Reaching beyond the classical receptive field of V1 neurons: horizontal or feedback axons?

https://doi.org/10.1016/j.jphysparis.2003.09.001Get rights and content

Abstract

It is commonly assumed that the orientation-selective surround field of neurons in primary visual cortex (V1) is due to interactions provided solely by intrinsic long-range horizontal connections. We review evidence for and against this proposition and conclude that horizontal connections are too slow and cover too little visual field to subserve all the functions of suppressive surrounds of V1 neurons in the macaque monkey. We show that the extent of visual space covered by horizontal connections corresponds to the region of low contrast summation of the receptive field center mechanism. This region encompasses the classically defined receptive field center and the proximal surround. Beyond this region, feedback connections are the most likely substrate for surround suppression. We present evidence that inactivation of higher order areas leads to a major decrease in the strength of the suppressive surround of neurons in lower order areas, supporting the hypothesis that feedback connections play a major role in center–surround interactions.

Introduction

The traditional, feedforward, model of the visual system invokes a cascade of processing stages, beginning with relay of retinal input to the neurons of area V1, via the lateral geniculate nucleus (LGN), and subsequent processing through a hierarchy of cortical areas [102]. According to this model, cells at each successive stage process inputs from increasingly larger regions of space, and code for increasingly more complex aspects of visual stimuli, a complex abstraction of the visual world achieved by cells at the highest visual centers. The selectivity of a neuron to a given parameter1 is supposed to result from the ordered convergence of afferents from the lower stages. One example is the emergence of orientation selectivity in area V1 neurons. Following the initial model of Hubel and Wiesel [39], several authors have claimed that orientation selectivity can be entirely generated by the ordered arrangement of afferents from the LGN. This view has been subject to debate for several decades (for reviews see [26], [94]). In particular, there is evidence that local intracortical excitatory and inhibitory mechanisms play a role in amplifying and sharpening the orientation tuning of V1 neurons.

Although feedforward models can perform a surprising number of object recognition tasks in a simple environment [98], they fail to identify an object in a cluttered environment, where the object can be partially masked or occluded by other objects. Proper interpretation of partially occluded figures requires global information about the 3D interpretation of the scene to guide the alignment of edges. The computations related to such global-to-local interactions should be flexible, i.e. different sets of global cues should lead to different types of local processing of edges. Furthermore, to mediate perceptual completion across several degrees of visual angle, these computations should involve exchange of information across distant regions of the visual field (often >10°). Finally, to allow for interpretation of an image within the timeframe of inter-saccadic times (200 ms), they should be fast.

There is compounding evidence that flexible and rapid long-distance computations do occur in the visual system, even at the lower levels of cortical processing. Thus, for example V2 and V1 cells respond to illusory contours2 [34], [87], [103], and to occluded contours defined by contextual depth cues [7], [96]. More generally, the response of cells in V1 (and extrastriate cortex) can be modified by contextual stimuli lying far outside the neurons’ receptive field [1], [10], [30], [60], [67]. What are the neural circuits underlying these fast, long-distance, global-to-local interactions? In this chapter we present a set of anatomical and physiological data exploring this issue.

Section snippets

Spatial properties of center–surround interactions in V1: where is the center, where is the surround?

The typical neural signature of long-range spatial computations at the single cell level is represented by center–surround interactions. These have been described for the first time in the retina 50 years ago [52], and are thought to endow retinal ganglion cells with the ability of signaling relative contrast rather than local luminance. Center–surround interactions are also observed in the responses of cortical cells, for which it is possible to define a discharge center or receptive field,

Spatial properties of V1 horizontal connections

V1 horizontal connections are long-range, reciprocal, intra-laminar projections prominent in the upper layers of V1 [28], [75] (but also present in layers 4B/upper 4Cα of primates and 5 and 6 of primates and carnivores). These connections arise from excitatory neurons, show a periodic, patchy pattern of termination, and contact predominantly (about 80%) excitatory neurons but also (about 20%) inhibitory neurons [48], [62], [64]. Horizontal axons have been found to link preferentially cortical

