In many day-to-day situations humans manifest a marked tendency to hold the head vertical while performing sensori-motor actions. For instance, when performing coordinated whole-body motor tasks, such as skiing, gymnastics or simply walking, and even when driving a car, human subjects will strive to keep the head aligned with the gravito-inertial vector. Until now, this phenomenon has been thought of as a means to limit variations of sensory signals emanating from the eyes and inner ears. Recent theories suggest that for the task of aligning the hand to a target, the CNS compares target and hand concurrently in both visual and kinesthetic domains, rather than combining sensory data into a single, multimodal reference frame. This implies that when sensory information is lacking in one modality, it must be 'reconstructed' based on information from the other. Here we asked subjects to reach to a visual target with the unseen hand. In this situation, the CNS might reconstruct the orientation of the target in kinesthetic space or reconstruct the orientation of the hand in visual space, or both. By having subjects tilt the head during target acquisition or during movement execution, we show a greater propensity to perform the sensory reconstruction that can be achieved when the head is held upright. These results suggest that the reason humans tend to keep their head upright may also have to do with how the brain manipulates and stores spatial information between reference frames and between sensory modalities, rather than only being tied to the specific problem of stabilizing visual and vestibular inputs.
Keywords: head posture; human motor control; multimodal perception; optimal sensori-motor integration; virtual reality.
Copyright © 2013 IBRO. Published by Elsevier Ltd. All rights reserved.