Switching Between Objects Improves Precision in Haptic Perception of Softness

Abstract

Haptic perception involves active exploration usually consisting of repeated stereotypical movements. The choice of such exploratory movements and their parameters are tuned to achieve high perceptual precision. Information obtained from repeated exploratory movements (e.g. repeated indentations of an object to perceive its softness) is integrated but improvement of discrimination performance is limited by memory if the two objects are explored one after the other in order to compare them. In natural haptic exploration humans tend to switch between the objects multiple times when comparing them. Using the example of softness perception here we test the hypothesis that given the same amount of information, discrimination improves if memory demands are lower. In our experiment participants explored two softness stimuli by indenting each of the stimuli four times. They were allowed to switch between the stimuli after every single indentation (7 switches), after every second indentation (3 switches) or only once after four indentations (1 switch). We found better discrimination performance with seven switches as compared to one switch, indicating that humans naturally apply an exploratory strategy which might reduce memory demands and thus leads to improved performance.

A Kalman filter model for predicting discrimination performance in free and restricted haptic explorations

“…consisting of more indentations) and larger symbols indicate explorations with more switches between the stimuli. B) Normalized observed discrimination performance in instructed explorations (study 2 [22]) as a function of normalized predicted performance. Here brighter colors indicate more number of switches.…”

“…There were four references and for each reference there were two comparisons one softer and the other harder. The compliances of the stimuli were the following: reference r 1 : 0.14 mm/N, comparisons c 1 1 : 0.12 mm/N, c We further took the behavioral data from [22] (later referred to as study 2). Here 11 participants completed a 2AFC task.…”

“…Both studies were conducted at a visuo-haptic setup, in which vision of the stimuli and the exploring hand was occluded and position of the index finger of the dominant hand and the force exerted on the stimuli were recorded during exploration. For more details please see [20] and [22] respectively.…”

“…And we could recently show that given the same amount of information (e.g. same number of indentations to compare the softness of two stimuli), more switches lead to better performance [22], suggesting that participants tune their exploratory behaviour to optimize performance. In the model of [13] the authors assume that after the exploration of the first stimulus a representation of its softness is built up.…”

“…We test the model by correlating empirically estimated performance in haptic softness discrimination and the predicted one. We take behavioral data from free softness explorations containing large variety of exploration patterns [20] and constrained softness explorations in which the number of switches was experimentally varied [22]. Additionally we assess whether natural haptic explorations are tuned to optimize performance by comparing the theoretically possible performance of different exploratory patterns and the frequency with which participants used those.…”