This website requires JavaScript.

Hybrid Representation Learning for Cognitive Diagnosis in Late-Life Depression Over 5 Years with Structural MRI

Lintao ZhangLihong WangMinhui Yu ...+3 Mingxia Liu
Dec 2022
Late-life depression (LLD) is a highly prevalent mood disorder occurring inolder adults and is frequently accompanied by cognitive impairment (CI).Studies have shown that LLD may increase the risk of Alzheimer's disease (AD).However, the heterogeneity of presentation of geriatric depression suggeststhat multiple biological mechanisms may underlie it. Current biologicalresearch on LLD progression incorporates machine learning that combinesneuroimaging data with clinical observations. There are few studies on incidentcognitive diagnostic outcomes in LLD based on structural MRI (sMRI). In thispaper, we describe the development of a hybrid representation learning (HRL)framework for predicting cognitive diagnosis over 5 years based on T1-weightedsMRI data. Specifically, we first extract prediction-oriented MRI features viaa deep neural network, and then integrate them with handcrafted MRI featuresvia a Transformer encoder for cognitive diagnosis prediction. Two tasks areinvestigated in this work, including (1) identifying cognitively normalsubjects with LLD and never-depressed older healthy subjects, and (2)identifying LLD subjects who developed CI (or even AD) and those who stayedcognitively normal over five years. To the best of our knowledge, this is amongthe first attempts to study the complex heterogeneous progression of LLD basedon task-oriented and handcrafted MRI features. We validate the proposed HRL on294 subjects with T1-weighted MRIs from two clinically harmonized studies.Experimental results suggest that the HRL outperforms several classical machinelearning and state-of-the-art deep learning methods in LLD identification andprediction tasks.