Grasp: combined contribution of object properties and task constraints on hand and finger posture.

We compared the effect of different object properties on human upper-limb posture during reach-to-grasp tasks. A combination of extrinsic (object position), intrinsic (object type) and contextual object properties (grasp type) was investigated. Three-dimensional reach posture was measured by the hand position and orientation relative to the object at the time of stable object contact (with the digits). Similarly, the grasp posture was quantified by the angular digit configuration at the time of stationary object contact. We found that hand position and hand orientation were not only dependent on object position, as previously hypothesized, but also on object type and grasp type. Similarly, angular digit configuration was also dependent on extrinsic and contextual properties, and not only on object type (the intrinsic property). Principal component analysis revealed that two principal components (PCs) explained >79 % of the variation in the reach posture, whereas four PCs explained >76 % of the variation of the grasp posture. Again, PCs represented combinations of the input variables, i.e., there was no clear separation between the extrinsic variable acting specifically on the reach component, and the intrinsic variable on the grasp component. Contrary to the Dual Visuomotor Channel theory, these results suggest that extrinsic, intrinsic and contextual object variables do not act separately and exclusively on the neural control of the reach component or on that of the grasp component, but interact on both.