Gender difference detection method based on electroencephalogram emotion recognition

Abstract

The invention provides an electroencephalogram feature recognition method based on a long-short term memory graph neural network, which comprises the steps of extracting differential entropy features from an electroencephalogram data set, converting the differential entropy features into a feature matrix of a representation graph, training a long-short term memory graph neural network model, collecting brain function connection information and a time sequence relationship of feature data at the same time, and finally, realizing emotion recognition by using the trained network model. According to the invention, brain function connection information and time sequence information are fully utilized through the long-short term memory graph neural network, multiple common and representative data sets are recognized, meanwhile, the most common differential performance is analyzed, and gender difference characteristics in emotion-related electroencephalogram activities are verified.

Description

technical field [0001] The present invention relates to a technology in the field of EEG signals, in particular to a gender difference detection method based on EEG signal emotion recognition. Background technique [0002] Compared with the facial expression, voice, body posture and other signal recognition often used by researchers in the past, the EEG signal can more delicately represent the brain activity related to the subject's emotion, so it is considered to be the most effective signal for emotion recognition. Form, and therefore, the research on emotion recognition using EEG signals is getting more and more attention from academia and industry. However, because there are not only differences in physical structures such as scalp impedance and head shape, but also differences in mental mechanisms such as thinking styles, mental states, and cognitive abilities among different individuals, the characteristic patterns of EEG signals are highly dependent on the individual ...

AI Technical Summary

Problems solved by technology

[0004] Although there have been research results on gender differences in EEG signal emotion recognition, the existing work has relatively large limitations: first, most of them only use data from a single data set, and the model performance is insufficient, so the conclusions are not universal Secondly, the data processing methods and operation steps in these studies are not the same, and there is no guarantee that the same conclusion can be obtained stably under other experimental configurations; in addition, although various EEG differences are found in these results form, but there are still inconsistent conclusions among different studies, which cannot be accepted as a general conclusion

Images