dorsal/arxiv
View SchemaAdaptive Filtering Enhances Information Transmission in Visual Cortex
| Authors | Tatyana O. Sharpee, Hiroki Sugihara, Andrei V. Kurgansky, Sergei P. Rebrik, Michael P. Stryker, Kenneth D. Miller |
|---|---|
| Categories | |
| ArXiv ID | q-bio/0611037 |
| URL | https://arxiv.org/abs/q-bio/0611037 |
| DOI | 10.1038/nature04519 |
| Journal | Nature, vol. 439, pp. 936- 942 (02/23/2006) |
Abstract
Sensory neuroscience seeks to understand how the brain encodes natural environments. However, neural coding has largely been studied using simplified stimuli. In order to assess whether the brain's coding strategy depend on the stimulus ensemble, we apply a new information-theoretic method that allows unbiased calculation of neural filters (receptive fields) from responses to natural scenes or other complex signals with strong multipoint correlations. In the cat primary visual cortex we compare responses to natural inputs with those to noise inputs matched for luminance and contrast. We find that neural filters adaptively change with the input ensemble so as to increase the information carried by the neural response about the filtered stimulus. Adaptation affects the spatial frequency composition of the filter, enhancing sensitivity to under-represented frequencies in agreement with optimal encoding arguments. Adaptation occurs over 40 s to many minutes, longer than most previously reported forms of adaptation.
{
"annotation_id": "8d1042ae-6f05-4cfc-97e9-36a4f3a40ab3",
"date_created": "2026-03-02T18:01:35.532000Z",
"date_modified": "2026-03-02T18:01:35.532000Z",
"file_hash": "4ea00c1b7fb6d2fcda36ce1fc405aa9479110fcd2f646cea38f9a13c78347cb1",
"private": false,
"record": {
"abstract": "Sensory neuroscience seeks to understand how the brain encodes natural\nenvironments. However, neural coding has largely been studied using simplified\nstimuli. In order to assess whether the brain\u0027s coding strategy depend on the\nstimulus ensemble, we apply a new information-theoretic method that allows\nunbiased calculation of neural filters (receptive fields) from responses to\nnatural scenes or other complex signals with strong multipoint correlations. In\nthe cat primary visual cortex we compare responses to natural inputs with those\nto noise inputs matched for luminance and contrast. We find that neural filters\nadaptively change with the input ensemble so as to increase the information\ncarried by the neural response about the filtered stimulus. Adaptation affects\nthe spatial frequency composition of the filter, enhancing sensitivity to\nunder-represented frequencies in agreement with optimal encoding arguments.\nAdaptation occurs over 40 s to many minutes, longer than most previously\nreported forms of adaptation.",
"arxiv_id": "q-bio/0611037",
"authors": [
"Tatyana O. Sharpee",
"Hiroki Sugihara",
"Andrei V. Kurgansky",
"Sergei P. Rebrik",
"Michael P. Stryker",
"Kenneth D. Miller"
],
"categories": [
"q-bio.NC"
],
"doi": "10.1038/nature04519",
"journal_ref": "Nature, vol. 439, pp. 936- 942 (02/23/2006)",
"title": "Adaptive Filtering Enhances Information Transmission in Visual Cortex",
"url": "https://arxiv.org/abs/q-bio/0611037"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "12a2f1fc-5850-4d9d-bd5f-1c77f8d705da",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}