Strength Of Lateral Inhibition Research Articles

Circuit and Cellular Mechanisms Facilitate the Transformation from Dense to Sparse Coding in the Insect Olfactory System.

Transformations between sensory representations are shaped by neural mechanisms at the cellular and the circuit level. In the insect olfactory system, the encoding of odor information undergoes a transition from a dense spatiotemporal population code in the antennal lobe to a sparse code in the mushroom body. However, the exact mechanisms shaping odor representations and their role in sensory processing are incompletely identified. Here, we investigate the transformation from dense to sparse odor representations in a spiking model of the insect olfactory system, focusing on two ubiquitous neural mechanisms: spike frequency adaptation at the cellular level and lateral inhibition at the circuit level. We find that cellular adaptation is essential for sparse representations in time (temporal sparseness), while lateral inhibition regulates sparseness in the neuronal space (population sparseness). The interplay of both mechanisms shapes spatiotemporal odor representations, which are optimized for the discrimination of odors during stimulus onset and offset. Response pattern correlation across different stimuli showed a nonmonotonic dependence on the strength of lateral inhibition with an optimum at intermediate levels, which is explained by two counteracting mechanisms. In addition, we find that odor identity is stored on a prolonged timescale in the adaptation levels but not in the spiking activity of the principal cells of the mushroom body, providing a testable hypothesis for the location of the so-called odor trace.

Synchrony, waves and ripple in spatially coupled Kuramoto oscillators with Mexican hat connectivity.

Spatiotemporal waves of synchronized activity are known to arise in oscillatory neural networks with lateral inhibitory coupling. How such patterns respond to dynamic changes in coupling strength is largely unexplored. The present study uses analysis and simulation to investigate the evolution of wave patterns when the strength of lateral inhibition is varied dynamically. Neural synchronization was modeled by a spatial ring of Kuramoto oscillators with Mexican hat lateral coupling. Broad bands of coexisting stable wave solutions were observed at all levels of inhibition. The stability of these waves was formally analyzed in both the infinite ring and the finite ring. The broad range of multi-stability predicted hysteresis in transitions between neighboring wave solutions when inhibition is slowly varied. Numerical simulation confirmed the predicted transitions when inhibition was ramped down from a high initial value. However, non-wave solutions emerged from the uniform solution when inhibition was ramped upward from zero. These solutions correspond to spatially periodic deviations of phase that we call ripple states. Numerical continuation showed that stable ripple states emerge from synchrony via a supercritical pitchfork bifurcation. The normal form of this bifurcation was derived analytically, and its predictions compared against the numerical results. Ripple states were also found to bifurcate from wave solutions, but these were locally unstable. Simulation also confirmed the existence of hysteresis and ripple states in two spatial dimensions. Our findings show that spatial synchronization patterns can remain structurally stable despite substantial changes in network connectivity.

Self-organized lateral inhibition improves odor classification in an olfaction-inspired network

The insect olfactory system is capable of classifying odorants by encoding and processing the neural representations of chemical stimuli. Odors are transformed into a neuronal representation by a number of receptor classes, each of which encodes a certain combination of chemical features. Those representations resemble a multivariate representation of the stimulus space [1]. The insect olfactory system thus provides an efficient basis for bio-inspired computational methods to process and classify multivariate data. Olfactory receptors typically have broad receptive fields, and the odor spectra of individual receptor classes overlap. From the viewpoint of multivariate data processing, overlapping receptive fields cause correlation between input variables (channel correlation). In previous work, we demonstrated how lateral inhibition in an olfaction-inspired network reduced channel correlation [2,3]. Decorrelation was achieved by setting the strength of lateral inhibition between two channels according to their correlation, which we pre-computed from the input data. Here, we propose unsupervised learning of the lateral inhibition structure. The lateral inhibition synapses support inhibitory spike-timing dependent plasticity (iSTDP) [4,5]. After exposing the network to a sufficient number of input samples, the inhibitory connectivity self-organizes to reflect the correlation between input channels. We show that this biologically realistic, local learning rule produces an inhibitory connectivity that effectively reduces channel correlation and yields superior network performance in a multivariate scent recognition scenario.

