Cerebellar learning using perturbations.

Published online

Journal Article

The cerebellum aids the learning of fast, coordinated movements. According to current consensus, erroneously active parallel fibre synapses are depressed by complex spikes signalling movement errors. However, this theory cannot solve the credit assignment problem of processing a global movement evaluation into multiple cell-specific error signals. We identify a possible implementation of an algorithm solving this problem, whereby spontaneous complex spikes perturb ongoing movements, create eligibility traces and signal error changes guiding plasticity. Error changes are extracted by adaptively cancelling the average error. This framework, stochastic gradient descent with estimated global errors (SGDEGE), predicts synaptic plasticity rules that apparently contradict the current consensus but were supported by plasticity experiments in slices from mice under conditions designed to be physiological, highlighting the sensitivity of plasticity studies to experimental conditions. We analyse the algorithm's convergence and capacity. Finally, we suggest SGDEGE may also operate in the basal ganglia.

Full Text

Duke Authors

Cited Authors

  • Bouvier, G; Aljadeff, J; Clopath, C; Bimbard, C; Ranft, J; Blot, A; Nadal, J-P; Brunel, N; Hakim, V; Barbour, B

Published Date

  • November 12, 2018

Published In

Volume / Issue

  • 7 /

PubMed ID

  • 30418871

Pubmed Central ID

  • 30418871

Electronic International Standard Serial Number (EISSN)

  • 2050-084X

Digital Object Identifier (DOI)

  • 10.7554/eLife.31599

Language

  • eng

Conference Location

  • England