Neural manifold under plasticity in a goal driven learning behaviour

Abstract

Neural activity is often low dimensional and dominated by only a few prominent neural covariation patterns. It has been hypothesised that these covariation patterns could form the building blocks used for fast and flexible motor control. Supporting this idea, recent experiments have shown that monkeys can learn to adapt their neural activity in motor cortex on a timescale of minutes, given that the change lies within the original low-dimensional subspace, also called neural manifold. However, the neural mechanism underlying this within-manifold adaptation remains unknown. Here, we show in a computational model that modification of recurrent weights, driven by a learned feedback signal, can account for the observed behavioural difference between within-and outside-manifold learning. Our findings give a new perspective, showing that recurrent weight changes do not necessarily lead to change in the neural manifold. On the contrary, successful learning is naturally constrained to a common subspace.

Visual exposure enhances stimulus encoding and persistence in primary cortex

“…For example, spike-frequency adaptation was shown to expand the memory exhibited by recurrent circuits (43) and different forms of biologically plausible synaptic learning rules have been employed to enhance computational performance of recurrent networks in an unsupervised fashion (44)(45)(46)(47). Furthermore, several studies made direct attempts to link learning in recurrent networks to optimization of state space dynamics (48,49) or metalearning (50). While neither of these recurrent models tried to explain how persistent responses after brief stimulation can interact with learning, they provide valuable insights into the various means by which refinements in internal network dynamics result in improved output performance.…”

Self-healing codes: how stable neural populations can track continually reconfiguring neural representations

“…Homeostatic preservation of predictive models may allow the brain to benefit from large networks during learning [138–140], and optimize these networks without extensive re-training. The processes we examine here may also be similar to those that allow transfer of learned motor skills despite gradual change in the readout of a brain-machine interface [141–143].…”

Estimating null and potent modes of feedforward communication in a computational model of cortical activity

“…One factor not examined here in the mean-rate model is the dimensionality of neural modes within the recurrent network formed by sender neurons. While this consideration has been the subject of extensive theoretical work [26,[41][42][43][44], the focus of the current work was the feedforward propagation of neural modes, and not their origin within recurrent circuits. Further work that examines both aspects of communication in a unified framework would provide an increased understanding of how interactions both within and across brain areas give rise to sensory perception and motor planning.…”

“…Other models are hard-wired to perform feedforward gating of neural activity [25] but offer no systematic way to control the transmission of null and potent modes. Finally, some models learn to generate low-dimensional representations of a signal [26,27] but focus on activity within a single brain area.…”