Prespecified
After-effects Elicited from Robotic Force Fields
James L. Patton & Ferdinando A. Mussa-Ivaldi
The Sensory Motor Performance Program at The Rehabilitation
Institute of Chicago
Patton, J.
L., Mussa-Ivaldi, F.A., "Prespecified After-effects Elicited from Robotic
Force Fields" Society for Neroscience, 63.16, November, 2000, New Orleans,
LA, USA.
When subjects make reaching movements in the presence of a force
field that systematically disturbs motion, they progressively adapt their
feedforward commands. They exhibit characteristic "after-effects"
when the disturbing force field is unexpectedly removed. Interestingly, both
the initial adaptation and the after-effects can occur without subjects being
aware of the adaptive process. Dynamic models of the neuro-musculoskeletal
system can predict such after-effects and hence make it possible to design
force fields capable of inducing desired movements as after-effects. Our
algorithm first identifies a model of the human arm-and-controller using a
linear combination of slightly different arm-and-controller models. The model
is used to fit an array of regional force fields that, following training,
should result in a new, "desired" movement as a after-effect. We used
a hand-held, two-link, planar manipulandum robot that exerts forces as subjects
make point-to-point reaching movements in the horizontal plane. Results in the
after-effect trials show a clear shift from the original unperturbed
trajectories to the desired movements, even though subjects were never given
any information about the desired movement. This approach is an alternative to
training methods based on the explicit specification of the desired movement to
the learner. We are investigating the efficacy of this approach in patients
that exhibit stereotypical deficits in movement such as hemiparetic stroke.
Supported by NIH 5 T32 HD07418 & NIH 5 RO1 NS 35673 & 1 F32 HD08658-01