Discovering affordances that determine the spatial structure of reach-to-grasp movements

Abstract

Extensive research has identified the affordances used to guide actions, as originally conceived by Gibson (Perceiving, acting, and knowing: towards an ecological psychology. Erlbaum, Hillsdale, 1977; The ecological approach to visual perception. Erlbaum, Hillsdale, 1979/1986). We sought to discover the object affordance properties that determine the spatial structure of reach-to-grasp movements--movements that entail both collision avoidance and targeting. First, we constructed objects that presented a significant collision hazard and varied properties relevant to targeting, namely, object width and size of contact surface. Participants reached-to-grasp objects at three speeds (slow, normal, and fast). In Experiment 1, we explored a "stop" task where participants grasped the objects without moving them. In Experiment 2, we studied "fly-through" movements where the objects were lifted. We discovered the object affordance properties that produced covariance in the spatial structure of reaches-to-grasp. Maximum grasp aperture (MGA) reflected affordances determined by collision avoidance. Terminal grasp aperture (TGA)--when the hand stops moving but prior to finger contact--reflected affordances relevant to targeting accuracy. A model with a single free parameter predicted the prehensile spatial structure and provided a functional affordance-based account of that structure. In Experiment 3, we investigated a "slam" task where participants reached-to-grasp flat rectangular objects on a tabletop. The affordance structure of this task was found to eliminate the collision risk and thus reduced safety margins in MGA and TGA to zero for larger objects. The results emphasize the role of affordances in determining the structure and scaling of reach-to-grasp actions. Finally, we report evidence supporting the opposition vector as an appropriate unit of analysis in the study of grasping and a unit of action that maps directly to affordance properties.

A geometric and dynamic affordance model of reaches-to-grasp: Men take greater risks than women.

“…The third is variations in the orientation of the grasp aperture as the hand approaches the target object. Mon-Williams and Bingham (2011) found those variations to be both random and systematic, varying as a function of the Maximum Object Extent. The variability in the orientation of the grasp aperture contributes to the variability in the spatial position of the fingers along the trajectory.…”

“…The goal of this study was to provide an account for the remaining free paramenter in an affordance model developed by Mon-Williams and Bingham (2011) to predict the spatial structure of reaches-to-grasp. Formulating a geometric model in line with Warren's affordance scaling models (e.g.…”

“…Pisoni and Remez, 2007), throwing (e.g. Bingham, Schmidt & Rosenblum, 1989;Zhu & Bingham, 2010;2011) and reaches-to-grasp (e.g. Jeannerod, 1984;Wing, Haggard & Flanagan, 1996).…”

“…The process models have provided only qualitative approximations to reach-to-grasp spatial structure that was assumed to vary only in 2D (that is, a strictly horizontal orientation of the MGA). The size of the MGA is known to co-vary systematically with the object size, but the specific scaling relation remained unknown until Mon-Williams and Bingham (2011) set out to discover the relevant object affordance properties and then, model the scaling of the Maximum Grasp Aperture (MGA) and the Terminal Grasp Aperture (TGA)…”

“…This requires the grasp aperture between thumb and index, for instance, to be opened wide enough and centered on the object during the approach of the hand. Mon-Williams and Bingham (2011) found that the Terminal Grasp Aperture (TGA), occurring at the end of a reach, reflected the targeting goal while the Maximum Grasp Aperture (MGA), occurring roughly midway through a reach, reflected the collision avoidance goal. When considering formation of the MGA, MonWilliams and Bingham realized that the hand and grasp aperture are not typically level or horizontal during the approach.…”