Post-stroke rehabilitation evaluation deep learning model construction method based on brain muscle network graph theory characteristics

Abstract

The invention discloses a post-stroke rehabilitation evaluation deep learning model construction method based on brain muscle network graph theory characteristics, and relates to the crossing field of neurophysiology and machine learning. According to the method, a pathological topological structure after stroke is represented through a brain muscle closed-loop function network, and on the basis, a deep learning model is further established based on graph theory characteristics to evaluate the recovery degree of a stroke patient and predict the recovery process; the method mainly considers consistency characteristics of hooked small-world network characteristics and a neural network in evaluation and prediction of dyskinesia, and how to realize multi-objective learning and joint optimization and the like. According to the method, a novel post-stroke hospitalization recovery period motor function evaluation and return visit period rehabilitation effect prediction method is constructed by utilizing electroencephalogram and myoelectricity bimodal neural electrophysiological information, and the clinical rehabilitation evaluation efficiency is expected to be improved, so that the method has an important application value.

Description

technical field [0001] The invention relates to the cross field of neurophysiology and machine learning, in particular to a method for building a deep learning model for post-stroke rehabilitation evaluation based on the graph theory feature of a closed-loop functional network of brain muscle. Background technique [0002] Stroke is the leading cause of death and disability worldwide. There are about 2 million new stroke patients in my country every year, of which 70%-80% have functional impairment. The long-term accompanying motor dysfunction after stroke is an important part of restricting patients' return to normal life. [0003] Rehabilitation therapy is the main diagnosis and treatment method for patients returning to society and family, but the premise is that the process of motor function remodeling after stroke is accurately assessed before rehabilitation training, which plays an important guiding role for clinicians to formulate treatment strategies. The principle...

AI Technical Summary

Problems solved by technology

The existing brain-muscle coupling is mainly based on single-channel analysis, without considering the local nature of the lesion after brain injury, and the quantitative characteristics of the coupling cannot fully represent the closed-loop evolution process of the bidirectional pathway

[0007] 2. The analysis of the connectivity of brain functional networks based on graph theory features provides a new idea, which has not been applied to the closed-loop process of brain-muscle coupling two-way pathways

Images