TY - JOUR
T1 - Evidence for a specific internal representation of motion-force relationships during object manipulation
AU - Mah, Christopher D.
AU - Mussa-Ivaldi, Ferdinando A.
PY - 2003/1
Y1 - 2003/1
N2 - Human subjects learned a tracking task which required them to point at a moving target with the free end of an inverted pendulum object. In order to determine how subjects represented this object internally, we studied learning interference between variants of this task in which the pendulum object had either stable or unstable dynamics. Using a novel method, agreement between possible internal representations of the two tasks was estimated by analysis of the motion-to-torque relationships experienced by each subject as they manipulated each object. It was possible to predict retention of the primary task on day 2 from our measure of agreement between primary and interfering tasks on day 1. This result suggests that the subjects learned the correct torque patterns to use to produce specific desired patterns of motion as they learned the balancing task. Surprisingly, the analyses indicate that retention was not impaired when similar motions of the two objects required retrieval of incompatible torque responses, but retention was impaired when similar patterns of motion in the two tasks required similar patterns of applied torque. These findings can be accounted for by a simple model of how multiple similar torque responses are selected and retrieved from memory when responses are freely chosen.
AB - Human subjects learned a tracking task which required them to point at a moving target with the free end of an inverted pendulum object. In order to determine how subjects represented this object internally, we studied learning interference between variants of this task in which the pendulum object had either stable or unstable dynamics. Using a novel method, agreement between possible internal representations of the two tasks was estimated by analysis of the motion-to-torque relationships experienced by each subject as they manipulated each object. It was possible to predict retention of the primary task on day 2 from our measure of agreement between primary and interfering tasks on day 1. This result suggests that the subjects learned the correct torque patterns to use to produce specific desired patterns of motion as they learned the balancing task. Surprisingly, the analyses indicate that retention was not impaired when similar motions of the two objects required retrieval of incompatible torque responses, but retention was impaired when similar patterns of motion in the two tasks required similar patterns of applied torque. These findings can be accounted for by a simple model of how multiple similar torque responses are selected and retrieved from memory when responses are freely chosen.
UR - http://www.scopus.com/inward/record.url?scp=0037990402&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0037990402&partnerID=8YFLogxK
U2 - 10.1007/s00422-002-0347-9
DO - 10.1007/s00422-002-0347-9
M3 - Article
C2 - 12545283
AN - SCOPUS:0037990402
SN - 0340-1200
VL - 88
SP - 60
EP - 72
JO - Biological Cybernetics
JF - Biological Cybernetics
IS - 1
ER -