according to the real-time change of the brain nerve activity, the nerve activity of different brain areas has complex phase relation, but the traditional technology can not utilize the phase information to carry out phase compensation and optimization on the stimulation signal, so that the synchronism of the stimulation signal and the brain nerve activity is poor, and the exertion of the stimulation effect is influenced; moreover, the existing sleep quality monitoring system lacks an effective feedback mechanism, and when a user tolerates transcranial alternating current stimulation, the user cannot judge timely and accurately and take corresponding measures, so that the stimulation effect is gradually reduced after long-term use, and the sleep quality of the user cannot be continuously improved. Disclosure of Invention Aiming at the defects of the prior art, the application provides a non-invasive user sleep quality monitoring method and system. In a first aspect, the application provides a non-invasive user sleep quality monitoring system, which comprises a characteristic analysis module, a parameter optimization module, a phase synchronization module and a double-loop feedback module; the feature analysis module is used for acquiring the EEG signals, extracting the space-time features of the EEG signals, and generating a feature vector group based on the space-time features of the EEG signals connected by the graph neural network; The parameter optimization module is used for calculating the similarity between the feature vector set and the transcranial alternating current stimulation parameters, generating a stimulation parameter set, and adjusting an individual stimulation threshold value and the stimulation parameter set according to the metabolic rate and the heart rate variability of the user; the phase synchronization module is used for determining a stimulation signal waveform according to the stimulation parameter set, receiving phase coherence across a brain region to interfere the stimulation signal waveform, determining the relation between a sleep stage and a stimulation mode based on the feature vector set and an individual stimulation threshold value, and performing multi-round transcranial alternating current stimulation on a user according to the stimulation mode; After each round of transcranial alternating current stimulation, extracting the variation of the feature vector group, and judging whether to reset the generated stimulation parameter group based on the variation of the feature vector group and the variation of the metabolic rate of the user; the double-loop feedback module is used for judging whether stimulus tolerance occurs according to the variation of the feature vector group, the variation amplitude of the metabolic rate of the user and the stimulus effect so as to adjust the stimulus mode and the individual stimulus threshold. As an alternative embodiment, the generating logic of the feature vector group includes: Acquiring EEG signals and preprocessing the EEG signals; Extracting spatiotemporal features of EEG signals, the spatiotemporal features of EEG signals comprising Spatial-temporal distribution of wave power,Wave power ratio and phase coherence across brain regions; Constructing a graph structure, determining nodes and edges of the graph structure, inputting the graph structure into a graph neural network, updating the nodes through convolution operation, outputting feature vectors of each node, and combining the feature vectors of all the nodes to generate a feature vector group. As an alternative embodiment, the logic for generating the