Forward and reverse delay-driven hippocampal replay without symmetric plasticity
arXiv:2608.21814
2026
Architecture
2 ideas extracted · analyzed Aug 29, 2026
What the math gives to ML
The paper gives a constructive delayed neural-field mechanism in which a circular recurrent state supports both forward and reverse replay, even with temporally asymmetric plasticity. Its transferable assets are delayed circular convolution, traveling-wave Fourier modes, and a derivative-based Hebbian rule that can learn sequence structure from a single noisy exposure. In neural networks, this suggests a compact recurrent replay-memory module whose direction and speed are dynamical modes rather than separately stored directed chains. The circular operator can also be implemented with FFTs, reducing recurrent computation from quadratic to near-linearithmic cost.
Ideas from this paper
✗ Mechanism failed
2026
Replace a directed sequence-memory chain with a circular recurrent state propagated by a learned delayed convolution. The same learned kernel can support forward and reverse replay because replay direction is a dynamical mode of the ring, rather than requiring plasticity to explicitly learn both forward and backward synapses.
Useful7/10
Difficulty5/10
Novelty8/10
✗ Failed on benchmark
2026
Regularize a circular recurrent kernel by directly controlling the growth rate and phase velocity of its Fourier modes. This converts replay-speed selection into a low-dimensional spectral control problem and can suppress unstable or excessively slow modes without adding recurrent parameters.
Useful6/10
Difficulty6/10
Novelty8/10