Abstract

This study used the process-dissociation procedure (Jacoby, 1991) to examine the contribution of automatic and
controlled uses of memory to a stem completion task in 16 patients with Alzheimer’s disease (AD) and a matched group of healthy elderly subjects (EC). In an inclusion task subjects attempted to use a studied word to complete three-letter word stems, in an exclusion task they were instructed to complete stems with unstudied words. Relative to patients with AD, EC subjects produced more target word completions under inclusion conditions, and less target word completions under exclusion conditions. The probability of the AD group using studied words to complete stems was invariant across inclusion and exclusion conditions. Estimates derived from the process-dissociation calculations, showed that the performance of the AD patients was mediated entirely by automatic uses of memory, whereas for EC subjects controlled and automatic
processes codetermined task performance. Both estimates of controlled and to a lesser extent automatic uses of memory
were greater for the EC than the AD subjects, indicating that the stem completion impairment in AD may not be entirely attributable to a deficiency in controlled memory processes but also due to reduced automatic processing.