Contralateral Receptive Fields Research Articles

Analysis of convolutional neural networks reveals the computational properties essential for subcortical processing of facial expression

Perception of facial expression is crucial for primate social interactions. This visual information is processed through the ventral cortical pathway and the subcortical pathway. However, the subcortical pathway exhibits inaccurate processing, and the responsible architectural and physiological properties remain unclear. To investigate this, we constructed and examined convolutional neural networks with three key properties of the subcortical pathway: a shallow layer architecture, concentric receptive fields at the initial processing stage, and a greater degree of spatial pooling. These neural networks achieved modest accuracy in classifying facial expressions. By replacing these properties, individually or in combination, with corresponding cortical features, performance gradually improved. Similar to amygdala neurons, some units in the final processing layer exhibited sensitivity to retina-based spatial frequencies (SFs), while others were sensitive to object-based SFs. Replacement of any of these properties affected the coordinates of the SF encoding. Therefore, all three properties limit the accuracy of facial expression information and are essential for determining the SF representation coordinate. These findings characterize the role of the subcortical computational processes in facial expression recognition.

Finger Vein Image Enhancement Based on Guided Tri-Gaussian Filters

In the process of image acquisition, the contrast between veins and non-veins in finger vein images is not high due to the influence of the fuzzy light source, skin scattering and finger movement. To solve this problem, a finger vein image enhancement method is proposed (GTGFs), which enhances finger vein patterns by setting guided image as input image firstly. On this basis, the tri-Gaussian model is based on disinhibitory properties of the concentric receptive field used to locally enhancing the image. The parameters of the tri-Gaussian model are determined based on the finger vein width information. The experiment results show that the proposed enhancement method can significantly enhance the finger vein patterns and improve the recognition effect of the methods based on vein pattern segmentation.

Explore double-opponency and skin color for saliency detection

Recent advances in salient detection have exploited the foreground or background information to assist other saliency cues such as contrast to achieve state-of-the-art results. However the problem remains challenging. For example, human skin color is easily overlooked during saliency detection. Simulating the human visual mechanism to improve the current algorithm, we propose a saliency model based on the color-opponent mechanisms of a certain type of color-sensitive double-opponent (DO) cells in the primary visual cortex (V1) of human visual system. Firstly, DO cells with concentric receptive fields (RFs) can detect region contrast for yielding foreground saliency map. Our approach is intuitive and easy to interpret, and it allows fast implementation. Then, skin saliency map is built in a way that is different from high-level factors such as face detection, combining skin region and spatial Euclidean distance weight in the RGB space, and the significant skin features can be obtained effectively. Finally, a linear fusion strategy is proposed to integrate different saliency maps. Experimental results with three well-known benchmark databases demonstrate that the proposed method can achieve competitive performance when compared to state-of-the-art methods. Saliency regions contain important skin features compared to other traditional methods.

The multifunctional lateral geniculate nucleus.

Providing the critical link between the retina and visual cortex, the well-studied lateral geniculate nucleus (LGN) has stood out as a structure in search of a function exceeding the mundane 'relay'. For many mammals, it is structurally impressive: Exquisite lamination, sophisticated microcircuits, and blending of multiple inputs suggest some fundamental transform. This impression is bolstered by the fact that numerically, the retina accounts for a small fraction of its input. Despite such promise, the extent to which an LGN neuron separates itself from its retinal brethren has proven difficult to appreciate. Here, I argue that whereas retinogeniculate coupling is strong, what occurs in the LGN is judicious pruning of a retinal drive by nonretinal inputs. These nonretinal inputs reshape a receptive field that under the right conditions departs significantly from its retinal drive, even if transiently. I first review design features of the LGN and follow with evidence for 10 putative functions. Only two of these tend to surface in textbooks: parsing retinal axons by eye and functional group and gating by state. Among the remaining putative functions, implementation of the principle of graceful degradation and temporal decorrelation are at least as interesting but much less promoted. The retina solves formidable problems imposed by physics to yield multiple efficient and sensitive representations of the world. The LGN applies context, increasing content, and gates several of these representations. Even if the basic concentric receptive field remains, information transmitted for each LGN spike relative to each retinal spike is measurably increased.

