Beyond visual processing: A multi-method investigation into the role of early visual areas during reaching, grasping, and haptic exploration without vision

The human hand enables goal-directed actions, like reaching and grasping, and haptic object exploration, allowing us to interact with but also to perceive and contextualize our surroundings. Most of the time, we use vision to guide these processes as it is the dominant sense for extracting information about object properties that are relevant for appropriate hand-object interactions. Yet there are many situations in which visual information is not available, for example when we search for our keys in a pocket, or we reach for a door handle in complete darkness. In such circumstances we need to rely on memory and previous experience, as well as on somatosensory feedback to guide our actions and execute appropriate movements to achieve our goals. The present work is about the neural mechanisms behind grasping, reaching, and haptic exploration, particularly focusing on the role of the ventral visual stream and Early Visual Cortex (EVC), when visual input is unavailable. In fact, while the existing literature overall agrees on the neural underpinnings of visually guided, skilled hand actions, which consistently recruit dorsal stream areas in parietal cortex as well as somatomotor-related areas, no consensus has been reached on the brain areas subtending the execution of the same actions in the absence of visual information. Specifically, it remains a matter of debate whether early visual and ventral stream areas are consistently involved in the execution of reaching and grasping actions without visual input. In addition, the EVC, including the primary visual cortex (V1), also seems to be involved in haptic object exploration even in the absence of visual information, further pointing to a role of early visual areas that goes beyond unisensory processing of visual information. Yet it is unclear whether attributes of objects explored only through active touch might be represented in early visual areas, and whether this could be a mere epiphenomenon of top-down visual imagery. Finally, it has not yet been thoroughly investigated whether haptic processing in early visual areas might also be functionally relevant for behaviour. Therefore, the contribution of this work is three-fold. First, I show that across published neuroimaging studies there is consistent univariate activation of dorsal stream areas during the execution of skilled hand actions irrespective of the availability of visual information, and of ventral visual stream areas, along with the EVC, only when online visual input is available. Second, I provide evidence for the role of V1 in haptic processing of object size by showing that activity patterns in V1 can be used to decode the size of unseen, haptically explored stimuli, but not visually imagined ones, even if participants never had visual experience of the stimuli, thereby suggesting that haptic processing in V1 is not merely due to visual imagery. Finally, I present behavioural data on manual size estimation of unseen haptically explored stimuli. The behavioural results show that the location of the hand exploring the unseen target, relative to gaze direction and body midline, does not significantly modulate haptic size estimation, and suggest that haptic size processing likely occurring in V1 might be functionally relevant for behavior. However, future research is needed to replicate our findings and further elucidate whether and how V1 contributes to haptic size processing. Overall, this work contributes to our understanding of the neural mechanisms underlying the execution of skilled hand actions and haptic processing of object features in the absence of visual information, particularly addressing the controversial role of early visual areas. The findings of this thesis suggest that the visual system, particularly V1, acts in concert with well-established action-related and somatosensory brain networks. In particular, early visual areas show cross-modal modulations that can be explained within the predictive coding framework, wherein they might process haptic and action-related information to guide behaviour even when visual input is absent.