Formation of receptive fields in realistic visual environments according to the Bienenstock, Cooper, and Munro (BCM) theory.

Abstract

The Bienenstock, Cooper, and Munro (BCM) theory of synaptic plasticity has successfully reproduced the development of orientation selectivity and ocular dominance in kitten visual cortex in normal, as well as deprived, visual environments. To better compare the consequences of this theory with experiment, previous abstractions of the visual environment are replaced in this work by real visual images with retinal processing. The visual environment is represented by 24 gray-scale natural images that are shifted across retinal fields. In this environment, the BCM neuron develops receptive fields similar to the fields of simple cells found in kitten striate cortex. These fields display adjacent excitatory and inhibitory bands when tested with spot stimuli, orientation selectivity when tested with bar stimuli, and spatial-frequency selectivity when tested with sinusoidal gratings. In addition, their development in various deprived visual environments agrees with experimental results.In 1982 Bienenstock, Cooper, and Munro (BCM) proposed a concrete synaptic-modification hypothesis in which two regions of modification (Hebbian and anti-Hebbian) are stabilized by the addition of a sliding modification threshold (1). The theory was created originally to explain the development of orientation selectivity of visual cortical neurons in various visual environments. This theory has since proven capable of also explaining ocular dominance and selective response properties in the most diverse visual environments, one ofthe most thoroughly studied areas in neuroscience. In this paper we examine the consequences ofthe replacement ofprevious abstractions of the visual environment by real visual images with retinal processing.During a critical period of postnatal development, the response properties of neurons in striate cortex of the kitten can be modified by manipulating the visual experience of the animal (2-4). Clothiaux, Bear, and Cooper (CBC) (5) showed that simulations based on the BCM theory, with a fixed set of parameters, reproduce both the kinetics and equilibrium states of experience-dependent modifications that are observed experimentally in these neurons. The rearing conditions that they simulated include normal rearing, monocular deprivation, reverse suture, strabismus, and binocular deprivation.An important simplification used in all previous BCM simulations is the assumed activity of neurons in the lateral geniculate nucleus (LGN) resulting from visual experience. In CBC, for example, an abstract set of 12 patterns represents the activity on geniculo-cortical afferents supposedly resulting from contoured stimuli with 12 different orientations.LGN activity in a normal visual environment is modeled as this patterned input distorted by noise. In the absence of visual contours, LGN-cortical input activity is assumed to be uncorrelated noise distributed around spontaneous activity. Thejustification for this simplification is that the visual fields of neurons in primary visual cortex are small, so that reproducibl...

Experience‐dependent plasticity of visual acuity in rats

“…The demonstration that monocular deprivation produces a larger ocular dominance shift than equivalent monocular inactivation supports the basic tenants of this theory (Rittenhouse et al, 1999). Computational models of the Bienenstock et al (1982) rule (Law & Cooper, 1994;Blais et al, 1999) applied to the developing visual cortex predict that the temporal patterning of normal binocular activity is capable of maintaining a high threshold for synaptic modi®cation in each eye's inputs, thus enabling the development of shared cortical ocular dominance and the formation of receptive ®elds with orientation and spatial frequency selectivity. The unpatterned activity generated by the deprived eye during monocular deprivation would not be suf®cient to maintain the threshold at a high value, resulting in rapid active depression of its synapses.…”

Natural scene statistics at the centre of gaze

“…Thus it has been proposed that sensory systems have specifically evolved to encode their natural stimuli efficiently, by taking advantage of stimulus invariances [5,6]. There is recent evidence for this hypothesis in the computations performed by the first stages of visual systems [7][8][9][10][11][12][13][14][15][16][17][18][19].…”

Contrast normalization contributes to a biologically-plausible model of receptive-field development in primary visual cortex (V1)

“…Law and Cooper (1994) showed that the BCM rule, for a single neuron trained on natural images, produces a model receptive field which has many similarities to those of real simple cells. However, in its simplest form, the BCM rule provides no way for neurons to communicate with each other in order to form a meaningful coding population .…”