Abstract
In vision, only the information projected onto the central portion of our retina is perceived with high resolution. Therefore, the visual system needs to process the full visual scene with coarse resolution through peripheral vision and shift the eye in order to analyse a selected portion in detail (foveation). This process allows to reduce the complexity of visual processing by serializing detailed analysis.
A haptic process analogous to foveation has been described in the behavior of the blind star-nosed mole, who detects potential prays with any of its tactile appendages but analyzes it with a specific pair, characterized by higher tactile resolution.
Here we tested the hypothesis of haptic foveation behavior in humans. Nine participants searched for a particular configuration of symbols on a planar rigid tactile display. We computed the probability for each finger of touching a potential target after it was previously encountered by any of the other fingers, and the exploration speed of each finger while exploring a potential target.
Independent of which finger encountered a potential target first, there was higher probability that subsequent exploration was performed by the index or the middle finger. At the same time, these fingers dramatically slowed down, suggesting that these specialized fingers are involved in detailed analysis.
In a second experiment we tested the hypothesis that foveation is performed to gain information. Ten participants searched either for an easy target (a rough patch among smooth ones) or a difficult one (a hole in a certain corner of a patch). Overall, we replicated the results of the first experiment. Corraborating our hypothesis, specialized detailed analysis was reduced in easy search, suggesting that foveation behavior was employed less if it provided less information gain. Our results suggest that in haptic search humans employ foveation-like behavior similar as in vision.