Modeling temporal response characteristics of V1 neurons with a dynamic normalization model

Samuel Mikaelian, Eero P. Simoncelli

Research output: Contribution to journalArticle

Abstract

We present a dynamic normalization model to characterize both the transient and the steady-state components of V1 simple and complex cell responses. Primary receptive field properties are chiefly determined by the convergence of LGN afferents. These linear responses are rectified, and subjected to shunting inhibition through cortical feedback, which accounts for the non-linear characteristics of the neuronal responses. The duration of the transient response is determined by the time delay and the low-pass filtering of the cortical feedback. In addition to accounting for basic non-linear behaviors such as response saturation and cross-orientation inhibition, the model is also able to reproduce several short-term contrast and pattern-selective adaptation effects.

Original languageEnglish (US)
Pages (from-to)1461-1467
Number of pages7
JournalNeurocomputing
Volume38-40
DOIs
StatePublished - Jun 1 2001

    Fingerprint

Keywords

  • Dynamics
  • Feedback
  • Normalization
  • Transients
  • V1

ASJC Scopus subject areas

  • Computer Science Applications
  • Cognitive Neuroscience
  • Artificial Intelligence

Cite this