CN-121987223-A - Closed-loop analysis system and method for sound stimulation-electroencephalogram response

CN121987223ACN 121987223 ACN121987223 ACN 121987223ACN-121987223-A

Abstract

The invention provides a closed-loop analysis system and a method for sound stimulation-electroencephalogram response, which relate to the technical field of electroencephalogram data analysis and comprise the steps of acquiring real-time electroencephalogram data under sound stimulation based on a standardized interface, and performing artifact removal and data standardization operation to obtain standard electroencephalogram data; the method comprises the steps of obtaining standard electroencephalogram data, carrying out response latency analysis on the standard electroencephalogram data to obtain response latency data, carrying out response significance analysis on the standard electroencephalogram data to obtain response significance data, extracting an abnormal response event set from the standard electroencephalogram data, carrying out context feature modeling and abnormal root cause analysis on the abnormal response event set based on the response latency data and the response significance data to obtain abnormal root cause data, generating an electroencephalogram response analysis graph interface corresponding to the response latency data, the response significance data and the abnormal root cause data based on nerve signal analysis logic, and carrying out visual display on the electroencephalogram response analysis graph interface.

Inventors

CHEN HAO
Lv Linyang
ZHU DEJIANG
FAN LIWEN

Assignees

杭州皓世天辉科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260410

Claims (10)