Regulation of Spike Timing-Dependent Plasticity of Olfactory Inputs in Mitral Cells in the Rat Olfactory Bulb

The recent history of activity input onto granule cells (GCs) in the main olfactory bulb can affect the strength of lateral inhibition, which functions to generate contrast enhancement. However, at the plasticity level, it is unknown whether and how the prior modification of lateral inhibition modulates the subsequent induction of long-lasting changes of the excitatory olfactory nerve (ON) inputs to mitral cells (MCs). Here we found that the repetitive stimulation of two distinct excitatory inputs to the GCs induced a persistent modification of lateral inhibition in MCs in opposing directions. This bidirectional modification of inhibitory inputs differentially regulated the subsequent synaptic plasticity of the excitatory ON inputs to the MCs, which was induced by the repetitive pairing of excitatory postsynaptic potentials (EPSPs) with postsynaptic bursts. The regulation of spike timing-dependent plasticity (STDP) was achieved by the regulation of the inter-spike-interval (ISI) of the postsynaptic bursts. This novel form of inhibition-dependent regulation of plasticity may contribute to the encoding or processing of olfactory information in the olfactory bulb.

Bio-inspired solutions to the challenges of chemical sensing

Bio-inspired solutions to the challenges of chemical sensing

The Operating Regime of Local Computations in Primary Visual Cortex

In V1, local circuitry depends on the position in the orientation map: close to pinwheel centers, recurrent inputs show variable orientation preferences; within iso-orientation domains, inputs are relatively uniformly tuned. Physiological properties such as cell's membrane potentials, spike outputs, and temporal characteristics change systematically with map location. We investigate in a firing rate and a Hodgkin-Huxley network model what constraints these tuning characteristics of V1 neurons impose on the cortical operating regime. Systematically varying the strength of both recurrent excitation and inhibition, we test a wide range of model classes and find the likely models to account for the experimental observations. We show that recent intracellular and extracellular recordings from cat V1 provide the strongest evidence for a regime where excitatory and inhibitory recurrent inputs are balanced and dominate the feed-forward input. Our results are robust against changes in model assumptions such as spatial extent and strength of lateral inhibition. Intriguingly, the most likely recurrent regime is in a region of parameter space where small changes have large effects on the network dynamics, and it is close to a regime of "runaway excitation," where the network shows strong self-sustained activity. This could make the cortical response particularly sensitive to modulation.

A model of contour extraction including multiple scales, flexible inhibition and attention

A mathematical model of contextual integration and contour extraction in the primary visual cortex developed in a recent work [Ursino, M., & La Cara, G. E. (2004). A model of contextual interactions and contour detection in primary visual cortex. Neural Networks, 17, 719-735] has been significantly improved to include two fundamental additional aspects, i.e., multi-scale decomposition and attention. The model incorporates two independent paths for visual processing corresponding to two different scales. Attention from higher hierarchical levels works by modifying different properties of the network: by selecting the portion of the image to be scrutinized and the appropriate scale, by modulating the threshold of a gating mechanism, and by modifying the width and/or strength of lateral inhibition. Through computer simulations of real complex and noisy black-and-white images, we demonstrate that appropriate selection of the above factors allows accurate analysis of image contours at different levels, from global perception of the overall objects without details, down to a fine examination of minute particulars (such as the lips in a face or the fingers of a hand). Attentive reconfiguration of lateral inhibition plays a key role in the analysis of images at different detail levels.