Vector model for mapping of visual space to subjective 4-D sphere

Here we present a mathematical model of binocular vision that maps a visible physical world to a subjective perception of it. The subjective space is a set of 4-D vectors whose components are outputs of four monocular neurons from each of the two eyes. Monocular neurons have one of the four types of concentric receptive fields with Gabor-like weighting coefficients. Next this vector representation of binocular vision is implemented as a pool of neurons where each of them is selective to the object's particular location in a 3-D visual space. Formally each point of the visual space is being projected onto a 4-D sphere. Proposed model allows determination of subjective distances in depth and direction, provides computational means for determination of Panum's area and explains diplopia and allelotropia.

A synaptic signature for ON- and OFF-center parasol ganglion cells of the primate retina.

In the primate retina, parasol ganglion cells contribute to the primary visual pathway via the magnocellular division of the lateral geniculate nucleus, display ON and OFF concentric receptive field structure, nonlinear spatial summation, and high achromatic temporal-contrast sensitivity. Parasol cells may be homologous to the alpha-Y cells of nonprimate mammals where evidence suggests that N-methyl-D-aspartate (NMDA) receptor-mediated synaptic excitation as well as glycinergic disinhibition play critical roles in contrast sensitivity, acting asymmetrically in OFF- but not ON-pathways. Here, light-evoked synaptic currents were recorded in the macaque monkey retina in vitro to examine the circuitry underlying parasol cell receptive field properties. Synaptic excitation in both ON and OFF types was mediated by NMDA as well as α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate glutamate receptors. The NMDA-mediated current-voltage relationship suggested high Mg2+ affinity such that at physiological potentials, NMDA receptors contributed ∼20% of the total excitatory conductance evoked by moderate stimulus contrasts and temporal frequencies. Postsynaptic inhibition in both ON and OFF cells was dominated by a large glycinergic "crossover" conductance, with a relatively small contribution from GABAergic feedforward inhibition. However, crossover inhibition was largely rectified, greatly diminished at low stimulus contrasts, and did not contribute, via disinhibition, to contrast sensitivity. In addition, attenuation of GABAergic and glycinergic synaptic inhibition left center-surround and Y-type receptive field structure and high temporal sensitivity fundamentally intact and clearly derived from modulation of excitatory bipolar cell output. Thus, the characteristic spatial and temporal-contrast sensitivity of the primate parasol cell arises presynaptically and is governed primarily by modulation of the large AMPA/kainate receptor-mediated excitatory conductance. Moreover, the negative feedback responsible for the receptive field surround must derive from a nonGABAergic mechanism.

A Mathematical Model of Retinal Ganglion Cells and Its Applications in Image Representation

Contrast stimuli are the essential components that constitute visual scenes. This paper studies how retinal ganglion cells respond to a contrast stimulus covering their concentric receptive fields with center-surround antagonism, and thereby proposes a mathematical model that describes the response by the multiplication of the contrast of the stimulus and a normalized response function with respect to the coverage ratio and the center/surround ratio. The obtained response curves turn out to be consistent with physiological data. This model partially accounts for the contrast sensitivities of ganglion cells and the invariance of visual perception to the contrast. A computational approach based on this neural model is developed for orientation detection, and is further applied to image representation. The experiments achieve convincing results on challenging image datasets. Moreover, it is revealed that the produced orientation maps remarkably enhance the efficiencies and the effectiveness of segmentation algorithms.

