Maximum entropy models as a tool for building precise neural controls

Cristina Savin, Gašper Tkačik

Research output: Contribution to journalReview article

Abstract

Neural responses are highly structured, with population activity restricted to a small subset of the astronomical range of possible activity patterns. Characterizing these statistical regularities is important for understanding circuit computation, but challenging in practice. Here we review recent approaches based on the maximum entropy principle used for quantifying collective behavior in neural activity. We highlight recent models that capture population-level statistics of neural data, yielding insights into the organization of the neural code and its biological substrate. Furthermore, the MaxEnt framework provides a general recipe for constructing surrogate ensembles that preserve aspects of the data, but are otherwise maximally unstructured. This idea can be used to generate a hierarchy of controls against which rigorous statistical tests are possible.

Original languageEnglish (US)
Pages (from-to)120-126
Number of pages7
JournalCurrent Opinion in Neurobiology
Volume46
DOIs
StatePublished - Oct 1 2017

Fingerprint

Entropy
Population Characteristics
Population

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Maximum entropy models as a tool for building precise neural controls. / Savin, Cristina; Tkačik, Gašper.

In: Current Opinion in Neurobiology, Vol. 46, 01.10.2017, p. 120-126.

Research output: Contribution to journalReview article

@article{335fa9287e42449c81656769e86db043,
title = "Maximum entropy models as a tool for building precise neural controls",
abstract = "Neural responses are highly structured, with population activity restricted to a small subset of the astronomical range of possible activity patterns. Characterizing these statistical regularities is important for understanding circuit computation, but challenging in practice. Here we review recent approaches based on the maximum entropy principle used for quantifying collective behavior in neural activity. We highlight recent models that capture population-level statistics of neural data, yielding insights into the organization of the neural code and its biological substrate. Furthermore, the MaxEnt framework provides a general recipe for constructing surrogate ensembles that preserve aspects of the data, but are otherwise maximally unstructured. This idea can be used to generate a hierarchy of controls against which rigorous statistical tests are possible.",
author = "Cristina Savin and Gašper Tkačik",
year = "2017",
month = "10",
day = "1",
doi = "10.1016/j.conb.2017.08.001",
language = "English (US)",
volume = "46",
pages = "120--126",
journal = "Current Opinion in Neurobiology",
issn = "0959-4388",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Maximum entropy models as a tool for building precise neural controls

AU - Savin, Cristina

AU - Tkačik, Gašper

PY - 2017/10/1

Y1 - 2017/10/1

N2 - Neural responses are highly structured, with population activity restricted to a small subset of the astronomical range of possible activity patterns. Characterizing these statistical regularities is important for understanding circuit computation, but challenging in practice. Here we review recent approaches based on the maximum entropy principle used for quantifying collective behavior in neural activity. We highlight recent models that capture population-level statistics of neural data, yielding insights into the organization of the neural code and its biological substrate. Furthermore, the MaxEnt framework provides a general recipe for constructing surrogate ensembles that preserve aspects of the data, but are otherwise maximally unstructured. This idea can be used to generate a hierarchy of controls against which rigorous statistical tests are possible.

AB - Neural responses are highly structured, with population activity restricted to a small subset of the astronomical range of possible activity patterns. Characterizing these statistical regularities is important for understanding circuit computation, but challenging in practice. Here we review recent approaches based on the maximum entropy principle used for quantifying collective behavior in neural activity. We highlight recent models that capture population-level statistics of neural data, yielding insights into the organization of the neural code and its biological substrate. Furthermore, the MaxEnt framework provides a general recipe for constructing surrogate ensembles that preserve aspects of the data, but are otherwise maximally unstructured. This idea can be used to generate a hierarchy of controls against which rigorous statistical tests are possible.

UR - http://www.scopus.com/inward/record.url?scp=85028597997&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028597997&partnerID=8YFLogxK

U2 - 10.1016/j.conb.2017.08.001

DO - 10.1016/j.conb.2017.08.001

M3 - Review article

VL - 46

SP - 120

EP - 126

JO - Current Opinion in Neurobiology

JF - Current Opinion in Neurobiology

SN - 0959-4388

ER -