Abstract
Purpose. We routinely perform tasks while limited in our freedom to move, be it by carrying groceries or by an injury that prohibits normal use of a limb. Recent work indicates that we have access to accurate estimates of motor uncertainty in unconstrained movements (Maloney et al, VSS, 2006). Here we examine whether we can anticipate our movement uncertainty even when our freedom of movement is abruptly constrained. Methods. Participants (n=7) made fast reaches (35 cm in 250ms) with their right index finger to touch a target (circular, 0.4 cm) on a horizontal touch-screen. Prior to each trial, participants estimated their chances (0–100%) of hitting the target. Participants performed three blocks: B1) 40 unconstrained trials, B2) 30 trials with an elbow constraint that fixed the joint at a 90° angle, and B3) 30 unconstrained trials. Analysis. We calculated across-subject averages of the endpoint errors in the last 20 trials of each block to estimate the objective probability Oi of hitting the target at asymptotic performance in Bi. We averaged subjective estimates over the same ranges to obtain corresponding subjective probabilities Si. Results: Statistical tests indicated that O1=O2=O3: Participants had compensated for the constraint within the first ten constrained trials and showed only slight, and non-significant, improvements in the final post-constraint block. Participants accurately judged their performance before the constraint was imposed (O1=S1). However, S1[[gt]][[gt]]S2=S3: Subjects' predictions of constrained performance in B2 and of post-constraint performance in B3 were lower by almost a factor of two than their objective performance. Conclusions: Restraining the arm's degrees of freedom disrupted the ability to accurately estimate motor uncertainty and lead to a pronounced underestimation of objective performance. The finding has implications for motor impairments that are characterized by enduring losses in the freedom of movement.
Support: NIH NS047178-01A1 (UW), NIH EY08266 (LTM).