Method for removing multi-site functional magnetic resonance imaging heterogeneity for predicting diseases

Abstract

The invention discloses a method for removing multi-site functional magnetic resonance imaging heterogeneity for predicting diseases. The method is characterized by comprising the following steps: acquiring multi-site resting state fMRI data and a phenotype data set; extracting fMRI data by using an HO brain map, performing Fisher transformation on a brain network function connection matrix, and performing feature engineering; brain network features are decoupled and expressed as site invariant and site specific embedding through a double-end encoder, and the encoder is trained by using regular constraint; constructing a population map based on site specific embedding and phenotypic information, and performing propagation and conversion of site invariant embedding on the population map through a map convolutional neural network; and taking a category with a higher probability in the two-dimensional vector obtained by final conversion from a large number of unlabeled nodes contained in the population map, and the like. Compared with the prior art, the method has a high-accuracy diagnosis effect, errors caused by a large amount of manual intervention are effectively avoided, and the working efficiency and the accuracy of the diagnosis result are greatly improved.

Description

technical field [0001] The invention relates to the technical field of computer-aided diagnosis, in particular to a method for removing heterogeneity of multi-site functional magnetic resonance imaging for predicting diseases. Background technique [0002] Resting-state functional magnetic resonance imaging (rs-fMRI) has the ability to capture the interaction of regions of interest in the brain. The findings suggest that brain functional connectivity patterns can serve as diagnostic biomarkers for a range of psychiatric disorders, including Alzheimer's disease, depression, and autism. Due to the difficulty of diagnosing mental illnesses, computer-aided diagnosis has high hopes. However, in the process of using large-scale shared multi-site fMRI data to train a diagnostic model, the diagnostic model is difficult to achieve the desired effect due to data heterogeneity caused by many differences in acquisition protocols and scanner types used at each site. Therefore, it is of...

AI Technical Summary

Problems solved by technology

But this decoupling represents learning constraints, propagating heterogeneity-removing fMRI information across population maps, and combining the site-specific embeddings obtained from the decoupling to construct population maps for large-scale ethos removal of multi-site fMRI heterogeneity Disease diagnosis, so far there is no public report on related technologies

Images