Application of deep canonically correlated sparse autoencoder for the classification of schizophrenia

Abstract Background and objective Imaging geneticshas been widely used to help diagnose and treat mental illness, e.g., schizophrenia, by combining magnetic resonance imagingof the brainand genomic information for comprehensive and systematic analysis. As a result, utilizing the correlation between magnetic resonance imagingof the brainand genomic information is becoming an important challenge. Methods In this paper, the joint analysis of single nucleotide polymorphisms and functional magnetic resonance imagingis conducted for comprehensive study of schizophrenia. We developed a deep canonically correlatedsparse autoencoderto classify schizophrenia patients from healthy controls, which can address the limitation of many existing methods such as canonical correlationanalysis, deep canonical correlationanalysis and sparse autoencoder. Results The proposed deep canonically correlatedsparse autoencodercan not only use complex nonlinear transformation and dimension reduction, but also achieve more accurate classifications. Our experiments showed the proposed method achieved an accuracy of 95.65% for SNP data sets and an accuracy of 80.53% for fMRI data sets. Conclusions Experiments demonstrated higher accuracy of using the proposed method over other conventional models when classifying schizophrenia patients and healthy controls.