Discrimination learning, reversal, and set-shifting in first-episode schizophrenia: stability over six years and specific associations with medication type and disorganization syndrome

Abstract

Background: The intradimensional/extradimensional (IDED) task assesses different forms of learning from feedback. Limited evidence suggests that attentional set-shifting deteriorates over time in schizophrenia. We tested this hypothesis and examined the specificity of learning impairments identified by this task.

Method: Two hundred sixty-two first-episode patients and 76 healthy control subjects, matched for age and premorbid IQ, were tested; 104 patients and 25 control subjects were reassessed 1 and 3 years later, and 31 patients were reassessed additionally 6 years later.

Results: Patients showed impaired set-shifting that correlated with current IQ and working memory, but there were no impairments when subgroups were matched on current IQ. In contrast, patients showed marked impairments in rule reversal learning that survived correction for IQ, were present in the context of intact rule abstraction, and correlated with disorganization symptoms. Patients prescribed second-generation antipsychotics were worse on set-shifting compared with first-generation, a finding not explained by demographic data, illness characteristics, or IQ. Patients and control subjects showed stable IDED performance over the first 6 years of illness, although set-shifting was inconsistent over the first year. Those with residual negative symptoms were more likely to fail the set-shifting stage at follow-up.

Conclusions: First-episode schizophrenia patients can learn and generalize rules but are inflexible when rules change, reflecting reduced responsiveness to negative feedback and difficulty in switching attention. Rule-reversal is a promising target for translational studies, because it is specific, clinically relevant, and might reflect orbitofrontal dysfunction. Set-shifting is related to poor function more generally but might be sensitive to medication effects and valuable for clinical trials.

Figure 1

Baseline attentional set shifting task performance. Upper panel: for patients and control subjects, mean errors with standard error bars are shown on the y axis for each stage denoted along the x axis. For each stage, Kruskal–Wallis comparisons are given. Lower panel: on the left the cumulative pass rate (%) is given on the y axis for each stage denoted along the x axis. For each stage this refers to the total number of patients and control subjects who have passed up until and including the current stage. On the right, the pass rate for the individual stage (%) is given on the x axis for each stage denoted on the y axis. This refers to the number of patients and control subjects who actually attempted that particular stage, having passed the previous stage. SD, simple discrimination; SR, simple reversal; C_D, compound discrimination; CD, compound discrimination; CR, compound reversal; IDS, intradimensional shift; IDR, intradimensional reversal; EDS, extradimensional shift; EDR, extradimensional reversal.

Figure 2

The upper panel shows cumulative pass rates for patients and control subjects at baseline, 1 year follow-up, and 3-year follow-up for those subjects who completed all three assessments. This denotes the percentage of subjects passing up to and including the current stage. The lower panel shows the number of errors committed at each stage at the three time points for patients and control subjects. Abbreviations as in Figure 1.

Figure 3

Cumulative pass rates by stage at baseline, 1 year, 3-year, and 6-year follow-up assessments in the subset of patients who had completed all four assessments (n = 31). This denotes the % subjects passing up to and including the current stage. Abbreviations as in Figure 1.