Spatial properties of feedback connections to V1

On the basis of laminar origin and termination, connections between visual cortical areas have been classified as feedforward or feedback, and a hierarchical organization of cortical areas has been proposed [22], [25], [76]. Area V1, at the bottom of the hierarchy, receives its main feedforward inputs from the thalamus, and sends feedforward projections to several extrastriate cortical areas which, in turn, send projections back to V1. In both cat and monkey, feedback connections to V1 arise

Conclusions

There is thus evidence for involvement of feedback connections as well as horizontal connections in center–surround interactions in area V1 neurons. The different extents of the divergences of horizontal and feedback connections from different areas (Fig. 7) provide a progressive tiling of the surround by horizontal and feedback connections from areas located further and further away from V1. This extensive overlap between the different feedback connections may explain why effects of V2

Acknowledgments

This work was supported by: European Community Grant Viprom Biomed 2, Wellcome Trust Grant 061113, and Research to Prevent Blindness, Inc., New York, INSERM, CNRS, GIS Cognisciences.

References (107)

  • U. Polat et al.

    The architecture of perceptual spatial interactions

    Vis. Res.

    (1994)
  • K.S. Rockland et al.

    Laminar origins and terminations of cortical connections to the occipital lobe in the rhesus monkey

    Brain Res.

    (1979)
  • G. Sclar et al.

    Coding of image contrast in central visual pathways of the macaque monkey

    Vis. Res.

    (1990)
  • F. Sengpiel et al.

    Different mechanisms underlie three inhibitory phenomena in cat area 17

    Vis. Res.

    (1998)
  • H. Sompolinsky et al.

    New perspectives on the mechanisms for orientation specificity

    Curr. Opin. Neurobiol.

    (1997)
  • D.D. Stettler et al.

    Lateral connectivity and contextual interactions in macaque primary visual cortex

    Neuron

    (2002)
  • J. Allman et al.

    Stimulus specific responses from beyond the classical receptive field: neurophysiological mechanisms for local–global comparisons in visual neurons

    Ann. Rev. Neurosci.

    (1985)
  • A. Angelucci et al.

    Convergence of color motion and form pathways in macaque V3

    Soc. Neurosci. Abstr. Online

    (2002)
  • A. Angelucci et al.

    Circuits for local and global signal integration in primary visual cortex

    J. Neurosci.

    (2002)
  • A. Angelucci et al.

    Functional specificity of feedforward and feedback connections between primate V1 & V2

    Soc. Neurosci. Abstr. Online

    (2003)
  • J.S. Bakin et al.

    Visual responses of monkey areas V1 and V2 to three-dimensional surface configurations

    J. Neurosci.

    (2000)
  • H.B. Barlow et al.

    The neural mechanisms of binocular depth discrimination

    J. Physiol.

    (1967)
  • P. Barone et al.

    Laminar distribution of neurons in extrastriate areas projecting to visual areas V1 and V4 correlates with the hierarchical rank and indicates the operation of a distance rule

    J. Neurosci.

    (2000)
  • C. Blakemore et al.

    Lateral inhibition between orientation detectors in the cat’s visual cortex

    Exp. Brain Res.

    (1972)
  • V. Bringuier et al.

    Horizontal propagation of visual activity in the synaptic integration field of area 17 neurons

    Science

    (1999)
  • J. Bullier, Communications between cortical areas of the visual system, in: L. Chalupa, J. Werner (Eds.), The Visual...
  • J. Bullier et al.

    Branching and laminar origin of projections between visual cortical areas in the cat

    J. Comp. Neurol.

    (1984)
  • C. Chen et al.

    Contrast response characteristics of long-range lateral interactions in cat striate cortex

    Neuroreport

    (2001)
  • H.J. Chisum et al.

    Emergent properties of layer 2/3 neurons reflect the collinear arrangement of horizontal connections in tree shrew visual cortex

    J. Neurosci.

    (2003)
  • T.A. Coogan et al.

