Task-Parametrized Dynamics: Representation of Time and Decisions in Recurrent Neural Networks
Journal:
bioRxiv
Published Date:
May 22, 2026
Abstract
How do recurrent neural networks (RNNs) internally represent elapsed time to initiate responses after learned delays? To address this question, we trained RNNs on delayed decision-making tasks with progressively increasing temporal demands, including binary decisions, context-dependent decisions, and perceptual integration. We analyzed trained networks using connectivity statistics, eigenvalue spectra, readout alignment, and low-dimensional population trajectories. Across tasks, networks converged to qualitatively distinct but behaviourally comparable dynamical solutions, including oscillatory and non-oscillatory (ramping/decaying) regimes, consistent with solution degeneracy. Population activity remained low-dimensional and distributed across recurrent units rather than localized to individual neurons. Readout alignment was strongly epoch-dependent: activity evolved largely in the readout-null subspace prior to response generation and became increasingly aligned with the output dimension near decision time. In sign-symmetric tasks, trained networks preserved an approximate sign-flip equivariance inherited from architecture and training symmetry, despite independent noisy perturbations across trials, yielding mirrored population responses across stimulus sign. Together, these results show that temporal and decision-related computations can emerge through multiple dynamical regimes, while maintaining structured low-dimensional representations and comparable behavioural performance, mirroring biological principles of degeneracy and functional redundancy.
