Visual influence on bimanual haptic slant adaptation

Abstract

Adapting to particular features of a haptic shape, for example, the slant of a surface, affects how a subsequently touched shape is perceived (aftereffect). Previous studies showed that this adaptation is largely based on our proprioceptive sense of hand posture, yet the influence of vision on haptic shape adaptation has been relatively unexplored. Here, using a slant-adaptation paradigm, we investigated whether visual information affects haptic adaptation and, if so, how. To this end, we varied the available visual cues during the adaptation period. This process ranged from providing visual information only about the slant of the surface, or the reference frame in which it is presented, to only providing visual information about the location of the fingertips. Additionally, we tested several combinations of these visual cues. We show that, as soon as the visual information can be used as a spatial reference to link the own fingertip position to the surface slant, haptic adaptation is very much reduced. This result means that, under these viewing conditions, vision dominates touch and is one reason why we do not easily adapt to haptic shape in our daily life, because we usually have visual information about both hand and object available simultaneously.

Figure 1.

Experimental setup. (A) The visuohaptic workbench. The participant was seated in front of the workbench with the body midline aligned with the center of the workbench. Both index fingers were fixed in thimble-like holders attached to the PHANToMs. The visual stimuli were presented on the CRT and viewed via mirror. The system was calibrated to easily present objects both visually and haptically in a spatially aligned way. (B) Front view of the virtual space showing the response zones, the threshold for starting a trial and the virtual surface.

Figure 2.

Adaptation conditions. Haptics Only: Subjects could feel the slanted surface but no visual cues were presented. Vision only: Subjects could see the slanted surface but not touch it. Boxes: Subjects could see eight boxes framing the area in which the haptic surface was presented; however, the slanted surface itself could only be felt but not seen. Cursors: Subjects could see the position of their index fingers as cursors, but the slanted surface was only rendered haptically. Slant and cursors: Subjects could see the position of their index fingers as cursors as well as the surface slant, which was also rendered haptically.

Figure 3.

Results of experiment 1. The y axis shows the haptic adaptation aftereffects as a result of different visual and haptic adaptation conditions. Each bar represents a different adaptation condition as indicated on the x axis. Error bars are standard errors. The dotted line marks the point at which full adaptation would occur. Note that, on the test trials for all adaptation conditions, the surface was rendered only haptically.

Figure 4.

The conditions of experiment 2. Haptics Only: The subjects could only feel the surface, but no visual information was available. Boxes and Cursors: Both the cursors representing the position of the fingertips as well the 8 visual boxes that spanned the available exploration space were presented (for details, see descriptions for experiment 1). The surface itself was not visually represented, but was available for haptic touch.

Figure 5.

Results of experiment 2. The x axis shows the labels for each condition, the y axis the size of the aftereffect, and error bars are the standard error. The dotted line marks the point at which full adaptation would occur.