Toggle Main Menu Toggle Search

Open Access padlockePrints

Spatio-temporal prediction and inference by V1 neurons

Lookup NU author(s): Dr Kun Guo, Dr Robert Robertson, Maribel Pulgarin Montoya, Dr Angel Nevado, Dr Stefano Panzeri, Professor Alexander Thiele, Professor Malcolm Young

Downloads

Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


Abstract

In normal vision, visual scenes are predictable, as they are both spatially and temporally redundant. Evidence suggests that the visual system may use the spatio-temporal regularities of the external world, available in the retinal signal, to extract information from the visual environment and better reconstruct current and future stimuli. We studied this by recording neuronal responses of primary visual cortex (area V1) in anaesthetized and paralysed macaques during the presentation of dynamic sequences of bars, in which spatio-temporal regularities and local information were independently manipulated. Most V1 neurons were significantly modulated by events prior to and distant from stimulation of their classical receptive fields (CRFs); many were more strongly tuned to prior and distant events than they were to CRFs bars; and several showed tuning to prior information without any CRF stimulation. Hence, V1 neurons do not simply analyse local contours, but impute local features to the visual world, on the basis of prior knowledge of a visual world in which useful information can be distributed widely in space and time. © The Authors (2007).


Publication metadata

Author(s): Guo K, Robertson RG, Pulgarin M, Nevado A, Panzeri S, Thiele A, Young MP

Publication type: Article

Publication status: Published

Journal: European Journal of Neuroscience

Year: 2007

Volume: 26

Issue: 4

Pages: 1045-1054

ISSN (print): 0953-816X

ISSN (electronic): 1460-9568

Publisher: Wiley-Blackwell Publishing Ltd.

URL: http://dx.doi.org/10.1111/j.1460-9568.2007.05712.x

DOI: 10.1111/j.1460-9568.2007.05712.x

PubMed id: 17714195


Altmetrics

Altmetrics provided by Altmetric


Share