Uncertainty and spatial updating in posterior parietal cortex

Abstract

Vision is an active process that requires frequently reorienting the central, high-acuity part of the retina toward relevant objects in the world to enable further processing. About three times per second we make saccadic eye movements that shift the entire image on the retina, yet the world appears stable and we can effortlessly keep track of where things are. The question of how the brain achieves visual stability across eye movements (or self-motion in general) has fascinated many great thinkers, including von Helmholtz (von Helmholtz, 1867) who had already recognized that our sense of location must come from a synthesis of retinal coordinates with information about eye position and movement. In recent years much progress has been made in uncovering the neural and computational mechanisms of visual stability. From a computational point of view, modeling studies have shown how populations of neurons could maintain a stable mapping between an internal retinotopic representation and spatiotopic (world-centered) coordinates of objects in the world (Casarotti et al., 2012;Denève et al., 2007;Zipser & Andersen, 1988). These models rely either on proprioceptive signals from the eye to build an explicit spatiotopic map or on a copy of the motor command sent to move the eye, an efferent copy, that would be used to update the location of the target of interests on the internal retinotopic map. The latter mechanism can support a working spatiotopy that keeps track of locations across eye movements without postulating an explicit spatiotopic internal map. What concerns the neural mechanisms, a seminal study (Duhamel, Colby, et al., 1992) established a link between the processing of efferent copy signal and the posterior parietal cortex (PPC), by demonstrating the existence of neurons in the monkey lateral intraparietal area (LIP, which contains a retinotopic representation of the visual field) that, even before a saccade start, respond to the stimuli that will land in their receptive field after the eye movement. This result inspired investigations of visual stability in human patients with PPC damage (e.g. Heide et al., 1995;Vuilleumier et al., 2007), leading to the conclusion that an intact parietal cortex is crucial for maintaining a stable sense of location across eye movements. A new study by Fabius and colleagues (Fabius et al., in press) however, present data that challenge this conclusion and call instead for a more nuanced view about the role of PPC in visual stability.Many of the original studies on visual stability used an experimental protocol known as doublestep task. In this task, participants are asked to make two successive saccades toward the location of two sequentially flashed targets, both of which disappear before the first saccade begins. Thus,

Towards assessing extra-retinal uncertainty: A reply to M. Lisi (2020)

“…In a commentary on our study, Lisi (2020) provided an interesting alternative interpretation of our results and the results of the double-step saccade studies. He posits the hypothesis that these results can potentially be explained with a simple model that he and his colleagues developed (Lisi, Solomon, & Morgan, 2019).…”

“…From a theoretical perspective, the asymmetric cost function is an elegant model because it captures several aspects of saccadic behavior with only a single free parameter: the degree of asymmetry in the cost of forward versus backward corrective saccades. Applying the model to extra-retinal signals and PPC lesions, Lisi (2020) argued that the lesion in the patients of our study could have decreased the precision of extra-retinal information, irrespective of whether there are different types of extra-retinal signals involved. The hypothesis is that a reduced precision of extra-retinal signals will lead to a lower precision in the ISD task.…”