Format

Send to

Choose Destination
Trends Cogn Sci. 2010 Mar;14(3):119-30. doi: 10.1016/j.tics.2010.01.003. Epub 2010 Feb 12.

Statistically optimal perception and learning: from behavior to neural representations.

Author information

1
National Volen Center for Complex Systems, Brandeis University, Volen 208/MS 013, Waltham, MA 02454, USA. fiser@brandeis.edu

Abstract

Human perception has recently been characterized as statistical inference based on noisy and ambiguous sensory inputs. Moreover, suitable neural representations of uncertainty have been identified that could underlie such probabilistic computations. In this review, we argue that learning an internal model of the sensory environment is another key aspect of the same statistical inference procedure and thus perception and learning need to be treated jointly. We review evidence for statistically optimal learning in humans and animals, and re-evaluate possible neural representations of uncertainty based on their potential to support statistically optimal learning. We propose that spontaneous activity can have a functional role in such representations leading to a new, sampling-based, framework of how the cortex represents information and uncertainty.

PMID:
20153683
PMCID:
PMC2939867
DOI:
10.1016/j.tics.2010.01.003
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Elsevier Science Icon for PubMed Central
Loading ...
Support Center