A Neural Model of Figure–Ground Organization

Abstract

Psychophysical studies suggest that figure-ground organization is a largely autonomous process that guides--and thus precedes--allocation of attention and object recognition. The discovery of border-ownership representation in single neurons of early visual cortex has confirmed this view. Recent theoretical studies have demonstrated that border-ownership assignment can be modeled as a process of self-organization by lateral interactions within V2 cortex. However, the mechanism proposed relies on propagation of signals through horizontal fibers, which would result in increasing delays of the border-ownership signal with increasing size of the visual stimulus, in contradiction with experimental findings. It also remains unclear how the resulting border-ownership representation would interact with attention mechanisms to guide further processing. Here we present a model of border-ownership coding based on dedicated neural circuits for contour grouping that produce border-ownership assignment and also provide handles for mechanisms of selective attention. The results are consistent with neurophysiological and psychophysical findings. The model makes predictions about the hypothetical grouping circuits and the role of feedback between cortical areas.

Occipital network for figure/ground organization

“…The image region categorized by the brain as a Wgure 'owns' the contour, appears closer to the observer, has a speciWed shape, and has increased salience; while the ground has no contour, appears farther away, is 'shapeless' and is perceived as extending behind the Wgure and the contour (Rubin 1921(Rubin / 2001. The Wgure/ground categorization thus is considered to precede and govern attention (Palmer 1999;Craft et al 2007).…”

“…Neurophysiologically, a role for MT/MST in Wgure/ground assignment for structure-from-motion has been suggested for the past two decades (Allman et al 1985;Olavarria et al 1992;Eifuku and Wurtz 1999;Born 2000;Born and Bradley 2005). Rapid feedback to V1 from areas with large receptive Welds, such as V2 or MT, was proposed in many cases Lamme 1995;Zipser et al 1996;Lee et al 1998;Nothdurft et al 1999Nothdurft et al , 2000Hupé et al 2001;Craft et al 2007). The connections between V1 and MT were found to conduct at the high speed of »10 m/s Wiesel 1983, 1989;Born and Tootell 1991;DeAngelis et al 1994;Grinvald et al 1994;Fitzpatrick 2000;Anderson et al 2001;Girard et al 2001;Bair et al 2002).…”

“…This extended suppression must derive from a mechanism able to categorize all regions of the cortical representation on the basis of the global Gestalt relations between them and that can send or withhold the inhibitory signal back to the requisite locations of V1 and V2 accordingly. As will be detailed in the General Discussion, there are both neurophysiological and anatomical data suggesting that most likely a top-down feedback from a higher area, that is able to implement the global Gestalt principles, underlies the Wgure/ground process (Mumford 1992;Lamme 1995;Lamme et al 1998;Hupé et al 2001;Craft et al 2007). …”

“…What neural architecture allows local and global information to be combined rapidly, to provide for Wgure/ground categorization? In terms of neural architecture, does the mechanism employ lateral interactions (e.g., Gerrits and Vendrik 1970;Grossberg and Mingolla 1985;Grossberg 1994;Baek and Sajda 2005;Pao et al 1999;Zhaoping 2005), top-down feedback (e.g., Lamme and Roelfsema 2000;Lamme et al 2002;Roelfsema et al 2002;Craft et al 2007) or more sophisticated combinations? In terms of its coding mechanism, do the Wgure/ground conWgurations use border-based coding (e.g., von der Heydt et al 2003Heydt et al , 2005Craft et al 2007) or region-based coding (e.g., Lamme 1995;Zipser et al 1996;Lamme et al 1998;Super et al 2001;Roelfsema et al 2002)?…”

The role of oscillatory brain activity in object processing and figure–ground segmentation in human vision

“…The role of gamma rhythms in figure-ground segregation is hypothesised to result from 486 one or two general processes: (i) region labelling, achieved by labelling corresponding 487 elements in an isomorphic surface representation (Lamme, 1995); and/or (ii) border 488 ownership coding, achieved through contour representation following the activity of 489 orientation-selective units (Craft, et al, 2007). These different schemes may explain the 490 opposing power changes in gamma reported here.…”