Many motor skills require a precise time sequence of commands from the brain. Recent studies suggested that human subjects tend to learn state representation rather than time representation while adapting to a time varying force field. Sequence of force fields could be regarded as a discrete form of time dependent field. Can subjects learn such a sequence? More specifically, can they form a representation of a variable environment when the environment changes in a very predictable manner from one movement to the next?
We have tried to reveal the ability to represent time while learning to perform a reaching movements in a two dimensional workspace. Subjects performed reaching movements, in a sinusoidal time varying force field and in an alternating viscous force field.
We found that (a) subjects did not adapt to the time varying force field and (b) were unable to form a consistent representation of the simple sequence of force fields. In the latter case they did adapt to a state dependent field that produced similar forces when the hand passed by the same region of state space. This first result, (a), refines our understanding of the adaptation mechanism and strengthens the hypothesis that this system lacks the ability for time representation. The second result, (b), suggests that the motor control adaptation mechanism is incapable of sequence representation, even a simple sequence that could easily be learned implicitly in other cognitive tasks.
We present a theoretical framework for the distinction between time, state and sequence representation and speculate that the observed ability to produce timely accurate movements in musical and dance performance is the result of other higher level learning mechanisms.
Keywords: Time representation; Internal model; Sequence learning; Arm movement; Reaching movements; Force fields; Adaptation; Learning; Manipulandum
[poster, pdf ]
[Publication list, Amir Karniel ]