Disentangling structure-function relationships between the human hippocampus and the whole brain using track-weighted dynamic functional connectivity.

Abstract

Understanding how structural and functional connectivity shape hippocampal interactions with the rest of the brain is critical for elucidating its role in cognition. Here, we combine high-resolution diffusion MRI, a novel fibre-tracking pipeline designed to specifically probe anatomical connectivity of the in vivo human hippocampus, and track-weighted dynamic functional connectivity (TW-dFC) to investigate how direct anatomical connections between the hippocampus and the rest of the brain relate to time-varying functional interactions. In Study 1, TW-dFC maps were computed for 10 participants from the Human Connectome Project and subjected to ICA and k-means clustering to derive a data-driven parcellation of the hippocampus based on its structure-function relationships. This revealed circumscribed clusters distributed along anterior-posterior and medial-lateral axes, which broadly aligned with hippocampal subfields. In Study 2, we examined the resting-state functional connectivity profiles of each TW-dFC derived cluster in an independent sample of 100 participants. Each hippocampal cluster displayed distinct patterns of functional connectivity with specific substructures within medial temporal, parietal, frontal and occipital cortices as well as subcortical and cerebellar regions. Our findings demonstrate that TW-dFC provides a powerful framework for anatomically informed functional parcellation of the hippocampus and offers new insights into the structural-functional organisation underlying hippocampal-(sub)cortical interactions. Our approach opens new avenues for probing memory systems in health and their disruption in aging and disease.