Three Forms of Spatial Temporal Feedforward Inhibition Are Common to Different Ganglion Cell Types in Rabbit Retina

There exist more than 30 different morphological amacrine cell types, but there may be fewer physiological types. Here we studied the amacrine cell outputs by measuring the temporal and spatial properties of feedforward inhibition to four different types of ganglion cells. These ganglion cells, each with concentric receptive field organization, appear to receive a different relative contribution of the same three forms of feed-forward inhibition, namely: local glycinergic, local sustained GABAergic, and broad transient GABAergic inhibition. Two of these inhibitory components, local glycinergic inhibition and local sustained GABAergic inhibition were localized to narrow regions confined to the dendritic fields of the ganglion cells. The third, a broad transient GABAergic inhibition, was driven from regions peripheral to the dendritic area. Each inhibitory component is also correlated with characteristic kinetics expressed in all ganglion cells: broad transient GABAergic inhibition had the shortest latency, local glycinergic inhibition had an intermediate latency, and local sustained GABAergic inhibition had the longest latency. We suggest each of these three inhibitory components represents the output from a distinct class of amacrine cell, mediates a specific visual function, and each forms a basic functional component for the four ganglion cell types. Similar subunits likely exist in the circuits of other ganglion cell types as well.

Function of the appendages in tentacled snakes (Erpeton tentaculatus)

We investigated the function of the tentacles in aquatic, piscivorous tentacled snakes (Erpeton tentaculatus) by examining anatomy, peripheral innervation, and the response properties of primary afferents. We also investigated visual and somatosensory responses in the optic tectum and documented predatory strikes to visual stimuli and under infrared illumination. Our results show the tentacles are sensitive mechanoreceptors that respond to water movements. They are innervated by rami of the maxillary and ophthalmic branches of the trigeminal nerve and contain a dense array of fine terminal neurites that cross the interior of the tentacle orthogonal to its long axis. The optic tectum contained a retinotopic map of contralateral receptive fields with superior fields represented dorsally in the tectum, inferior fields represented laterally, nasal fields represented rostrally, and temporal fields represented caudally. Large somatosensory receptive fields were identified in deeper layers of the tectum and were in approximate register with overlying visual fields. Tentacled snakes struck accurately at a simulated digital fish, indicating that visual cues are sufficient to guide strikes, but they also captured fish under infrared illumination, suggesting water movements alone could be used to localize prey. We conclude the tentacles are mechanosensors that are used to detect fish position based on water movements and that visual and mechanosensory cues may be integrated in the tectum to enhance localization when visual cues are reduced.

Organization of Concentric Double-Opponent Receptive Fields in the Primary Visual Cortex by Using Independent Component Analysis

Event Abstract Back to Event Organization of Concentric Double-Opponent Receptive Fields in the Primary Visual Cortex by Using Independent Component Analysis Yasunobu Igarashi1*, Takaaki Aoki2, Rui Kimura3 and Shin Ishii4 1 National Center for Neurology and Psychiatry, Japan 2 Nara Institute of Science and Technology, Japan 3 RIKEN Brain Science Institute, Japan 4 Kyoto University, Japan Previous experimental studies showed that neurons in the primary visual cort ex (V1) have four firing properties to represent visual information. Four fi ring properties are named luminance gabor, oriented double-opponent, concent ric single-opponent, and concentric double-opponent receptive fields. Previo us theoretical studies demonstrated that redundancy compressions like indepe ndent component analysis (ICA) and sparse coding can derive basic functions corresponding to luminance gabor and oriented double-opponent receptive fiel ds from natural images. On the other hand, emergences of basic functions cor responding to two concentric receptive fields via redundancy compression of natural images have not been reported. In this study, we propose a low pass filtering function of neurons in lateral geniculate nucleus (LGN) is importa nt to derive concentric basic functions in addition to redundancy compressio n. Neurons in LGN project their axons to neurons in V1. Therefore, it is pla usible that LGN contributes to form receptive fields of V1. We show that the concentric double-opponent basis function can be obtained by applying the l ow pass filtering and Fast ICA to natural images in a computer simulation. H owever, the concentric single-opponent basis function cannot be not obtained . This reason is thought that the concentric double-opponent basis function functionally includes the concentric single-opponent basis function from the view point of redundancy compression. We also clarify how shapes of basic f unctions depend on ones of learning images, and how color components of basi c functions depend on a color distribution of learning images. Conference: Neuroinformatics 2010 , Kobe, Japan, 30 Aug - 1 Sep, 2010. Presentation Type: Poster Presentation Topic: Computational Neuroscience Citation: Igarashi Y, Aoki T, Kimura R and Ishii S (2010). Organization of Concentric Double-Opponent Receptive Fields in the Primary Visual Cortex by Using Independent Component Analysis. Front. Neurosci. Conference Abstract: Neuroinformatics 2010 . doi: 10.3389/conf.fnins.2010.13.00055 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 11 Jun 2010; Published Online: 11 Jun 2010. * Correspondence: Yasunobu Igarashi, National Center for Neurology and Psychiatry, Tokyo, Japan, igalab1@j-adni.org Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Yasunobu Igarashi Takaaki Aoki Rui Kimura Shin Ishii Google Yasunobu Igarashi Takaaki Aoki Rui Kimura Shin Ishii Google Scholar Yasunobu Igarashi Takaaki Aoki Rui Kimura Shin Ishii PubMed Yasunobu Igarashi Takaaki Aoki Rui Kimura Shin Ishii Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