1. A method of closed loop analysis of acoustic stimulus-electroencephalogram response, the method comprising: acquiring real-time brain electrical data under sound stimulation based on a standardized interface of an brain electrical acquisition system, and performing artifact removal and data standardization operation on the real-time brain electrical data based on a stimulation event identifier to obtain standard brain electrical data; Performing response latency analysis on the standard electroencephalogram data based on the triggering time of sound stimulation and the peak time of electroencephalogram response to obtain response latency data; Performing response significance analysis on the standard electroencephalogram data based on a dynamic response intensity threshold value to obtain response significance data; Extracting an abnormal response event set from the standard electroencephalogram data, and carrying out context feature modeling and abnormal root cause analysis on the abnormal response event set based on the response latency data and the response significance data to obtain abnormal root cause data; And generating an electroencephalogram response analysis chart interface corresponding to the response latency data, the response significance data and the abnormal root data based on the nerve signal analysis logic, and visually displaying the electroencephalogram response analysis chart interface.
2. The closed loop analysis method of sound stimulation-electroencephalogram response according to claim 1, wherein the process of acquiring real-time electroencephalogram data under sound stimulation based on the standardized interface of the electroencephalogram acquisition system comprises: carrying out multi-channel signal acquisition on electroencephalogram acquisition equipment in a sound stimulation scene to obtain an original electroencephalogram signal, carrying out stimulation event association and data encapsulation on the original electroencephalogram signal based on stimulation presentation logic to obtain encapsulated electroencephalogram data, and publishing the encapsulated electroencephalogram data to a preset nerve signal data platform based on a standardized interface of an electroencephalogram acquisition system; Receiving the real-time brain electrical information of the nerve signal data platform in a signal subscription mode, carrying out signal analysis on the real-time brain electrical information to obtain analysis brain electrical data, and carrying out stimulus response context association on the analysis brain electrical data to obtain real-time brain electrical data.
3. The method of closed loop analysis of acoustic stimulus-electroencephalogram response of claim 2, wherein the process of artifact removal and data normalization of the real-time electroencephalogram data based on stimulus event identification comprises: The method comprises the steps of classifying the real-time electroencephalogram data according to the stimulus type, adding a stimulus tag, carrying out signal integrity check on the real-time electroencephalogram data based on the stimulus tag, and carrying out data deduplication on the real-time electroencephalogram data after the integrity check based on a stimulus event identifier and a signal acquisition time difference threshold value to obtain deduplication electroencephalogram data; the method comprises the steps of carrying out missing signal completion on the de-duplicated electroencephalogram data by utilizing time sequence linear interpolation to obtain full electroencephalogram data, carrying out electro-oculogram artifact and myoelectric artifact detection on the full electroencephalogram data based on a sliding time window to obtain artifact window data, separating corresponding artifact components from effective electroencephalogram components based on an independent component analysis algorithm, retaining the effective electroencephalogram components and reconstructing an electroencephalogram signal to obtain de-artifact electroencephalogram data, carrying out normalization processing on the de-artifact electroencephalogram data based on a preset signal amplitude range, and carrying out data structure normalization operation on the normalized de-artifact electroencephalogram data based on a preset electroencephalogram field template to obtain standard electroencephalogram data.
4. A method of closed loop analysis of acoustic stimulus-brain electrical response according to claim 3, wherein the process of performing response latency analysis on the standard brain electrical data based on the trigger time of acoustic stimulus and the peak time of brain electrical response comprises: The stimulus event identification is used as a primary key, a corresponding electroencephalogram response event group is extracted from the standard electroencephalogram data, and then a stimulus trigger event and a response peak event are extracted to obtain a trigger-peak event group set; The method comprises the steps of carrying out stimulation type grouping on a latency sequence set according to a stimulation label to obtain a latency group set, carrying out multidimensional statistical analysis on each latency group in the latency group set to obtain a latency statistical data set, carrying out latency anomaly analysis on the latency group set based on the latency statistical data set to obtain an abnormal latency sequence table, and taking the abnormal latency sequence table and the latency statistical data set as response latency data.
5. The method of closed loop analysis of acoustic stimulus-brain electrical response of claim 4, wherein the process of performing a response saliency analysis on the standard brain electrical data based on a dynamic response intensity threshold comprises: Sequencing the amplitude sequence group sets according to the stimulus presentation sequence to obtain an amplitude event sequence set, and calculating a peak intensity sequence set corresponding to the amplitude event sequence set; The method comprises the steps of obtaining a peak intensity sequence set, obtaining a historical intensity sequence set corresponding to the peak intensity sequence set, and generating corresponding dynamic response intensity thresholds for all intensity groups based on the historical intensity sequence set; And performing salient response duty ratio analysis on each intensity group in the intensity group set based on the salient response group set to obtain a response duty ratio data set, and taking the salient response event list and the response duty ratio data set as response salient data.
6. The method of claim 5, wherein extracting an abnormal response event set from the standard electroencephalogram data, and performing context feature modeling and abnormal root cause analysis on the abnormal response event set based on the response latency data and the response significance data comprises: The standard electroencephalogram data is subjected to graph structure modeling and abnormal context modeling based on the electroencephalogram time sequence characteristics and the abnormal response event set, so that an electroencephalogram associated graph structure is obtained; Training a neural network model comprising a graph attention layer by taking the abnormal response event set as an abnormal label and taking the abnormal label as a supervision signal to obtain an electroencephalogram attention network model, forward transmitting the electroencephalogram associated graph structure based on the electroencephalogram attention network model to obtain attention abnormal contribution weights, and carrying out abnormal attribution analysis on the attention abnormal contribution weights based on a graph neural network interpreter to obtain abnormal root data.
7. The method of closed loop analysis of acoustic stimulus-brain electrical response of claim 6, wherein the process of multi-dimensional time series feature extraction of the standard brain electrical data based on the response latency data and the response significance data comprises: Extracting electroencephalogram response event groups corresponding to the stimulus event identifiers from the standard electroencephalogram data by taking the stimulus event identifiers as a main key, and sequencing the electroencephalogram response events in the electroencephalogram response event groups according to the stimulus presentation time sequence to obtain an electroencephalogram event sequence set; The method comprises the steps of respectively carrying out multi-scale time sequence convolution on response latency data and response significance data based on the electroencephalogram event sequence set to obtain latency time sequence features and significance time sequence features, extracting an electroencephalogram frequency band parameter sequence set from the electroencephalogram event sequence set, carrying out time sequence feature extraction on the electroencephalogram frequency band parameter sequence set based on a bidirectional long-short-term memory network to obtain frequency band time sequence features, and splicing the latency time sequence features, the significance time sequence features and the frequency band time sequence features into the electroencephalogram time sequence features.
8. The method of claim 7, wherein modeling the graph structure and modeling the abnormal context of the standard electroencephalogram data based on the electroencephalogram timing characteristics and the set of abnormal response events comprises: The electroencephalogram time sequence features are used as graph node features, an electroencephalogram node feature set is generated, a node edge set corresponding to the electroencephalogram node feature set is generated based on the sequence of the electroencephalogram event sequence set, a plurality of front electroencephalogram events corresponding to an abnormal response event set in the electroencephalogram event sequence set are collected into an abnormal event sequence set, abnormal context modeling is conducted on the abnormal response event set based on response significance data corresponding to the abnormal event sequence set and the response latency data, an abnormal context event sequence set is obtained, abnormal feature embedding is conducted on the node edge set based on the abnormal context event sequence set, an edge feature set is obtained, and an electroencephalogram associated graph structure is generated based on the electroencephalogram node feature set, the node edge set and the edge feature set.
9. The method of closed loop analysis of acoustic stimulus-brain electrical response of claim 8, wherein generating the response latency data, response significance data, and anomaly basis data based on neural signal analysis logic comprises: The stimulus event identification is used as a primary key, the response latency data, the response significance data and the abnormal root data are associated and integrated to obtain integrated analysis data, and event supplementary information is added to the integrated analysis data based on the standard electroencephalogram data to obtain enhanced analysis data; based on the neural signal analysis logic and the data type of the enhancement analysis data, mapping the corresponding chart type of the enhancement analysis data to obtain a chart configuration rule set; configuring a billboard layout template for the chart configuration rule set based on a preset user role to obtain a billboard structure template; Performing chart instantiation operation on the enhanced analysis data according to the chart configuration rule set to obtain an analysis chart set, and adding dynamic interaction logic for the analysis chart set to obtain a dynamic analysis chart set; and carrying out interface assembly on the dynamic analysis chart set according to the billboard structure template to obtain an electroencephalogram response analysis chart interface, and carrying out visual display on the electroencephalogram response analysis chart interface.
10. A closed-loop analysis system for sound stimulation-electroencephalogram response, for implementing a closed-loop analysis method for sound stimulation-electroencephalogram response according to any one of claims 1 to 9, characterized in that the system comprises a data acquisition module, a data preprocessing module, a data processing module, a data analysis module and an intelligent visualization module; the data acquisition module acquires real-time brain electrical data under sound stimulation based on a standardized interface of the brain electrical acquisition system; the data preprocessing module is used for carrying out artifact removal and data standardization operation on the real-time electroencephalogram data based on the stimulus event identification to obtain standard electroencephalogram data; The data processing module is used for carrying out response latency analysis on the standard electroencephalogram data based on the triggering time of sound stimulation and the peak time of electroencephalogram response to obtain response latency data; The data analysis module is used for extracting an abnormal response event set from the standard electroencephalogram data, and carrying out context feature modeling and abnormal root cause analysis on the abnormal response event set based on the response latency data and the response significance data to obtain abnormal root cause data; And the intelligent visualization module is used for generating the response latency data, the response significance data and the electroencephalogram response analysis chart interface corresponding to the abnormal root data based on the neural signal analysis logic and carrying out visual display on the electroencephalogram response analysis chart interface.

