Cortical dynamics of lateral inhibition: Visual persistence and ISI

Abstract

Psychophysical studies show that increasing the interstimulus interval (lSI) between two stimuli decreases persistence of the first stimulus. Whilesome researchers account for these results with interactions of transient and sustained inhibition, this paper describes an alternative explanation. In a neural-network model of boundary detection called the boundary contour system, persistence is the result of feedback-generated reverberations. Mechanisms to control these reverberations include lateral inhibition, which computer simulations show allows persistence in the network to qualitatively match the psychophysicaldata. Additionalsimulations predict that increasing the duration of the second stimulus should cause persistence to increase with lSI. The model links psychophysical data on visual persistence with computational requirements of spatial vision and properties of cells in visual cortex. Burr (1980) observed that moving stimuli were not as blurred as would be expected. Studies of visual persistence show that stimuli persist for hundreds of milliseconds (e.g., Bowen, Pola, & Matin, 1974), which would produce substantial blurring of moving stimuli. However, Burr suggested that mechanisms responsible for detecting motion inhibited persisting representations ofa stimulus to deblur the moving percept. Di Lollo and Hogben (1987) found an inverse proximity effect, where nearby stimuli that were spatially close and temporally contiguous reduced the persistence of nonmoving stimuli. This finding suggested that motion-detection mechanisms were not critical for deblurring moving stimuli, and that lateral inhibition interacting across space and time inhibited persisting representations of past stimuli. Farrell, Pavel, and Sperling (1990) showed similar results using stimuli that moved in apparent motion. They also developed a descriptive model of visual persistence that outlined how lateral inhibition from a new stimulus affects the decay of old stimulus representations. In this model, persistence is the result ofa fading trace, and lateral inhibition decreases the strength of the trace so that it more quickly decays below a constant threshold.While the idea of persistence as a fading trace is intuitively appealing, the mechanisms underlying the trace and their interaction with mechanisms oflateral inhibition appear to be more complicated than the Farrell et al. (1990) description. Di Lollo andHogben (1987) found that the effects of nearby stimuli were stronger with longer stimulus onset asynchronies (SOAs), and Castet (1994), using a method very similar to Farrell et al.'s, demonstrated that persistence ofa leading stimulus decreases as Correspondence should be addressed to the author at Department of Psychological Sciences, 1364 Psychological Science Building, Purdue University, West Layfayette, IN 47907-1364 (e-mail: gfrancis@psych. purdue.edu). the interstimulus interval (lSI) between it and a trailing stimulus increases. Figure la shows data for 2 subjects from Castet. Farrell et al.'s model did not consider the...

Recent models and findings in visual backward masking: A comparison, review, and update

“…When stimuli change or move, this persistent activity would be problematic in that it could give rise to forward masking (Breitmeyer, 1980(Breitmeyer, , 1984 or to motion smear (Chen et aI., 1995). In the BCS model, such persistence can be curtailed in one of two ways: (I) by a gateddipole mechanism (Francis et aI., 1994) that, at stimulus offset, produces a reset signal inhibiting the persisting reverberatory activity or (2) by lateral inhibition operating mainly at the first competitive stage (Francis, 1996a).…”

“…According to Francis (1996aFrancis ( , 1997, the latter mechanism of lateral inhibition is the more significant contributor to the dynamics of stimulus boundary erosion over time under metacontrast conditions. Figure 5 shows schematically how the model accounts for key properties of metacontrast.…”

“…Unfortunately, a practical limitation of this model, as well as of similar network models, is that the solutions to the network's differential equations are very time consuming (see the appendix of Francis, 1997), and this may prohibit extensive testing of the model. Despite this practical limitation, the model already manifests a wide explanatory scope, providing an integrative scheme accounting for major findings on visual persistence, temporal integration, metacontrast masking, and apparent motion (Francis, 1996a(Francis, , 1996b(Francis, , 1997(Francis, , 1999Francis & Grossberg, 1996b;Francis et aI., 1994). Moreover, the model makes several explicit predictions regarding psychophysical and neurophysiological findings.…”

Backward Referral, Flash-Lags, and Quantum Free Will: A Response to Commentaries on Articles by Pockett, Klein, Gomes, and Trevena and Miller

“…masked the flash one video frame after its presentation and found that a flash lag was still observed. Masking after 15 ms is reported to reduce visual persistence (Castet, 1994;Francis, 1996) so the fact that a flash-lag effect was still observed under these conditions can probably be legitimately interpreted as showing that greater persistence of the flash than the bar is not the cause of the flash-lag effect.…”

A neural network model for the development of simple and complex cell receptive fields within cortical maps of orientation and ocular dominance

“…This sort of habituative synapse has been used to explain a variety of perceptual data that involve visual cortex (Francis, 1996a(Francis, , 1996bGrossberg, 1996a, 1996b;Francis et a!., 1994;Grossberg, 1976bGrossberg, , 1980Grossberg, , 1987Grossberg, , 1997Grunewald and Grossberg, 1996). Abbott et al (1997) have reported habituating cells in rat visual cortex and model them using (15).…”