Item

Abstract

In a grasping task, Johansson et al. [1] found that subjects look at the position to which the finger tips are guided. However, in their experiment, only the contact position of the thumb
was visible. We investigated what happens if the contact positions of both finger and thumb are
visible. We recorded eye and finger movements. In a first experiment, subjects always grasped
with the index finger at the top and the thumb at the bottom of a flat shape that was mounted on
a horizontal bar. In order to see whether a salient feature of a shape would affect the fixation
positions, we presented an (asymmetric) cross in 4 orientations (with the crossing of the bars
representing the salient feature). In order to see whether gaze is attracted to the position where
the finger has to be guided relatively precisely, we presented a triangle in two orientations that
subjects had to contact at the base and at the pointed top (i.e., a higher required precision to
contact the top than the base). We found that the crossing of the bars attracted the gaze whereas
the top of a triangle did not. A prominent result was that subjects fixated above the center of
the shape. In order to distinguish between subjects fixating the upper part of the shape versus
being attracted by the index finger, we mounted a square and a triangle in two orientations on a
vertical bar. We asked subjects to grasp first with one hand and then with the other so that the
shape remained constant but the contact positions of the index finger and thumb were reversed.
Subjects still looked above the center. In addition, for the square the gaze was now attracted to
the thumb so that the average fixation location was to the left or to the right half of the square,
depending on the used hand. The latter result indicates that gaze not only depends on the shape
but also on features of the grasp. This is also supported by the finding that the variability in
fixation locations was more in the horizontal direction for the second experiment than for the
first, i.e. the variability tends to be along the grasping axis. We conclude that both features of
the shape and the grasp affect gaze location.