Constraint on the Number of Synaptic Inputs to a Visual Cortical Neuron Controls Receptive Field Formation

To date, Hebbian learning combined with some form of constraint on synaptic inputs has been demonstrated to describe well the development of neural networks. The previous models revealed mathematically the importance of synaptic constraints to reproduce orientation selectivity in the visual cortical neurons, but biological mechanisms underlying such constraints remain unclear. In this study, we addressed this issue by formulating a synaptic constraint based on activity-dependent mechanisms of synaptic changes. Particularly, considering metabotropic glutamate receptor-mediated long-term depression, we derived synaptic constraint that suppresses the number of inputs from individual presynaptic neurons. We performed computer simulations of the activity-dependent self-organization of geniculocortical inputs with the synaptic constraint and examined the formation of receptive fields (RFs) of model visual cortical neurons. When we changed the magnitude of the synaptic constraint, we found the emergence of distinct RF structures such as concentric RFs, simple-cell-like RFs, and double-oriented RFs and also a gradual transition between spatiotemporal separable and inseparable RFs. Thus, the model based on the synaptic constraint derived from biological consideration can account systematically for the repertoire of RF structures observed in the primary visual cortices of different species for the first time.

A comparison of the response properties of periodontal mechanosensitive neurons in each relay nucleus from the trigeminal ganglion to the somatosensory cortex in the cat and rat

This is a review of the response properties of periodontal mechanosensitive (PM) neurons in each relay nucleus of the ascending pathway from the primary afferents (the trigeminal ganglion) to the primary somatosensory cortex in the cat and rat. The major characteristics are as follows: the great majority of cat and rat primary PM afferents adapt slowly to mechanical stimulation of the tooth and are sensitive to one tooth only (single-tooth neurons). The threshold intensities to tooth stimulation are less than 0.01 N. The incidence of slowly adapting and single-tooth neurons decreases and threshold intensities become higher, as PM information is relayed to higher relay nuclei of the brain. The receptive fields of PM neurons in the trigeminal ganglion and the trigeminal sensory complex are ipsilateral. However, the occurrence of thalamic PM neurons with contralateral receptive fields increases to about 50% in the cat and 100% in the rat. The difference in the projection side of thalamic neurons between the cat and rat might be related to species differences and/or specialization of mastication in each animal.

“Small-tufted” ganglion cells and two visual systems for the detection of object motion in rabbit retina

