The learning of generalized motor program and parameterization : practice composition effects on retention of a motor skill

Abstract

Evidence that acquisition practice with variations of the criterion task (variable practice) leads to better retention than practice on the criterion task alone (specific practice) has been provided by Shea and Kohl (1990, 1991). These findings are consistent with the variability of practice hypothesis derived from schema theory (Schmidt, 1975, 1976) , in which movement production is based on a generalized motor program (GMP) and variable practice leads to a stronger, more robust schema than specific practice. The purpose of the present experiments was to reinvestigate the variable practice effect found by Shea and Kohl (1990,1991) and to examine whether manipulations of practice composition would have effects on the learning of GMP and parameter specification. Subjects practiced either a single movement (C) or three movement patterns (A, C, & E) with the same relative forces and the same relative timing, but with different absolute forces (Exp. 1) or different absolute timing (Exp. 2). Retention was assessed on task C. The retention results indicated that the specification of the fixed parameter was facilitated by variable practice with the parameter that was varied during practice. However, overall proficiency analyzed by RMS error and the learning of GMP assessed by scaled RMS error failed to indicate the benefit of variable practice on retention. These results are inconsistent with the findings of Shea and Kohl (1990, 1991) in terms of overall performance and Wulf and colleagues (Wulf & Lee,1993; Wulf & Schmidt, 1989, 1993; Wulf, Schmidt, & Deubel, 1993; Wulf, 1992) in terms of the learning of GMP.