Brain Activation During Visually Guided Finger Movements

Abstract

Computer interaction via visually guided hand movements often employs either abstract cursor-based feedback or virtual hand (VH) representations of varying degrees of realism. The effect of changing this visual feedback in virtual reality settings is currently unknown. In this study, 19 healthy right-handed adults performed index finger movements ("action") and observed movements ("observation") with four different types of visual feedback: a simple circular cursor (CU), a point light (PL) pattern indicating finger joint positions, a shadow cartoon hand (SH) and a realistic VH. Finger movements were recorded using a data glove, and eye-tracking was recorded optically. We measured brain activity using functional magnetic resonance imaging (fMRI). Both action and observation conditions showed stronger fMRI signal responses in the occipitotemporal cortex compared to baseline. The action conditions additionally elicited elevated bilateral activations in motor, somatosensory, parietal, and cerebellar regions. For both conditions, feedback of a hand with a moving finger (SH, VH) led to higher activations than CU or PL feedback, specifically in early visual regions and the occipitotemporal cortex. Our results show the stronger recruitment of a network of cortical regions during visually guided finger movements with human hand feedback when compared to a visually incomplete hand and abstract feedback. This information could have implications for the design of visually guided tasks involving human body parts in both research and application or training-related paradigms.

Keywords: action observation; functional magnetic resonance imaging; healthy adults; virtual reality; visually-guided finger movements.

Figure 1

Illustration of the experimental setup. (A) In the MRI, participants wearing a data glove and grasping a tube (with the right hand). Visual feedback was delivered via mirror projection from a monitor. (B) Visual feedback consisted of the display of the starting position (light-blue circle), of the movement cursor (skin-colored circle, approximately Type III on the Fitzpatrick scale), and the target (red circle) on grey background. (C–F) The experiment comprised eight conditions, with four different types of visual feedback. (C) Cursor (CU), (D) point light (PL), (E) shadow hand (SH), and (F) virtual hand (VH) feedback.

Figure 2

Group-activations of action > baseline (blue) and of observation, > baseline (green) both overlaid on a rendered brain. (A) Left lateral, (B) right lateral, (C) left medial, and (D) right medial view. Primary motor (M1) and somatosensory (S1) cortex, dorsal (PMd) and ventral (PMv) premotor cortex, supplementary motor area (SMA), precuneus, inferior parietal lobule (IPL), occipitotemporal cortex (OTC), and cerebellum predominantly activated. Activations are shown at p < 0.05 (FWE corrected).

Figure 3

Illustration of the significant BOLD signal differences comparing feedback types. Activations are shown for both action (A–C) and observation (D–F) conditions overlaid on a left lateral rendered brain. Right-hemispheric activations were small and not shown. Activations of all plots were FWE corrected (p < 0.05). This figure illustrates that during the action and observation realistic displays of a human hand (SH and VH) activate the visual cortex more strongly than point-light (PL) or cursor (CU) displays.

Figure 4

Activations in four literature-driven (EBA, hMT+, LOC, and LOTC, see Table 1) left-hemispheric ROIs. Beta value means and standard errors extracted for VH (black), SH (gray), PL (dark red), and CU (light red) conditions during both action (top) and observation (bottom). Asterisks indicate significant changes between conditions (p < 0.05). Abbreviations: EBA, extrastriate body area; hMT+, human area MT+; LOC, lateral occipital complex; LOTC, lateral occipitotemporal complex. This figure indicates that during the action and observation realistic displays of a human hand (SH and VH) activate the literature-driven ROIs more strongly than point-light (PL) or cursor (CU) displays.