Multimodal brain network feature fusion method based on multi-task learning

Abstract

The invention discloses a multimodal brain network feature fusion method based on multi-task learning, and the multimodal brain network feature fusion method based on the multi-task learning includes the steps of preprocessing the obtained functional magnetic resonance imaging (fMRI) images and diffusion tensor imaging (DTI) images, registrating the preprocessed fMRI image to the standard AAL template, carrying out a fiber tracking for preprocessed DTI images, calculating fiber anisotropy (FA) value, and constructing structure connection matrix through the AAL template. Clustering coefficient of each brain area in a function connection matrix and the structure connection matrix is calculated to be regarded as function features and structure features. As two different tasks, the function features and the structure features assess an optimal feature set by solving the problem of multi-task learning optimization. The method uses information with multiple modalities complementing each other to learn simultaneously and to classify, improves the classification accuracy, solves the problems that a single task feature does not consider the correlation between features, and the fact that only one modality feature is used for pattern classification can bring to insufficient amount of information.

Description

technical field [0001] The present invention belongs to the technical field of biological information, relates to multimodal pattern recognition technology, and specifically relates to resting-state functional magnetic resonance imaging (resting-state functional magnetic resonance imaging, rs-fMRI) and diffusion tensor imaging (diffusion tensor imaging, DTI) feature fusion method. Background technique [0002] In order to provide more clinical imaging indicators, many current studies use the analysis of the differences of magnetic resonance images, and use the significant difference indicators as the feature set of the support vector machine, and classify the samples through the support vector machine. [0003] Magnetic resonance imaging can obtain not only functional information, but also rich structural information. However, most of the current MRI pattern classification studies are based on a single modality (such as functional magnetic resonance, structural magnetic res...

AI Technical Summary

Problems solved by technology

However, most current MRI pattern classification studies are based on a single modality (such as functional magnetic resonance, structural magnetic resonance, and diffusion tensor imaging), and rarely integrate functional information with structural information to further improve the accuracy of pattern classification; and In the feature selection process, most of them are based on univariate t statistical test. This feature selection method will ignore the subtle differences between the characteristics of the two groups of samples, and does not consider the relationship between the characteristics of the modes.

Images