Dynamics of Winner-Take-All Competition in Recurrent Neural Networks With Lateral Inhibition

This paper studies the behavior of recurrent neural networks with lateral inhibition. Such network architecture is important in biological neural systems. General conditions determining the existence, number, and stability of network equilibria are derived. The manner in which these features depend upon steepness of neuronal activation functions and the strength of lateral inhibition is demonstrated for a broad range of nondecreasing activation functions including the discontinuous threshold function which represents the infinite gain limit. For uniform lateral inhibitory networks, the lateral inhibition is shown to sharpen neuron output patterns by increasing separation of suprathreshold activity levels of competing neurons. This results in the tendency of one neuron's output to dominate those of the others which can afford a "winner-take-all" (WTA) mechanism. Importantly, multiple stable equilibria may exist and shifts in inputs levels may yield network state transitions that exhibit hysteresis. A limitation of using lateral inhibition to implement WTA is further demonstrated. The possible significance of these identified network dynamics to physiology and pathophysiology of the striatum (particularly in Parkinsonian rest tremor) is discussed.

Propagation of action potentials from the soma to individual dendrite of cultured rat amacrine cells is regulated by local GABA input.

Retinal amacrine cells are interneurons that make lateral and vertical connections in the inner plexiform layer of the retina. Amacrine cells do not possess a long axon, and this morphological feature is the origin of their naming. Their dendrites function as both presynaptic and postsynaptic sites. Half of all amacrine cells are GABAergic inhibitory neurons that mediate lateral inhibition, and their light-evoked response consists of graded voltage changes and regenerative action potentials. There is evidence that the amount of neurotransmitter release from presynaptic sites is increased by spike propagation into the dendrite. Thus understanding of how action potentials propagate in dendrites is important to elucidating the extent and strength of lateral inhibition. In the present study, we used the dual whole cell patch-clamp technique on the soma and the dendrite of cultured rat amacrine cells and directly demonstrated that the action potentials propagate into the dendrites. The action potential in the dendrite was TTX sensitive and was affected by the local membrane potential of the dendrite. Propagation of the action potential was suppressed by local application of GABA to the dendrite. Dual dendrite whole cell patch-clamp recordings showed that GABA suppresses the propagation of action potentials in one dendrite of an amacrine cell, while the action potentials propagate in the other dendrites. It is likely that the action potentials in the dendrites are susceptible to various external factors resulting in the nonuniform propagation of the action potential from the soma of an amacrine cell.

Apomorphine modifies the visual responses of cells in the rabbit's lateral geniculate nucleus

It has been shown that enhancing or reducing dopaminergic activity in the retina modifies the balance between center and surround responses of retinal neurons such as ganglion cells. We investigated how these changes are reflected in the dorsal lateral geniculate nucleus (dLGN) by studying the effects of injections of apomorphine, a mixed D1 and D2 agonist of dopamine, on the visual responses of geniculate cells. Experiments were carried out on anesthetized adult pigmented rabbits. A varnished tungsten microelectrode was used to record single-unit activity in the dLGN. The flash electroretinogram was also recorded to monitor retinal changes and to confirm the success of the injections. Apomorphine was injected intravitreally or intravenously. The results can be summarized as follows. Apomorphine decreased the amplitude of the b-wave of the electroretinogram. For most dLGN cells, apomorphine produced a strong reduction in response amplitude evoked by sine-wave grating stimuli, presented at various spatial frequencies. Responses to flashing spots were also reduced but to a much lesser extent than those to gratings. In addition, the balance between the responses to small and large spots changed in favor of large stimuli. Consequently, after injection of apomorphine, the geniculate cells were preferentially activated by large-sized flashing stimuli. These data suggest that apomorphine can reduce the spatial contrast sensitivity of cells in the dLGN. This effect could be mediated by the reduction of the strength of lateral inhibition at the retinal level.

