Abstract:Abstract Many cognitive processes involve transformations of distributed representations in neural populations, creating a need for population-level models. Recurrent neural network models fulfill this need, but there are many open questions about how their connectivity gives rise to dynamics that solve a task. Here, we present a method for finding the connectivity of networks for which the dynamics are specified to solve a task in an interpretable way. We apply our method to a working memory task by synthesizing a network that implements a drift-diffusion process over a ring-shaped manifold. We also use our method to demonstrate how inputs can be used to control network dynamics for cognitive flexibility and explore the relationship between representation geometry and network capacity. Our work fits within the broader context of understanding neural computations as dynamics over relatively low-dimensional manifolds formed by correlated patterns of neurons. Author Summary Neurons in the brain form intricate networks that can produce a vast array of activity patterns. To support goal-directed behavior, the brain must adjust the connections between neurons so that network dynamics can perform desirable computations on behaviorally relevant variables. A fundamental goal in computational neuroscience is to provide an understanding of how network connectivity aligns the dynamics in the brain to the dynamics needed to track those variables. Here, we develop a mathematical framework for creating recurrent neural network models that can address this problem. Specifically, we derive a set of linear equations that constrain the connectivity to afford a direct mapping of task-relevant dynamics onto network activity. We demonstrate the utility of this technique by creating and analyzing a set of network models that can perform a simple working memory task. We then extend the approach to show how additional constraints can furnish networks whose dynamics are controlled flexibly by external inputs. Finally, we exploit the flexibility of this technique to explore the robustness and capacity limitations of recurrent networks. This network synthesis method provides a powerful means for generating and validating hypotheses about how task-relevant computations can emerge from network dynamics.

Structure of activity in multiregion recurrent neural networks

Connectivity structure and dynamics of nonlinear recurrent neural networks

Engineering recurrent neural networks from task-relevant manifolds and dynamics

Cognitive network interactions through communication subspaces in large-scale models of the neocortex

Linking Neural Manifolds to Circuit Structure in Recurrent Networks

Self-organization of microcircuits in networks of neurons with plastic synapses

Dimension of activity in random neural networks

Multistability, local pattern formation, and global collective firing in a small-world network of non-leaky integrate-and-fire neurons

Theory of Coupled Neuronal-Synaptic Dynamics

Linking structure and activity in nonlinear spiking networks

Dynamic effective connectivity of inter-areal brain circuits

Identification of Recurrent Dynamics in Distributed Neural Populations

Modeling communication and switching nonlinear dynamics in multi-region neural activity

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

Neuromodulators generate multiple context-relevant behaviors in a recurrent neural network by shifting activity flows in hyperchannels

Dynamic control of neural manifolds

Synergistic pathways of modulation enable robust task packing within neural dynamics

Spatio-temporal activity patterns induced by triadic interactions in an in silico neural medium

Coherent chaos in a recurrent neural network with structured connectivity

Network structure of cerebral cortex shapes functional connectivity on multiple time scales

A Step Towards Uncovering The Structure of Multistable Neural Networks