Visual information about object size and object position are retained differently in the visual brain: Evidence from grasping studies.

Abstract

Many experiments have examined how the visual information used for action control is represented in our brain, and whether or not visually-guided and memory-guided hand movements rely on dissociable visual representations that are processed in different brain areas (dorsal vs. ventral). However, little is known about how these representations decay over longer time periods and whether or not different visual properties are retained in a similar fashion. In three experiments we investigated how information about object size and object position affect grasping as visual memory demands increase. We found that position information decayed rapidly with increasing delays between viewing the object and initiating subsequent actionsimpacting both the accuracy of the transport component (lower end-point accuracy) and the grasp component (larger grip apertures) of the movement. In contrast, grip apertures and fingertip forces remained well-adjusted to target size in conditions in which positional information was either irrelevant or provided, regardless of delay, indicating that object size is encoded in a more stable manner than object position. The findings provide evidence that different grasp-relevant properties are encoded differently by the visual system. Furthermore, we argue that caution is required when making inferences about object size representations based on alterations in the grip component as these variations are confounded with the accuracy with which object position is represented. Instead fingertip forces seem to provide a reliable and confound-free measure to assess internal size estimations in conditions of increased visual uncertainty.

Cross-talk connections underlying dorsal and ventral stream integration during hand actions

“…Thus, the VOF supports the transfer of object-related properties necessary for efficient grasp formation by allowing the communication between ventral regions that process instrinsic object properties and dorsal regions that map spatial location to action plans (Takemura et al 2016; but see Freud et al 2016 for dorsal stream in object perception). This is in line with studies showing a crucial role of the VOF fibers in the visual shape processing (Lee Masson et al 2017) and the ventral stream cortical areas in processing object size during grasping (Schenk 2012;Hesse et al 2016)explaining why patients with optic ataxia and dorsal stream damage are able to use object size for grasping (Cavina-Pratesi et al 2010b). Furthermore, recent evidence suggests that perceptual features, processed by the occipital ventral stream areas, are used as priors by the dorsal visuomotor stream to specify goal-directed grasping actions and are activated earlier in time than dorsal regions involved in specifying action plans (Zimmermann et al 2016).…”

“…The computation of these visual object properties is crucial for successful hand preshaping in grasping movements (Castiello 2005), and when visual feedback is not available exaggerated opening of the hands occurs (Jakobson and Goodale 1991;Hesse et al 2016).…”

Neural representations of haptic object size in the human brain revealed by multivoxel fMRI patterns

“…Second, in studies of vision-for-action, a distinction is made (e.g. Hesse et al, 2016) between processing intrinsic (e.g. size) and extrinsic (e.g.…”

Do we have distinct systems for immediate and delayed actions? A selective review on the role of visual memory in action

“…However, there might be a different and more nuanced version of the dorsal-amnesia hypothesis that deserves further consideration and investigation. This nuanced version would stipulate that the time-dependent decay of visual information might be different for different visual attributes and also depends on the format of the representation, with allocentric representations being less affected by time-dependent changes than egocentric representations (see Hesse, Miller, & Buckingham, 2016;Hesse & Schenk, 2014 for examples). Accordingly, one might expect that the mix of visual information used for a given visuomotor task might indeed depend on the time elapsed between the receipt of that information and the onset of the associated action.…”