Forward reactivation in human hippocampal CA1 supports repetition learning and is preserved in the aging brain
Abstract
Repetition is a fundamental mechanism for strengthening episodic memory, yet the neural computations engaged during re-exposure remain disputed. Competing accounts attribute repetition-related gains either to retrospective reinstatement of prior events or to the prospective use of learned structure to anticipate what comes next. Using high-resolution functional MRI and representational similarity analysis in a lifespan cohort (N = 418, ages 36 -100), we tracked hippocampal representations as participants twice encoded two fixed sequences of face-name pairs, resolving reactivation by direction, distance and sequence. During repeated encoding, CA1 and CA2/3/DG representations spontaneously anticipated immediate next states, unprompted by explicit task demands. Forward, but not backward, reactivation predicted repetition-related gains, and showed no detectable age dependence. Only CA1 forward reactivation predicted memory gains after accounting for reactivation in other medial temporal lobe regions. Repetition learning is thus an active, temporally local prospective computation with a distinctive CA1 contribution that appears preserved well into old age.
Related
- Selective Hippocampal Reactivation Driven by Neocortical Feedback Supports Memory Transfer in Predictive-Coding Networks
- ERBB4 Perturbation Supports a Developmental Cytoskeletal Reactivation Hypothesis in Alzheimer’s Disease
- Hippocampal stimulation timed to memory reactivation shapes human sleep oscillatory dynamics and consolidation
- Preserved brain hierarchy supports residual vision without awareness after visual cortex damage
- Postsynaptic TAK1-GluN2B axis supports hippocampal synaptic integrity and cognitive function during aging
BibTeX
@misc{gao2026forward,
title = {{Forward reactivation in human hippocampal CA1 supports repetition learning and is preserved in the aging brain}},
author = {Zhiyao Gao and Li Zheng and Vinod Menon and Weidong Cai},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.20.760541},
url = {https://www.biorxiv.org/content/10.64898/2026.10.20.760541v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - Forward reactivation in human hippocampal CA1 supports repetition learning and is preserved in the aging brain
AU - Zhiyao Gao
AU - Li Zheng
AU - Vinod Menon
AU - Weidong Cai
PY - 2026
DA - 2026/10/20
PB - bioRxiv
DO - 10.64898/2026.10.20.760541
UR - https://www.biorxiv.org/content/10.64898/2026.10.20.760541v1
AB - Repetition is a fundamental mechanism for strengthening episodic memory, yet the neural computations engaged during re-exposure remain disputed. Competing accounts attribute repetition-related gains either to retrospective reinstatement of prior events or to the prospective use of learned structure to anticipate what comes next. Using high-resolution functional MRI and representational similarity analysis in a lifespan cohort (N = 418, ages 36 -100), we tracked hippocampal representations as participants twice encoded two fixed sequences of face-name pairs, resolving reactivation by direction, distance and sequence. During repeated encoding, CA1 and CA2/3/DG representations spontaneously anticipated immediate next states, unprompted by explicit task demands. Forward, but not backward, reactivation predicted repetition-related gains, and showed no detectable age dependence. Only CA1 forward reactivation predicted memory gains after accounting for reactivation in other medial temporal lobe regions. Repetition learning is thus an active, temporally local prospective computation with a distinctive CA1 contribution that appears preserved well into old age.
SN - 2692-8205
ER -