Control of synaptic plasticity in deep cortical networks

Pieter R. Roelfsema,Anthony Holtmaat
DOI: https://doi.org/10.1038/nrn.2018.6
IF: 38.755
2018-02-16
Nature Reviews Neuroscience
Abstract:Key PointsIn addition to presynaptic and postsynaptic mechanisms, synaptic plasticity depends on neuromodulatory substances and feedback connections from higher-order cortical and thalamic brain regionsSynaptic plasticity in the brain depends on reward-prediction errors and on selective attention. Neuromodulatory systems code for the reward-prediction errors, and feedback connections from the response-selection stage mediate top-down attention effectsThe combined influence of feedback connections and neuromodulatory substances on plasticity enables powerful learning rules for the training of 'deep', multilayered neuronal networksFeedback connections project to cortical layers that are distinct from feedforward input, where they impinge on distal dendritic segments, separate excitatory neuronal populations or inhibitory interneuronsFeedback connections gate plasticity in cortical pyramidal neurons by promoting NMDA-receptor-driven calcium entry into dendrites and by disinhibiting the cortical column through activation of vasoactive-intestinal-peptide-positive interneurons (among others)Synaptic tags are biochemical processes that make synapses eligible for plasticity. Neuromodulators released later can interact with tagged synapses to increase or decrease synaptic strength
neurosciences
What problem does this paper attempt to address?