Chapter 2. Prehension

148 THE PHASES OF PREHENSION
ability. Similar results were found by Paulignan, MacKenzie,
Marteniuk, and Jeannerod (1991), who also showed that the
variability in movements is not evenly distributed through the
trajectory, noting more variability before peak velocity of the wrist is
reached and less afterwards. This relates to what Jeannerod noted
above, how an initial ballistic phase gets the hand into the vicinity of
the target, and a second phase using feedback guides the hand to the
target. It is worth noting that reaching to grasp differs from the
pointing tasks discussed earlier, in which contact is not being made
with a surface and results show symmetrical velocity profiles (Atkeson
& Hollerbach, 1985).
Other researchers have contrasted reaching to grasp with reaching
to point in order to examine the motor control of the arm. Using an
ELITE system, Corradini, Gentilucci, Leo and Rizzolatti (1992) found
only an effect of target size, not the distal task on the deceleration
phase of the arm movement. They provided a computational model of
the controller, ARMAX, with a model order which confiied invari-
ance of the control program with regard to movement amplitude, with
sensitivity to the size (diameter) of the object. In contrast, Carnahan
and colleagues (1992; Carnahan, Goodale & Marteniuk, 1993) present
evidence that pointing and grasping are fundamentally different.
Comparing moving and stationary targets, Figure 5.15 shows a dis-
ordinal interaction in which pointing is better for stationary targets and
grasping is faster for moving targets. That the grasping system was
better able to deal with target motion than the pointing system was also
demonstrated in experiments in which the object location was per-
turbed. Carnahan discusses the adaptive functions of the design of vi-
suomotor systems for pointing in a stationary environment (e.g.,
communication) and grasping to acquire objects in a dynamic envi-
ronment (e.g., food, prey, defence, locomotion, etc).
Object velocity was more extensively investigated by Chieffi,
Fogassi, Gallese and Gentilucci (1992) who showed that the speed of
the arm movement covaried with the speed of the approaching object.
Interestingly, they found that kinematic landmarks for grasping were
unaffected by object velocity, as was the acceleration phase of move-
ment, but transport parameters like peak velocity, and duration of de-
celeration phase were affected. Also, the spatial endpoints and paths
were different in that as velocity increased, subjects grasped objects
closer to the body. They suggested that object motion prior to grasp-
ing was an extrinsic object property in this grasping task and thus only
affected the transport component.
The above evidence suggests that the task requirements for the