TY - JOUR

T1 - Breaking the excitation-inhibition balance makes the cortical network’s space-time dynamics distinguish simple visual scenes

AU - Roland, Per E.

AU - Bonde, Lars H.

AU - Forsberg, Lars E.

AU - Harvey, Michael A.

PY - 2017/3

Y1 - 2017/3

N2 - Brain dynamics are often taken to be temporal dynamics of spiking and membrane potentials in a balanced network. Almost all evidence for a balanced network comes from recordings of cell bodies of few single neurons, neglecting more than 99% of the cortical network. We examined the space-time dynamics of excitation and inhibition simultaneously in dendrites and axons over four visual areas of ferrets exposed to visual scenes with stationary and moving objects. The visual stimuli broke the tight balance between excitation and inhibition such that the network exhibited longer episodes of net excitation subsequently balanced by net inhibition, in contrast to a balanced network. Locally in all four areas the amount of net inhibition matched the amount of net excitation with a delay of 125 ms. The space-time dynamics of excitation-inhibition evolved to reduce the complexity of neuron interactions over the whole network to a flow on a low-(3)-dimensional manifold within 80 ms. In contrast to the pure temporal dynamics, the low dimensional flow evolved to distinguish the simple visual scenes.

AB - Brain dynamics are often taken to be temporal dynamics of spiking and membrane potentials in a balanced network. Almost all evidence for a balanced network comes from recordings of cell bodies of few single neurons, neglecting more than 99% of the cortical network. We examined the space-time dynamics of excitation and inhibition simultaneously in dendrites and axons over four visual areas of ferrets exposed to visual scenes with stationary and moving objects. The visual stimuli broke the tight balance between excitation and inhibition such that the network exhibited longer episodes of net excitation subsequently balanced by net inhibition, in contrast to a balanced network. Locally in all four areas the amount of net inhibition matched the amount of net excitation with a delay of 125 ms. The space-time dynamics of excitation-inhibition evolved to reduce the complexity of neuron interactions over the whole network to a flow on a low-(3)-dimensional manifold within 80 ms. In contrast to the pure temporal dynamics, the low dimensional flow evolved to distinguish the simple visual scenes.

KW - Balanced network

KW - Cerebral cortex

KW - Dynamical systems

KW - Inhibition

KW - Mechanics of vision

KW - Voltage sensitive dye

U2 - 10.3389/fnsys.2017.00014

DO - 10.3389/fnsys.2017.00014

M3 - Journal article

C2 - 28377701

AN - SCOPUS:85018302081

VL - 11

JO - Frontiers in Systems Neuroscience

JF - Frontiers in Systems Neuroscience

SN - 1662-5137

M1 - 14

ER -