Integrating Multimodal Neuroimaging and Genetics: A Structurally-Linked Sparse Canonical Correlation Analysis Approach.

ABSTRACT: Neuroimaging genetics represents a multivariate approach aimed at elucidating the intricate relationships between high-dimensional genetic variations and neuroimaging data. Predominantly, existing methodologies revolve around Sparse Canonical Correlation Analysis (SCCA), a framework we expand to 1) encompass multiple imaging modalities and 2) promote the simultaneous identification of structurally linked features across imaging modalities. The structurally linked brain regions were assessed using diffusion tensor imaging, which quantifies the presence of neuronal fibers, thereby grounding our approach in biologically well-founded prior knowledge within the SCCA model. In our proposed structurally linked SCCA framework, we leverage T1-weighted MRI and functional MRI (fMRI) time series data

SUBMITTER: Chung J 

PROVIDER: S-EPMC11505868 | biostudies-literature | 2024

REPOSITORIES: biostudies-literature