THE ROLE OF WORKING-MEMORY IN IMPLICIT SE QUENCE LEARNING

The focus of the present work is on implicit sequence learning, that i s, learning that occurs in the absence of an intention to learn and wi thout explicit recall of what has been learned. Existing research has primarily concentrated on three main issues: (1) the independence of i mplicit and explicit learning modes, (2) the role of attention in impl icit learning, and (3) the characteristics of the implicit learning me chanism(s). One issue that has received little attention thus far is h ow implicit learning ties in with the general cognitive architecture. In 3 experiments, we studied whether the working memory model, formula ted by Baddeley and Hitch (1974), provides reasonable constraints for understanding implicit sequence learning. Subjects in all experiments were asked to perform a sequential matching task in which the sequence of target symbols followed a systematic structure. Experiment 1 demon strated that sequences of phonological targets and sequences of altern ating phonological-visual targets are learned better than sequences of visual targets. Experiment 2 showed that phonological and visual sequ ence information is learned independently in mixed phonological-visual target sequences and is learned better than mixed phonological-visual sequence information. Experiment 3 demonstrated that both familiarity with the visual targets and presence of a phonological label for the visual targets improve implicit learning of visual sequence informatio n. Taken together, the findings imply that implicit sequence learning occurs in a short-term store. The results are surprisingly consistent with the working memory conception of Baddeley and Hitch (1974). Impli cit sequence learning appears to occur independently in the two workin g memory slave systems, phonological loop and visuo-spatial sketch pad . Learning of sequences that cross the slave systems appears to be imp aired and may thus not be coordinated by the Central Executive.