Transcript Reaching & Grasping

Reaching & Grasping
Summary of points for WS04 –
from Weiss & Jeannerod (1998)
Nature of the problem
• How do we control multiple DoF (degrees of
freedom) within this motor task that requires a
reach and grasp...
– involves torso, shoulder, upper arm, lower arm, hand,
fingers all coordinated together to stretch out to a
target, close around it and lift it
• Could be a problem of getting to the end position
(system seeks control of end posture)
• Could be a problem of the way we move to the
end position (system seeks control of trajectory)
Nature of the problem
• Research looks at time
– Limb segments organized in terms of timing of muscle
contractions
• Research looks at synergies
– Muscle contractions are repeatedly activated together
within practiced movements, thus forming “joint
systems” that come together at the level of the
movement (coordinative structures)
– Thus movements might be planned at the level of
these synergies (Bernstein, 1967)
– This is probably reflected by the formation of neural
networks accompanying the specific joints in the
movement
Nature of the problem
• Viability of these solutions is examined by
creating models based on these
assumptions and seeing whether they
match up to people’s movement behaviors
• Several such models have been tested...
Some controversy
• Model 1:
– Movement planned via movement of hand
through space (Hogan & Flash, 1987) –
predicts flat trajectory
• Model 2:
– Movement planned as series of joint
configurations – predicts curved trajectories
• Overall, seems to depend on the task – we
use both kinds of trajectories.
Simplifying the problem
• Freezing degrees of freedom
– When the target position varies, only some
parts of the movement are changed. Others
remain the same – implying that we tend to
only change aspects of a movement when we
have to...simplifying the coordination issue.
Underlying representations
• One view suggests we memorize stored
final postures, in terms of stiffness of
muscles required for particular positions.
– Movements then consists of simply “gliding” to
that final stiffness of the muscles, and hence
the final posture.
Underlying representations
• Another view points out that the nature of
movement is not consistent with this, and
instead suggests that the memory includes
information about future task intention
– think of how you grasp a door handle, and
how the grasp changes depending on
whether you want to twist the handle
clockwise or anti-clockwise
More on representations
• Representation = feedforward control
– Something that can start a movement and
anticipate the feedback that it will generate
(based on past experience), using this to drive
further movement.
– Seems to be localized in the cerebellum
– Seems to rely on updating through
proprioception in order to continually control
multi-joint movements (with the implied
awkward set of torques involved)