    Hierarchical organization of areas in rat visual cortex

    J. Neurosci.

    (1993)
  • J.M. Crook et al.

    GABA-inactivation attenuates colinear facilitation in cat primary visual cortex

    Exp. Brain Res.

    (2002)
  • G.C. DeAngelis et al.

    Length and width tuning of neurons in the cat’s primary visual cortex

    J. Neurophysiol.

    (1994)
  • B. Dreher

    Thalamocortical and corticocortical interconnections in the cat visual system: relation to the mechanisms of information processing

  • F. Felisberti et al.

    Long-range interactions modulate the contrast gain in the lateral geniculate nucleus of cats

    Vis. Neurosci.

    (1999)
  • F. Felisberti et al.

    Long-range interactions in the lateral geniculate nucleus of the New-World monkey, Callithrix jacchus

    Vis. Neurosci.

    (2001)
  • D.J. Felleman et al.

    Distributed hierarchical processing in the primate cerebral cortex

    Cereb. Cortex

    (1991)
  • D. Ferster et al.

    Neural mechanisms of orientation selectivity in the visual cortex

    Ann. Rev. Neurosci.

    (2000)
  • C.D. Gilbert et al.

    Clustered intrinsic connections in at visual cortex

    J. Neurosci.

    (1983)
  • C.D. Gilbert et al.

    Columnar specificity of intrinsic horizontal and corticocortical connections in cat visual cortex

    J. Neurosci.

    (1989)
  • C. Gilbert et al.

    Spatial integration and cortical dynamics

    Proc. Natl. Acad. Sci. USA

    (1996)
  • P. Girard et al.

    Feedforward and feedback connections between areas V1 and V2 of the monkey have similar rapid conduction velocities

    J. Neurophysiol.

    (2001)
  • A. Grinvald et al.

    Cortical point-spread function and long-range lateral interactions revealed by real-time optical imaging of macaque monkey primary visual cortex

    J. Neurosci.

    (1994)
  • D.H. Grosof et al.

    Macaque V1 neurons can signal “illusory” contours

    Nature

    (1993)
  • H.K. Hartline

    The receptive fields of optic nerve fibers

    Am. J. Physiol.

    (1940)
  • G.H. Henry et al.

    Projections from area 18 and 19 to cat striate cortex: divergence and laminar specificity

    Eur. J. Neurosci.

    (1991)
  • R. Hess et al.

    Integration of contours: new insights

    Trends Cogn. Sci.

    (2000)
  • J.A. Hirsch et al.

    Synaptic physiology of horizontal connections in the cat’s visual cortex

    J. Neurosci.

    (1991)
  • D.H. Hubel et al.

    Receptive fields, binocular interaction and functional architecture in the cat visual cortex

    J. Physiol. (London)

    (1962)
  • J.-M. Hupé et al.

    Cortical feedback improves discrimination between figure and background by V1, V2 and V3 neurons

    Nature

    (1998)
  • J.-M. Hupé et al.

    Response modulations by static texture surround in area V1 of the macaque monkey do not depend on feedback connections from V2

    J. Neurophysiol.

    (2001)
  • Cited by (185)

    • Combining brain perturbation and neuroimaging in non-human primates

      2021, NeuroImage
      Citation Excerpt :

      NBRs are also found with frequency mapping of the auditory cortex, adjacent to areas of PBRs (Ortiz-Rios et al., 2017; Petkov et al., 2006; Tanji et al., 2010). Local inhibitory lateral connections (Angelucci and Bullier, 2003) as well as distal excitatory and inhibitory interactions might play a role in NBR effects through a push/pull mechanism (Angelucci and Bressloff, 2006). Although positive BOLD or MION-based fMRI responses tend to predominate during cortical electrical microstimulation, suppression of fMRI responses appears to be common in the few available NHP studies with subcortical stimulation.

    • 1.23 - Network Properties of Visual Cortex

      2020, The Senses: A Comprehensive Reference: Volume 1-7, Second Edition
    View all citing articles on Scopus
    View full text