Brain function network classification method based on conditional mutual information and kernel density estimation

Abstract

The invention discloses a brain function network classification method based on conditional mutual information and kernel density estimation, and the method comprises the steps: estimating typical brain causal networks of different groups according to the preprocessed fMRI time sequences of different groups, and reserving the function connection between ROIs with large causal connection difference; the difference of the probability density of the strength of the same function connection under different labels is obtained through kernel density estimation, so that the causal strength between different function connections and the labels is obtained, the function connection with the high causal strength of the labels can be amplified, and the function connection with the low causal strength of the labels can be reduced; and fusing the BFN obtained from the preprocessed fMRI time sequence by using a Pearson's correlation coefficient with the causal knowledge obtained in the previous two steps to obtain a BFN with rich classification information, and sending the BFN into a BrainNetCNN for classification. Causal knowledge can be extracted from two aspects of the preprocessed fMRI time sequence and the BFN by using conditional mutual information and kernel density estimation, and the causal knowledge is fused into the original BFN to obtain a BFN with rich classification information.

Description

technical field [0001] The invention relates to the causal knowledge extraction and knowledge fusion method of fMRI functional magnetic resonance imaging data. Aiming at the goal of computer-aided diagnosis of brain diseases based on fMRI, a brain function network classification model based on conditional mutual information and kernel density estimation is designed. Background technique [0002] Brain functional network (BFN) is usually constructed from functional magnetic resonance imaging (fMRI) data, which can reveal patterns of brain functional activity. Specifically, BFN consists of nodes and edges, where each node corresponds to a brain region of interest (ROI), and each edge represents the functional connection between ROIs. Studies have shown that brain diseases are often associated with BFN functional connectivity (functional connection, FC) abnormalities, and BFN classification has been successfully applied to computer-aided diagnosis (CAD) of many brain diseases, ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem of insufficient performance of TML method and poor interpretability of DL method in current brain function network classification, the present invention proposes a brain function network based on conditional mutual information and kernel density estimation (CMI-KDE) classification model

Images