Precedence of parvocellular- over magnocellular-biased information for 2D object-related shape processing

Abstract

The question of how we frame mental representations in particular coordinate systems to facilitate the kinds of computations necessary for our interactions with objects has long been the focus of research in cognitive science (e.g., Almeida et al., 2020; Buxbaum & Coslett, 1997; Caramazza & Hillis, 1990a, 1990b; Hillis & Caramazza, 1995; Medina et al., 2009; Tadin et al., 2002; Vannuscorps et al., 2021a, 2001b). Take the following examples: if you want to pick up your car keys, you will need to locate the keys within a coordinate system that relates to the position of your hands; however, if you want to identify those car keys within a set of other objects (e.g., house keys, a USB pen), you may need to encode the visible input in coordinates intrinsic to those objects in order to achieve perceptual constancy and better match them with stored object representations. In line with this, extant research has suggested that there are many coordinate systems onto which a representation can be framed depending on the computational goal of the processes at play, and that these frames of reference span different levels of processing (from lowerto higher-level representational frames; Almeida et al., 2020; Buxbaum & Coslett, 1997; Caramazza & Hillis, 1990a, 1990b; Hillis & Caramazza, 1995; Medina et al., 2009; Tadin et al., 2002; Vannuscorps et al., 2021a, 2021b). In the realm of visual perception, one basic reference frame relates to the retinal position of the stimulus – that is, at the earliest stages of visual processing, a stimulus is represented relative to its original retinotopic coordinates. However, other important computations and representational levels happen at later stages of processing, when a representation becomes progressively stripped of its most basic sensorial nature and is framed over more abstract types of coordinate systems (e.g., object-centered; hand-centered). Vannuscorps et al. (2021a; see also Vannuscorps et al., 2021b) recently suggested that there is an intermediate stage of processing in visual perception that requires framing 2D bounded regions of space using coordinate references that relate to the shape of that bounded region – what they call intermediate shapecentered representations (ISCRs). That is, the representation of that shape relates not to its location in the visual field/retina, but rather to the main axes of the shape itself. In their paper, they meticulously tested an individual – Davida – who shows a remarkable deficit for the perception of orientation of 2D shapes. Specifically, Davida perceives the orientation of shapes as continuously alternating between orthogonal orientations (i.e., between the correct orientation and orientations that result from mirroring the shape over its axes). This impairment is observed both when she makes visual orientation judgments, as well as when she is asked to manually perform putative actions over 2D shapes (e.g., point to the tip of an arrow presented on a screen; touch the extremities of a 2D shape presented on the screen). Because Davida’s misperception of orientation conforms exclusively to orientations that are possible if the shape is rotated or mirrored over its main axis (and not any other orientation), her