Dedifferentiation and substitute strategy: deconstructing the processing-speed impairment in schizophrenia

Abstract

Background: Recent research has identified impairment in processing speed, measured by the digit-symbol substitution task, as central to the cognitive deficit in schizophrenia. However, the underlying cognitive correlates of this impairment remain unknown.

Methods: A sample of cases (N=125) meeting DSM-IV criteria for schizophrenia and a sample of community controls (N=272) from the same geographical area completed a set of putative measures of processing-speed ability to which we implemented confirmatory factor and structural regression modelling in order to elucidate the latent structure of processing speed. Next, we tested the degree to which the structural and relational portions of the model were equal across groups.

Results: Processing-speed ability was best defined, in both controls and cases (χ(2)=38.59(26), p=0.053), as a multidimensional cognitive ability consisting of three latent factors comprising: psychomotor speed, sequencing and shifting, and verbal fluency. However, cases exhibited dedifferentiation (i.e., markedly stronger inter-correlations between factors; χ(2)=59.94(29), p<.01) and a reliance on an alternative ensemble of cognitive operations to controls when completing the digit-symbol substitution task.

Conclusion: Dedifferentiation of processing-speed ability in schizophrenia and subsequent overreliance on alternative (and possibly less than optimal) cognitive operations underlies the marked deficit observed on the digit-symbol substitution task.

Fig. 1

A schematic of the postulated A. one-factor, B. two-factor, C. three-factor and D. structural models.

Fig. 2

Schematic of the multigroup comparisons.

Fig. 3

The final multidimensional model of processing speed. The measurement model is shown on the right in blue and the structural model is shown on the left in orange.