FROM PERSONAL PAST TO FUTURE: EFFECTIVE CONNECTIVITY WITHIN BRAIN NETWORKS SUPPORTING SELF-PROJECTION ACROSS TIME

Autobiographical memory relies on the brain’s ability to navigate through the mental timeline, yet the directed neural interactions supporting this process remain elusive. Here, we used a data-driven approach to uncover the dynamic architecture of autobiographical retrieval. Directed functional connectivity was estimated using the Causal Dynamic Network (CDN) framework, a model-free method that infers directional interactions among brain regions directly from fMRI signals. Twenty-three participants performed a spatiotemporal interference task involving self-generated personal events from the remote and recent past, or the near and far future. Intrinsic connections analysis revealed a distributed baseline network across prefrontal, parietal, temporal, hippocampal, and cerebellar regions, supporting the control of self-referential and mnemonic processes. The contrast overall past versus overall future revealed positive cerebellar input to the angular gyrus for future events. The most robust condition-specific network-level reconfiguration emerged for far versus near past: projection towards the far past was characterized by a cascade of downregulatory functional influences, from the cerebellum to the hippocampus, and from the hippocampus to prefrontal and cingulate cortices. Finally, the contrast near past versus near future showed strengthened cerebellar cross-hemispheric communication and parietal-to-cerebellar influences for the past, and higher engagement of frontal areas for projection towards the future. Together, these findings demonstrate that autobiographical retrieval and future simulation rely on condition-sensitive, directed connectivity patterns, with hippocampal disengagement in remote memory and cerebellar contribution across both past and future projection.




Figure 4: Estimated cluster of directed connections for the Far past vs Near past condition.