Search PubMed⌕ Search
Home/Archive/PubMed

PubMed · 12074953

A multi-layer sparse coding network learns contour coding from natural images.

Patrik O Hoyer·Aapo Hyvärinen
View original source ↗Export metadata ↓

Abstract

An important approach in visual neuroscience considers how the function of the early visual system relates to the statistics of its natural input. Previous studies have shown how many basic properties of the primary visual cortex, such as the receptive fields of simple and complex cells and the spatial organization (topography) of the cells, can be understood as efficient coding of natural images. Here we extend the framework by considering how the responses of complex cells could be sparsely represented by a higher-order neural layer. This leads to contour coding and end-stopped receptive fields. In addition, contour integration could be interpreted as top-down inference in the presented model.

Explore related subjects

Cerebral CortexHumansModels, NeurologicalNerve NetNeurophysiologyVision, OcularJournal ArticleResearch Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Patrik O Hoyer, Aapo Hyvärinen. 2002. A multi-layer sparse coding network learns contour coding from natural images.. https://doi.org/10.1016/s0042-6989(02)00017-2

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2026

Histone H3K9 methyltransferases regulate cortical growth by coordinating heterochromatin formation and neural progenitor dynamics.

DNA packaging into heterochromatin is a fundamental mechanism of transcriptional silencing, yet its role in regulating neural progenitor behavior during brain development remains poorly understood. Trimethylation of histone H3 lysine 9 (H3K9me3), catalyzed by the methyltransferases SETDB1, SUV39H1, and SUV39H2, is a defining feature of heterochromatin, but functional redundancy among these enzymes has obscured their developmental roles. Here, we generated a cortex-specific triple knockout mouse model lacking Setdb1, Suv39h1, and Suv39h2 to directly interrogate H3K9me3 function during corticogenesis. Combined loss of H3K9 methyltransferases caused genome-wide depletion of H3K9me3, disruption of neural progenitor cell-cycle progression, and impaired cortical neurogenesis, resulting in microcephaly. H3K9 methyltransferases preserve neural progenitor identity and function by silencing clustered protocadherins, meiosis-associated genes, and a cell-cycle restraint program through H3K9me3 deposition. Loss of H3K9me3 promoted local chromatin opening and increased transcription factor occupancy, enabling transposable elements to acquire cryptic enhancer activity and modulate proximal gene expression. Together, these findings establish H3K9me3 heterochromatin as an active regulator of neural progenitor dynamics and lineage fidelity, revealing a central epigenetic mechanism that restricts aberrant transcriptional programs to ensure cortical growth.

Cerebral Cortex↗
PubMed2002

Color vision: how the cortex represents color.

Our understanding of how we see color has benefited from the long tradition of visual psychophysics. More recently, models and methods from psychophysics are guiding modern neuroimaging experiments on color vision. Combining the two techniques can lead to discoveries that neither can make alone.

Cerebral Cortex↗