Modelling of Human Flicker Detection at Various Light Levels

Root-mean-square flicker sensitivity was measured within 0.2 – 2500 phot td and 0.5 – 30 Hz for a small spot with an equiluminous surround using computer graphics and a 2AFC method. In temporal noise, flicker sensitivity as a function of temporal frequency had a low-pass shape at all illuminances. Without temporal noise, the flicker sensitivity functions showed a band-pass shape at high illuminances, but changed to a low-pass shape at low illuminances. Our data are well described (goodness of fit 88%) by a model comprising: (i) low-pass temporal filtering by the modulation transfer function (MTF) of the photoreceptors ( R), (ii) high-pass temporal filtering by the MTF of the neural visual pathways ( P) resulting from lateral inhibition, (iii) addition of the temporal equivalent ( Nit) of internal neural noise, and (iv) detection by a temporal matched filter, provided that we take into account the fact that receptor responses become weaker and slower with decreasing illuminance despite adaptation. In our model detection efficiency was \eta( f)=0.148 f−0.568, Nit=46.3 μs and P( f) equal to f, where f is flicker frequency. In addition, R= R0[1+( f/ fc)2]−3, where R0=(1+24.1/ I)−0.5, fc=6.33 I0.172, and I is retinal illuminance. The model indicates that for the detection of a flickering spot in cone vision (i) the strength of lateral inhibition is independent of light level and (ii) quantal noise and dark light always remain insignificant sources of noise. (Mathematical expressions may not appear as intended)

An Increase in Strength of Tilt Aftereffect Associated with Tryptophan Depletion

The effects of a tryptophan-free amino acid mixture on tilt aftereffect, movement aftereffect, and the Mueller-Lyer illusion were studied. 12 male subjects ingested either a balanced amino acid mixture or a tryptophan-free mixture which causes a marked depletion of brain tryptophan and serotonin. The tryptophan-free mixture significantly increased the strength of tilt aftereffect but had no effect on movement aftereffect or the Mueller-Lyer illusion. These results were discussed with reference to the pharmacological activity of serotonin and its influence on the strength of lateral inhibition in mammalian brains.

Disappearance of grating induction at scotopic luminances

The dependence of grating induction magnitude on retinal illuminance was examined in two subjects. Grating induction magnitude, as determined using the cancellation technique of McCourt ( Vision Research, 22, 119–134, 1982), declines monotonically with decreasing retinal illuminance, effectively disappearing at a value of 0.3–0.5 phot td. In a second experiment, sensitivity differences for test lights of 500 and 600 nm were measured as a function of background illuminance in order to gauge the luminance operating range for grating induction with respect to duplex photoreceptor function. Cancelling contrast (and hence grating induction magnitude) fell below detection threshold contrast at retinal illuminances coinciding with the transition from photopic to scotopic visual function. In a third experiment, spatial contrast sensitivity was measured using both spatially extended (10°) and truncated (2°) sinewave gratings at frequencies below 2 c/deg, at three values of retinal illuminance. Illuminance values corresponded to those where grating induction magnitude was, as determined from the first experiment, either maximal, intermediate or negligible. Similar to grating induction, the strength of lateral inhibition, as indexed by the slope of the low-frequency decline in contrast sensitivity, is progressively reduced with decreasing retinal illuminance, particularly for the 2° field. There was, however, using the same criteria, evidence of lateral inhibition at a value of retinal illuminance which did not support grating induction. The implications of these results are discussed with respect to classical brightness contrast phenomena, recent neuroanatomical and neurophysiological evidence of segregated parvo- and magnocellular mediated contrast processing systems, and with results from previous studies of the grating induction effect.

AN INCREASE IN STRENGTH OF TILT AFTEREFFECT ASSOCIATED WITH TRYPTOPHAN DEPLETION

The effects of a tryptophan-free amino acid mixture on tilt aftereffect, movement aftereffect, and the Mueller-Lyer illusion were studied. 12 male subjects ingested either a balanced amino acid mixture or a tryptophan-free mixture which causes a marked depletion of brain tryptophan and serotonin. The tryptophan-free mixture significantly increased the strength of tilt aftereffect but had no effect on movement aftereffect or the Mueller-Lyer illusion. These results were discussed with reference to the pharmacological activity of serotonin and its influence on the strength of lateral inhibition in mammalian brains.

UNSUPERVISED COMPETITIVE LEARNING IN NEURAL NETWORKS

We study an idealized neural network that may approximate certain neurophysiological features of natural neural systems. The network contains a mutual lateral inhibition and is subjected to unsupervised learning by means of a Hebb-type learning principle. Its learning ability is analysed as a function of the strength of lateral inhibition and the training set.