Abstract:In neural circuits, synaptic strengths influence neuronal activity by shaping network dynamics, and neuronal activity influences synaptic strengths through activity-dependent plasticity. Motivated by this fact, we study a recurrent-network model in which neuronal units and synaptic couplings are interacting dynamic variables, with couplings subject to Hebbian modification with decay around quenched random strengths. Rather than assigning a specific role to the plasticity, we use dynamical mean-field theory and other techniques to systematically characterize the neuronal-synaptic dynamics, revealing a rich phase diagram. Adding Hebbian plasticity slows activity in already chaotic networks and can induce chaos in otherwise quiescent networks. Anti-Hebbian plasticity quickens activity and produces an oscillatory component. Analysis of the Jacobian shows that Hebbian and anti-Hebbian plasticity push locally unstable modes toward the real and imaginary axes, respectively, explaining these behaviors. Both random-matrix and Lyapunov analysis show that strong Hebbian plasticity segregates network timescales into two bands, with a slow, synapse-dominated band driving the dynamics, suggesting a flipped view of the network as synapses connected by neurons. For increasing strength, Hebbian plasticity initially raises the complexity of the dynamics, measured by the maximum Lyapunov exponent and attractor dimension, but then decreases these metrics, likely due to the proliferation of stable fixed points. We compute the marginally stable spectra of such fixed points as well as their number, showing exponential growth with network size. Finally, in chaotic states with strong Hebbian plasticity, a stable fixed point of neuronal dynamics is destabilized by synaptic dynamics, allowing any neuronal state to be stored as a stable fixed point by halting the plasticity. This phase of freezable chaos offers a new mechanism for working memory. https://doi.org/10.1103/PhysRevX.14.021001 Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI. Published by the American Physical Society

Neural networks with transient state dynamics

Multiperiodicity of Periodically Oscillated Discrete-Time Neural Networks with Transient Excitatory Self-Connections and Sigmoidal Nonlinearities

Alternation of up and down states at a dynamical phase-transition of a neural network with spatiotemporal attractors

Learning in cognitive systems with autonomous dynamics

Metastable Dynamics of Neural Circuits and Networks.

Neural networks with dynamical synapses: from mixed-mode oscillations and spindles to chaos

Theory of Coupled Neuronal-Synaptic Dynamics

Attractor Metadynamics in Adapting Neural Networks

Irregular dynamics in up and down cortical states

Bistability and Tunable Gating in a Recurrent Neuronal Network with Synaptic Failure and Hidden Neurons

Drifting states and synchronization induced chaos in autonomous networks of excitable neurons

Learning and Structure of Neuronal Networks

Intrinsic adaptation in autonomous recurrent neural networks

Stimulus-Driven and Spontaneous Dynamics in Excitatory-Inhibitory Recurrent Neural Networks for Sequence Representation

How Can Anomalous-Diffusion Neural Networks Under Connectomics Generate Optimized Spatiotemporal Dynamics

Dynamics of Networks of Excitatory and Inhibitory Neurons in Response to Time-Dependent Inputs

Stability switches, oscillatory multistability, and spatio-temporal patterns of nonlinear oscillations in recurrently delay coupled neural networks

Stochastic cellular automata model of neural networks

Identification of Recurrent Dynamics in Distributed Neural Populations

Transitions among metastable states underlie context-dependent working memories in a multiple timescale network