Small-tufted (ST) ganglion cells of rabbit retina are divided into eight types based upon morphology, branching pattern, level of dendritic stratification, and quantitative dimensional analysis. Only one of these types has been previously characterized in Golgi preparations, and four may be discerned in the work of others. Given their small dendritic-field size, and assuming uniform mosaics of each across the retina, ST cells comprise about 45% of all rabbit ganglion cells, and are therefore of major functional significance. Four ST cells occur as two paramorphic (a/b) pairs, and thus belong to class III, as previously defined. Four branch in sublaminae a and b of the inner plexiform layer (IPL) and therefore belong to class IV. ST cells have small cell bodies 10-15 microm in diameter, small axons 0.7-1.3 microm in diameter, and small dendritic-field diameters, 40-110 microm in mid-visual streak. The dendrites of ST cells are highly branched, and bear spines and appendages of varying length, but vary from type to type. Class III.2 cells and class III.3 cells are partly bistratified. Class IV small-tufted cells differ characteristically in multiple features of dendritic branching and stratification. Class III small-tufted cells apparently have concentric (ON-center and OFF-center) receptive fields and may have "sluggish-transient" (class III.2) and "sluggish-sustained" (class III.3) physiology. Class IV cells include the "local-edge-detector" (LED) (class IVst1), and are all expected to give ON-OFF responses to small, centered, slowly moving visual stimuli. Based upon systematic variation in dendritic-field size across the retina, ST cells may be divided into two groups. In this "universal prey" species, they may belong to two systems of motion detection, typified by ON-OFF directionally selective and LED ganglion cells, respectively, specialized for detection of rapid motion at the horizon for land-based predators, and slow motion for airborne predators.

Intermanual and intramanual tactual grating discrimination

For intermanual tactual discrimination to occur, it is thought that neural signals have to cross the corpus callosum in some way at least once. In this note we address the question of whether this interhemispheric transfer affects grating discrimination in active touch. Eight right-handed participants made intermanual and intramanual tactual discriminations of sinusoidal gratings that had slightly different spatial periods. Intramanual discrimination included comparisons in which the gratings were felt with the same finger, as well as comparisons made with two different fingers of the same hand. During intermanual discrimination the gratings were felt with corresponding fingers or with non-corresponding fingers of different hands. We found that thresholds for intramanual conditions were lower than for intermanual conditions in active dynamic touch. This suggests that there is a component of the task that is unilateral, as might be mediated by a somatosensory region that has predominantly or exclusively contralateral receptive fields.

Pattern detection: interactions between oriented and concentric patterns

Target threshold vs. pedestal contrast (TvC) functions for a vertical and a concentric target were measured on a vertical, a horizontal, a plaid or a concentric Gabor pedestal. All patterns had Guassian envelopes. All except one of the TvC functions had a dipper shape, that is, the target threshold first decreased and then increased with the pedestal contrast. The TvC function for a vertical target on a horizontal pedestal was monotonically increasing. A divisive inhibition model with two orientation selective mechanisms, sensitive to vertical and horizontal, fits the data reasonably well. A significant improvement in fit is obtained by adding one or two more mechanisms. These additional mechanisms are different for different observers. Some are consistent with oblique receptive fields, some with concentric receptive fields, and some are indeterminate.

The influence of stimulus properties on multisensory processing in the awake primate superior colliculus.

Multisensory integration is a process whereby information converges from different sensory modalities to produce a response that is different from that elicited by the individual modalities presented alone. A neural basis for multisensory integration has been identified within a variety of brain regions, but the most thoroughly examined model has been that of the superior colliculus (SC). Multisensory processing in the SC of anaesthetized animals has been shown to be dependent on the physical parameters of the individual stimuli presented (e.g., intensity, direction, velocity) as well as their spatial relationship. However, it is unknown whether these stimulus features are important, or evident, in the awake behaving animal. To address this question, we evaluated the influence of physical properties of sensory stimuli (visual intensity, direction, and velocity; auditory intensity and location) on sensory activity and multisensory integration of SC neurons in awake, behaving primates. Monkeys were trained to fixate a central visual fixation point while visual and/or auditory stimuli were presented in the periphery. Visual stimuli were always presented within the contralateral receptive field of the neuron whereas auditory stimuli were presented at either ipsi- or contralateral locations. Many of the SC neurons responsive to these sensory stimuli (n = 66/84; 76%) had stronger responses when the visual and auditory stimuli were combined at contralateral locations than when the auditory stimulus was located on the ipsilateral side. This trend was significant across the population of auditory-responsive neurons. In addition, some SC neurons (n = 31) were presented a battery of tests in which the quality of one stimulus of a pair was systematically manipulated. A small proportion of these neurons (n = 8/31; 26%) showed preferential responses to stimuli with specific physical properties, and these preferences were not significantly altered when multisensory stimulus combinations were presented. These data demonstrate that multisensory processing in the awake behaving primate is influenced by the spatial congruency of the stimuli as well as their individual physical properties.

