Leakage and Second-Order Dynamics Improve Hippocampal RNN Replay

2026-02-20

AI Summary

Existing noisy recurrent neural network (RNN) models for simulating hippocampal neural replay suffer from low sampling efficiency and dynamic distortions, struggling to balance exploratory behavior with temporal compression. This work proposes an enhanced noisy RNN architecture that, for the first time, incorporates hidden-state momentum to enable temporally compressed replay. By integrating hidden-state leakage and a negative feedback adaptation mechanism, the model effectively mitigates replay lag caused by non-Markovian sampling. The approach demonstrates superior performance across multiple benchmarks—including 2D triangular trajectories, T-maze environments, and high-dimensional synthetic rat place cell data—significantly improving both the exploratory capacity and speed of replay while more accurately capturing key dynamical features observed in biological neural replay.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Reinforcement Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Biological neural networks (like the hippocampus) can internally generate"replay"resembling stimulus-driven activity. Recent computational models of replay use noisy recurrent neural networks (RNNs) trained to path-integrate. Replay in these networks has been described as Langevin sampling, but new modifiers of noisy RNN replay have surpassed this description. We re-examine noisy RNN replay as sampling to understand or improve it in three ways: (1) Under simple assumptions, we prove that the gradients replay activity should follow are time-varying and difficult to estimate, but readily motivate the use of hidden state leakage in RNNs for replay. (2) We confirm that hidden state adaptation (negative feedback) encourages exploration in replay, but show that it incurs non-Markov sampling that also slows replay. (3) We propose the first model of temporally compressed replay in noisy path-integrating RNNs through hidden state momentum, connect it to underdamped Langevin sampling, and show that, together with adaptation, it counters slowness while maintaining exploration. We verify our findings via path-integration of 2D triangular and T-maze paths and of high-dimensional paths of synthetic rat place cell activity.

Problem

• hippocampal replay
• recurrent neural networks
• path integration
• Langevin sampling
• temporal compression

Innovation

• hidden state leakage
• hidden state adaptation
• temporally compressed replay
• underdamped Langevin sampling
• noisy RNN