On the insufficiency of laterality-based accounts of face perception and corresponding visual field asymmetries
Loading...
Authors
Harrison, Matthew T
Issue Date
2021
Type
Dissertation
Language
Keywords
Cerebral laterality , Contralateral bias , Face recognition , Left visual field advantage , Right hemisphere superiority
Alternative Title
Abstract
It has been known for nearly a century that the left half of a face is better recognized than the right half (Wolff, 1933). This left half-face advantage is commonly thought to reflect a combination of right hemisphere (RH) superiority for face recognition and a contralateral hemifield-hemisphere correspondence between the RH and the left visual field (LVF). The purpose of this set of experiments was to determine whether RH superiority for faces and contralateral hemifield-hemisphere correspondence is sufficient to explain the LVF half-face advantage. We set out four aims to accomplish this: (1) Use behavioral and fMRI methods to demonstrate the LVF half-face advantage and identify its neural basis in ventral occipital-temporal cortex (VOTC); (2) use behavioral methods to show that RH superiority is insufficient to explain the LVF half-face advantage; (3) use behavioral methods to show that we perceive only one half of a face at a time; and (4), albeit not initially proposed, use methods developed to accomplish aims 1-3 to distinguish retinotopic face representation from face-centered representation.In our first set of experiments (behavioral and fMRI), we identified for the first time a neural LVF half-face bias in RH face-selective cortex. We also found that the neural LVF bias in right FFA underlies the relationship between FFA laterality and the LVF half-face advantage. This revealed an explicit neural mechanism to describe the commonly assumed basis of the LVF advantage for centrally-viewed faces. In our next set of experiments (behavioral) we addressed the second aim, and found that LVF half-face advantage is contingent upon the simultaneous presence of both an upright LVF and RVF half-face, and does not reflect inherently superior processing of LVF over RVF half-face information. This challenged the sufficiency of the mechanism we discovered in Aim 1 as an explanation of the LVF half-face advantage. In our next set of behavioral experiments (which addressed our third aim) we found that half-face identities compete for limited processing resources, and only one identity can be processed at a time. Furthermore, we found that this does not apply to faces in which half-face identities are similar enough to be perceived as a normal (i.e. non-chimeric) face. In our final set of experiments (behavioral) we addressed our additional Aim 4, and found that the LVF half-face advantage occurs regardless of the location of the face in the visual field. This suggests that faces are represented to some degree in an object-centered reference frame, and the LVF half-face bias reflects a bias to the left half of a face, rather than a retinotopic bias to the left half of visual space.