Efficacy of errorless learning in the acquisition of a new procedural skill in Alzheimer's disease

In Alzheimer’s disease (AD), implicit or procedural rehabilitation techniques would be more effective to train new skills than explicit or declarative learning methods (van Halteren-van Tilborg, 2007). Following Baddeley and Wilson (1994)’s assumption, Maxwell et al. (2001) showed that reducing errors during motor learning minimizes the building of declarative knowledge and would allow implicit knowledge accumulation. However, most studies on errorless learning focused on learning of face-name associations (Clare et al., , 2001), and very few studies have investigated errorless learning in procedural learning situations, even though some data suggest that errorless learning would be efficient for learning instrumental activities of daily living (e.g., Thivierge et al., 2008). The aim of this study was to examine the efficacy of an error-reducing versus an errorfull method in motor skill learning. We examined the acquisition of a new motor skill in 24 patients with AD and 24 healthy older adults matched for age, sex, and education. In this task, subjects had to follow with a reversed mouse the contour of a form (a star) displayed on a computer screen. Half the subjects learned in an error-reducing condition, and the others in an errorfull condition. After the learning phase, all the subjects had to complete a novel form. Results show an advantage for the error-reducing condition in the AD group, whereas the performance of the healthy participants did not differ between the two conditions, confirming the efficiency of errorless learning principles in AD for procedural learning situations