Bilateral receptive field neurons in the hindlimb region of the postcentral somatosensory cortex in awake macaque monkeys.

Single-neuron activities were recorded in the hindlimb region of the primary somatosensory cortex and part of area 5 in awake Japanese monkeys. A total of 1050 units were isolated from five hemispheres of four animals. Receptive fields (RFs) and submodalities were identified for 90% of isolated neurons in areas 3a and 3b. The percentage decreased as the recoding site moved to the more caudal areas. Deep or skin submodality neurons were dominant in area 3a or area 3b, respectively. Deep submodality neurons increased in more caudal areas and were the majority in areas 2 and 5. These observations were consistent with those in the hand and/or digit or arm and/or trunk region. The identified neurons were classified by their RF positions into four types: the foot, leg, foot and leg, or hindlimb and other body parts type. Among 831 identified neurons, 33 neurons had bilateral RFs, 14 had ipsilateral RFs, and the rest (N=784) had contralateral RFs. The relative incidence of neurons with bilateral or ipsilateral RFs among identified neurons was less than 1% in areas 3a, 3b, and 1, and 16% or 25% in areas 2 or 5, respectively. Within areas 2 and 5, the percentage of neurons with bilateral or ipsilateral RFs was significantly smaller in the foot type (5%) than in other RF types (24-57%). RFs of the foot type were on the sole or single toe but never on multiple toes. These observations contrasted with the previous findings that neurons with bilateral RFs were more frequently seen in the hand and/or digit region and that RFs on multiple digit tips were dominant there. The present study thus demonstrated that neurons with bilateral RFs do exist in the hindlimb region. Similarly to the forelimb region, they were found mostly in areas 2 and 5, the caudalmost areas of the postcentral gyrus and hierarchically higher stages in information processing. The relative paucity of neurons with bilateral RFs on the foot, especially those with RFs on multiple toes, may reflect functional differences between the foot and the hand.

Response properties of slowly and rapidly adapting periodontal mechanosensitive neurones in the primary somatosensory cortex of the cat

Periodontal mechanosensitive neurones in the primary somatosensory (SI) cortex are classified as either slowly or rapidly adapting. The responses of cortical neurones and their projection pathways were studied using mechanical and electrical stimulation of the teeth and electrical stimulation of the thalamic posteromedial ventral (VPM) nuclei and contralateral SI cortex. A total of 247 periodontal mechanosensitive units were recorded from the SI cortex in 35 anaesthetized cats, distributed mainly in area 3b: 14% were slowly adapting and 86% rapidly adapting units; 62% of the slowly adapting and 9% of the rapidly adapting units were single-tooth units sensitive to stimulation of only one tooth. The incidence of slowly adapting units with an ipsilateral receptive field was almost equal to that of slowly adapting units with a contralateral receptive field, and more than half of the units were directionally selective to mechanical tooth stimulation. The majority of rapidly adapting units had their receptive field in the contralateral teeth and were directionally non-selective to tooth stimulation. The latencies of the cortical responses of the slowly and rapidly adapting units were 7.3 and 10.7 ms, respectively, on electrical stimulation of the contralateral teeth, and 1.8 and 2.0 ms, respectively, on electrical stimulation of the ipsilateral VPM nucleus. From these findings, it is inferred that slowly adapting neurones are useful for discriminating the tooth stimulated, the stimulus direction, the stimulus intensity and the change of pressure applied to the tooth, while rapidly adapting neurones could function to signal initial contact with food or the opposing teeth.

Representation of the midline trunk, bilateral arms, and shoulders in the monkey postcentral somatosensory cortex.

