Sparse coding of sensory inputs.

Bruno A Olshausen; David J Field

doi:10.1016/j.conb.2004.07.007

Home/Archive/PubMed

PubMed · 15321069

Sparse coding of sensory inputs.

Bruno A Olshausen·David J Field

View original source ↗Export metadata ↓

Abstract

Several theoretical, computational, and experimental studies suggest that neurons encode sensory information using a small number of active neurons at any given point in time. This strategy, referred to as 'sparse coding', could possibly confer several advantages. First, it allows for increased storage capacity in associative memories; second, it makes the structure in natural signals explicit; third, it represents complex data in a way that is easier to read out at subsequent levels of processing; and fourth, it saves energy. Recent physiological recordings from sensory neurons have indicated that sparse coding could be a ubiquitous strategy employed in several different modalities across different organisms.

Explore related subjects

Action Potentials Afferent Pathways Animals Brain Humans Models, Neurological Neurons, Afferent Sensation Signal Transduction Visual Cortex Journal Article Review

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Bruno A Olshausen, David J Field. 2004. Sparse coding of sensory inputs.. https://doi.org/10.1016/j.conb.2004.07.007

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

KEEP EXPLORING

Related citations

PubMed2006

Bootstrap-based methods for testing factor-by-curve interactions in generalized additive models: assessing prefrontal cortex neural activity related to decision-making.

In many situations the effect of a continuous covariate on response varies across groups defined by levels of a categorical variable. This paper addresses generalized additive models incorporating the so-called factor-by-curve interaction. A local scoring algorithm based on local linear kernel smoothers was used to estimate the model. Two different types of bootstrap-based procedures are proposed for testing interaction terms, namely, the likelihood ratio test, and a procedure based on an estimate of the interaction terms. Given the high computational cost involved, binning techniques were used to speed up computation in the estimation and testing processes. A simulation study was conducted to assess the validity of these bootstrap-based tests. This methodology was applied to studying prefrontal cortex neural activity associated with decision-making in monkeys. The proposed statistical procedure proved very useful in revealing the neural activity correlates of decision-making strategies adopted by monkeys in accordance with different behavioural tasks.

Action Potentials↗

PubMed2006

Excitatory response of prefrontal cortical fast-spiking interneurons to ventral tegmental area stimulation in vivo.

Prefrontal cortical (PFC) pyramidal neurons (PN) and fast spiking interneurons (FSI) receive dopaminergic (DA) and non-DA inputs from the ventral tegmental area (VTA). Although the responses of PN to VTA stimulation and DA administration have been extensively studied, little is known about the response of FSI to mesocortical activation. We explored this issue using single and double in vivo juxtacellular recordings of medial PFC PN and FSI with chemical VTA stimulation. Electrophysiological characteristics combined with Neurobiotin staining and parvalbumin immunohistochemistry allowed identification of recorded cells as FSI or PN. NMDA injection into the VTA increased firing in all FSI tested (n = 7), whereas most PN (7/11) responded with an inhibition. Furthermore, FSI excitation matching the temporal course of PN inhibition was observed with FSI-PN paired recordings (n = 5). These divergent electrophysiological responses to mesocortical activation could reflect PFC GABAergic interneurons contributing to silencing PN. Thus, the mesocortical system could provide a critical control of PFC circuits by simultaneously affecting FSI and PN firing.

Action Potentials↗

PubMed2006

Characterization of chronic constriction of the saphenous nerve, a model of neuropathic pain in mice showing rapid molecular and electrophysiological changes.

Neuropathic pain is one of the most inextricable problems encountered in clinics, because few facts are known about its etiology. Nerve injury often leads to allodynia and hyperalgesia, which are symptoms of neuropathic pain. The aim of this study was to understand some molecular and electrophysiological mechanisms of neuropathic pain after chronic constriction of the saphenous nerve (CCS) in mice. After surgery, CCS mice displayed significant allodynia and hyperalgesia, which were sensitive to acute systemic injection of morphine (4 mg/kg), gabapentin (50 mg/kg), amitriptyline (10 mg/kg), and the cannabinoid agonist WIN 55,212-2 (5 mg/kg). These behavioral changes were accompanied after surgery by an increase of c-Fos expression and by an overexpression of mu-opioid and cannabinoid CB1 and CB2 receptors in the spinal cord and the dorsal hind paw skin. In combination with the skin-nerve preparation, this model showed a decrease in functional receptive fields downstream to the injury and the apparition of A-fiber ectopic discharges. In conclusion, CCS injury induced behavioral, molecular, and electrophysiological rearrangements that might help us in better understanding the peripheral mechanisms of neuropathic pain. This model takes advantage of the possible use in the future of genetically modified mice and of an exclusively sensory nerve for a comprehensive study of peripheral mechanisms of neuropathic pain.

Action Potentials↗