Related Experiment Video
Updated: May 22, 2025
06:35
Basics of Multivariate Analysis in Neuroimaging Data
Published on: July 24, 2010
16.8K
Characterizing multivariate regional hubs for schizophrenia classification, sex differences, and brain age estimation
Yuzheng Nie1,2, Taslim Murad1, Hui-Yuan Miao1
1Department of Radiology, Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO, USA.
Medrxiv : the Preprint Server for Health Sciences
|March 17, 2025
Summary
This study used artificial intelligence (AI) to identify brain patterns in schizophrenia (SZ) and aging. The deep learning model revealed specific regional brain changes in SZ patients, differing by sex, and highlighted aging-related brain regions.
Area of Science:
- Neuroimaging
- Artificial Intelligence
- Psychiatry
Background:
- Schizophrenia (SZ) diagnosis and treatment can be improved by understanding its neurobiological underpinnings.
- Sex differences in brain structure and aging processes are not fully understood in SZ.
Purpose of the Study:
- To investigate multivariate regional brain patterns for schizophrenia classification, sex differences, and brain age prediction.
- To utilize structural MRI, demographics, and explainable AI for enhanced analysis.
Main Methods:
- Employed and compared various AI models including Support Vector Classifier (SVC), k-nearest neighbor (KNN), and Deep Learning (DL) for classification and regression.
- Integrated the best-performing DL model with Shapley Additive Explanations (SHAP) to identify significant multivariate brain regional patterns.
Main Results:
- Deep Learning (DL) models outperformed other AI models in schizophrenia classification, sex difference identification, and brain age prediction.
- The integrated DL-SHAP approach identified specific anatomical changes in SZ patients (e.g., left pallidum, insula, hippocampus, putamen) that varied between sexes.
- Key brain regions associated with aging were identified in both healthy controls and SZ patients.
Conclusions:
- This study successfully employed predictive modeling and explainable AI to reveal complex multivariate brain regions involved in schizophrenia, sex differences, and brain aging.
- The findings offer deeper insights into the neurobiological mechanisms of schizophrenia, paving the way for precision medicine and improved diagnosis/treatment.