Single neuronal activities were recorded in the arm/trunk region of the postcentral gyrus in awake Japanese monkeys. A total of 1608 units were isolated from four hemispheres of two animals, and receptive fields (RFs) and submodalities were identified in 1162 units. Deep or skin submodality neurons were dominant in area 3a or area 3b, respectively. The deep/skin ratio increased as the recording site moved from area 3b to the more caudal areas. In areas 3a and 3b, neuronal RFs were almost exclusively on either the arm or trunk. In areas 2 and 5, neurons with RFs on the trunk decreased and those with RFs on the hand or covering more than one body part, etc. increased. We found a total of 107 neurons with bilateral RFs and 56 with ipsilateral RFs, while the rest (n=999) were with contralateral RFs. Bilateral or ipsilateral neurons of skin submodality (n=37) were found in areas 1, 2, and 5. Twenty six (70%) had RFs on the trunk and/or occiput, five on the forelimb, and the rest (n=6) on both the trunk and forelimb (the combined type). Among 33 skin bilateral neurons, 90% (n=30) had RFs across the midline. Bilateral or ipsilateral neurons responding to joint manipulation (n=104) were found in areas 2 and 5. Most of them were activated by manipulation of the shoulder and/or elbow (the proximal type, n=72, 69%). There were 25 neurons of the combined type (both the proximal and distal joints were effective, 24%). Bilateral or ipsilateral neurons of deep-others submodality (n=20) were found in areas 1, 2, and 5. The forelimb type (n=12, 60%) was dominant in this category. The combined-type neurons in both the skin- and joint-manipulation categories were found only or mostly in area 5. These results indicate the presence of hierarchical processing for bilateral as well as contralateral information within the arm/trunk region of the postcentral gyrus.

Illusions as neuro-signs

One of the reasons visual scientists find illusions fascinating is that their pictorial simplicity belies the complexity of visual processing. Illusions can direct us towards neurophysiological interpretations at many levels, which I have called ‘neuro-signs’. The most useful visual illusions are those that point to the simplest neural level possible, that is, those that provide signs of neural processing at early stages of the visual system, such as in the retina itself. The region of the retina that can change the activity of nerve cells throughout the visual system when light falls on it is called the ‘receptive field’ of those nerve cells. Primate retinal ganglion cells have receptive fields that are concentric and antagonistic — if the centre of a receptive field is excited by light the surround is inhibited (called ‘surround inhibition’), or vice versa. Are there any visual signs of such neural antagonism? There could be. In the black and white grid pattern shown below, a variety of illusory dots can be seen at the intersections of the black and white grid. In the upper half, dark grey dots appear at the white intersections between the black rectangles, although a light meter would not register such effects. This is called a Hermann grid; its reverse in the lower half, where illusory light grey dots are seen in the black intersections, is called a Hering grid. In both cases, the dots at an intersection disappear as soon as it is fixated, but they remain at those nearby. The dots are more difficult to see at the smallest intersections in both grids. Because of this effect, it is thought the illusory dots might be related to the distance from the centre of the retina (or ‘retinal eccentricity’); receptive fields have been found to be much smaller around the centre of the retina than more peripherally. Indeed, the optimal separation for producing the illusory dots increases with retinal eccentricity in a similar manner to the increase in receptive field size. But why do the illusory dots appear at all? One suggestion is that concentric receptive fields of the same dimensions will be stimulated differently when they fall at an intersection between four rectangles than when flanked by two rectangles — there would be more surround inhibition at an intersection than elsewhere. This would make a light intersection darker than its neighbours and a dark intersection lighter. Such an interpretation does not account for the dots that are visible between rectangles that are simply outlines, although these dots are not as compelling. When coloured rectangles are placed on a grey background (see above), the illusory dots at the intersections take on the colour of the rectangles: they appear to be light green in the grey intersections of the upper half, and reddish in the grey intersections of the lower half. This is because some receptive fields at the retinal ganglion cell level display colour opponency — the centre can be excited by, say, red and the surround inhibited by green; others operate on blue–yellow opponency. Hermann–Hering grids provide signs of a link between visual perception and neurophysiology. The link is made at an early stage of visual processing — the differences in the activity of the retinal ganglion cells — with the assumption that the neural signals for these differences are preserved in their transmission to the visual cortex of the brain. N Wade, Department of Psychology, University of Dundee, Dundee DD1 4HN, UK. This is the second in a short series of articles by Nicholas Wade, describing some common visual illusions.