Capacity limitations in short-term memory in schizophrenia: tests of competing hypotheses

Capacity limitation theories have proved to be surprisingly resilient in characterizing some of the cognitive deficits in schizophrenia. However, this perspective has not generally been applied to short-term verbal memory tasks. We explored this issue by first attempting to ascertain if gross misallocations of processing resources might explain impairments in short-term memory in schizophrenia on a classic digit span task and in a second study by attempting to determine what effects delay and memory set size had on a divided attention short-term verbal memory paradigm.

In the first study 16 patients with schizophrenia and 21 normal controls received 40 trials of a three digit task and 20 trials of a six digit span task. As the absolute number of digits presented and duration of presentation in two conditions were identical, subjects thus had equivalent 'opportunities' to make errors if distraction, in the sense of misallocation of cognitive resources, were at the root of poor performance. In the second study 15 patients with schizophrenia and 15 normal controls were tested in conditions in which two, four or six words were presented and in which rehearsal was prevented by an interference task (colour naming) for delays of 5, 10 or 15 s.

Patients had disproportionate difficulty on the six digit rather than the three digit condition, suggesting that deficits in the verbal working memory short-term store may not be the result of attentional factors. In the second study, patients' performance was differentially worsened by the interference task, by memory set size (i.e. a capacity limitation) and by delay, a measure of decay rate.

In concert, these studies demonstrate that schizophrenia patients have difficulties on verbal short-term memory span tasks not because of misallocation of resources, but rather because of limitations in 'representational capacity' and maintenance of information over delays.