Method for extracting brain function states

Abstract

The present invention discloses a method for extracting brain function states. The method comprises the following steps: 1) extracting time series from functional magnetic resonance imaging; 2) dividing the time series into overlapping sub-segments of fixed length and calculating correlation coefficients among the sub-segments to construct a time-dependent dynamic function network; 3) vectorizingthe dynamic function network and matching the vectorized dynamic function network into a functional connection aggregation matrix; 4) improving and training a deep autoencoder model and inputting thefunctional connection aggregation matrix into the deep autoencoder for dimensionality reduction to obtain several feature vectors; 5) determining the optimal number of clusters and clustering the matrix composed of the feature vectors to obtain public functional networks of different functional states; and 6) analyzing residence time, conversion times and stability of the various functional statesand determining the brain function states. The method has reference value for study of brain cognitive functions and is helpful for discovering the differences among different functional states and the changing laws among the functional states.

Description

technical field [0001] The invention belongs to the technical field of biomedical information, and specifically relates to a method for extracting brain function states. [0002] technical background [0003] The human brain is an extremely complex system that exists in nature. All kinds of neurons are connected together through synapses, forming a very complex network structure, which is the structural basis for various physiological and cognitive activities of the brain. During the active or passive activity of the brain caused by external stimuli, each neuron or neural dynamic process extends into a complex brain functional network, which is an intuitive description of the changes in brain neural activity. In previous studies, functional connectivity in brain functional networks was a research hotspot. Functional connectivity is closely related to structural connectivity and is limited by structural connectivity, but structural connectivity can also be used to predict the...

AI Technical Summary

Problems solved by technology

For example, the dimensionality reduction method of brain network state observation matrix based on deep autoencoder (Yang Baojie, Wang Bin, Xue Jie, Dai Zhaokun, Liu Hui, Xiong Xin. The dimensionality reduction method of brain network state observation matrix based on deep autoencoder. Sensor and Microsystems, 2017(1):9-12.), using standard deep autoencoders to reduce the dimensionality of the brain functional network, to a certain extent solves the problem of "dimension disaster" faced by clustering, but after dimensionality reduction The features did not improve the clustering performance very well

Images