Procedural learning in schizophrenia: reconciling the discrepant findings.

BACKGROUND

Studies of procedural learning in schizophrenia have been inconsistent, sometimes finding it to be preserved and sometimes impaired. This study examined three factors that could account for the variability among findings: type of task, presence of general intellectual impairment, and the extrapyramidal side effects of neuroleptic treatment.

METHODS

Forty-three patients with schizophrenia and 22 normal control subjects were examined with three different paradigms: the pursuit rotor, mirror reading, and probabilistic learning ("weather prediction"). A subgroup of intellectually preserved patients was also examined. Patients with and without tardive dyskinesia and with and without Parkinsonism were also compared.

RESULTS

The schizophrenic patients showed learning comparable to the control subjects on the pursuit rotor and mirror reading but were impaired on the probabilistic learning task. However, this last difference disappeared when the subgroup of intellectually preserved patients was compared with a subgroup of matched control subjects. Patients with and without tardive dyskinesia or Parkinsonism showed similar learning on all three tasks, but patients with tardive dyskinesia showed poorer overall performance than those without.

CONCLUSIONS

Procedural learning tends to be preserved in schizophrenia, and when impairment is found, differences in the overall level of intellectual function might be the determining factor.