Synergistic Processing of Visual Contours across Cortical Layers in V1 and V2

Abstract

Visual cortical areas are interconnected via layer-specific feedforward and feedback projections. Such intricate connections are thought to be essential for parsing complex visual images, but the synergy among different layers in different cortical areas remains unclear. By simultaneously mapping neuronal activities across cortical depths in V1 and V2 of behaving monkeys, we identified spatiotemporally dissociable processes for grouping contour fragments and segregating background components. These processes generated and amplified contour signals within specific layers in V1 and V2. Contour-related inter-areal interactions, measured as Granger causality, were also dominant between these cortical layers within a time window when the contour signals were rapidly augmented. The grouping process became much faster for isolated contour elements compared with visual contours embedded in a complex background. Our results delineate the mode whereby image components are grouped and segmented through synergistic inter-laminar and inter-areal processes that are dynamically adjusted during interpretation of sensory inputs.

Laminar Differences in Responses to Naturalistic Texture in Macaque V1 and V2

“…Recent studies have found similar laminar differences between V1 and V2 in the representation of contours embedded in a randomly oriented background (Chen et al, 2017). Like our results here, Chen et al (2017) found the highest sensitivity to contours in the supragranular and granular layers of V2 and in the supragranular and infragranular layers of V1, and they also attribute this pattern of sensitivity in V1 largely to top-down feedback from V2. However, our results provide two distinct insights about the selectivity and dynamics of laminar circuitry in early visual cortex.…”

“…Our experiments were performed under anesthesia, so represent only reflexive feedback mechanisms operating in the absence of goal-directed behavior. Second, sensitivity to embedded contours found in V2 is likely due to the larger receptive field sizes of V2 neurons (Chen et al, 2017). Thus, contour selectivity in V2 merely reflects integration of orientation signals within the classical receptive field, and not a more complex form of selectivity.…”

“…However, our results provide two distinct insights about the selectivity and dynamics of laminar circuitry in early visual cortex. First, the recordings in Chen et al (2017) are made from monkeys trained to perform a contour grouping task, and so their results may result from strong top-down influences reflecting cognitive states such as attention (van Kerkoerle et al, 2017). Our experiments were performed under anesthesia, so represent only reflexive feedback mechanisms operating in the absence of goal-directed behavior.…”

Modeling time-varying brain networks with a self-tuning optimized Kalman filter

“…As a note of caution, STOK can be used to derive directed functional connectivity measures within the Granger causality framework which has well-known strengths and limitations [ 101 – 103 ]. As such, it estimates linear temporal dependencies and statistical relationships among multiple signals in a data-driven way, without a guaranteed mapping onto the underlying neuronal circuitry [ 26 , 104 – 106 ]. However, STOK provides a novel formulation that is well-suited for incorporating model-based or physiologically-derived information that could favor more biophysically plausible interpretations.…”

Bottom-up saliency and top-down learning in the primary visual cortex of monkeys

“…Different temporal components of V1 responses are suggested to play different roles: the late components are related to feedback modulations ( 12 , 14 , 22 – 24 ), whereas the initial bursts are largely driven by bottom-up inputs. Training to detect camouflaged global contours has been observed to refine neuronal population code in V1 by affecting only the late responses ( 15 ).…”