Description

Closed-loop analysis system and method for sound stimulation-electroencephalogram response Technical Field The invention relates to the technical field of electroencephalogram data analysis, in particular to a closed-loop analysis system and method for sound stimulation-electroencephalogram response. Background In the sound stimulation electroencephalogram response analysis level, the abnormal incubation period is identified without combining with the multidimensional statistical characteristics, the response significance analysis mostly adopts a fixed threshold value, cannot be dynamically adjusted according to historical data, and is difficult to adapt to the difference of electroencephalogram signal baselines of different testees and the change of experimental environments. For root cause analysis of abnormal response events, structural modeling based on multi-dimensional time sequence features is lacking, association relation between abnormal features and root causes is difficult to quantify, and the root causes are fuzzy in positioning and poor in interpretability. The personalized requirements of different roles such as researchers, clinicians and the like on data interpretation are difficult to meet, and experimental scheme optimization and clinical intervention decision-making cannot be supported efficiently. Accordingly, a closed loop analysis system and method of acoustic stimulus-electroencephalogram response is now provided. Disclosure of Invention In order to solve the above technical problems, the present invention is directed to a closed-loop analysis system and method for acoustic stimulation-electroencephalogram response. In order to achieve the above purpose, the invention provides a closed-loop analysis method of sound stimulation-electroencephalogram response, which comprises the following steps: acquiring real-time brain electrical data under sound stimulation based on a standardized interface of an brain electrical acquisition system, and performing artifact removal and data standardization operation on the real-time brain electrical data based on a stimulation event identifier to obtain standard brain electrical data; Performing response latency analysis on the standard electroencephalogram data based on the triggering time of sound stimulation and the peak time of electroencephalogram response to obtain response latency data; Performing response significance analysis on the standard electroencephalogram data based on a dynamic response intensity threshold value to obtain response significance data; Extracting an abnormal response event set from the standard electroencephalogram data, and carrying out context feature modeling and abnormal root cause analysis on the abnormal response event set based on the response latency data and the response significance data to obtain abnormal root cause data; And generating an electroencephalogram response analysis chart interface corresponding to the response latency data, the response significance data and the abnormal root data based on the nerve signal analysis logic, and visually displaying the electroencephalogram response analysis chart interface. Further, the process of acquiring the real-time brain electrical data under the sound stimulation based on the standardized interface of the brain electrical acquisition system comprises the following steps: carrying out multi-channel signal acquisition on electroencephalogram acquisition equipment in a sound stimulation scene to obtain an original electroencephalogram signal, carrying out stimulation event association and data encapsulation on the original electroencephalogram signal based on stimulation presentation logic to obtain encapsulated electroencephalogram data, and publishing the encapsulated electroencephalogram data to a preset nerve signal data platform based on a standardized interface of an electroencephalogram acquisition system; Receiving the real-time brain electrical information of the nerve signal data platform in a signal subscription mode, carrying out signal analysis on the real-time brain electrical information to obtain analysis brain electrical data, and carrying out stimulus response context association on the analysis brain electrical data to obtain real-time brain electrical data. Further, the process of artifact removal and data normalization operations on the real-time electroencephalogram data based on the stimulus event identification includes: The method comprises the steps of classifying the real-time electroencephalogram data according to the stimulus type, adding a stimulus tag, carrying out signal integrity check on the real-time electroencephalogram data based on the stimulus tag, and carrying out data deduplication on the real-time electroencephalogram data after the integrity check based on a stimulus event identifier and a signal acquisition time difference threshold value to obtain deduplication electroencephalogram data; the method comprises the steps