Functional magnetic resonance imaging was used to examine the neural correlates of two linguistic tasks in schizophrenia.

Five dextral male schizophrenic patients and five volunteers matched for demographic variables and task performance participated. Echoplanar images were acquired over 5min at 1.5T while subjects performed two paced, covert tasks; (1) verbal fluency: silent generation of words beginning with an aurally presented cue letter, contrasted with silent repetition of the aurally presented word `rest'; (2) semantic decision: deciding whether a visually presented cue word was `living or non-living' and silently articulating the response, contrasted with rest. Both tasks entailed language processing; only verbal fluency requires the intrinsic generation of verbal material. Between-group differences in the mean power of experimental response to the semantic decision task were identified by a one-way analysis of covariance (ANCOVA), with a measure of stimulus-correlated motion as a covariate. Voxels demonstrating a significant interaction between task and group were identified using a two-way ANCOVA.

In controls, both tasks were associated with activation of prefrontal cortex. In patients with schizophrenia there was a significantly reduced power of response in several prefrontal regions during verbal fluency relative to controls, a difference that was not evident for the semantic decision task. There was a significant group×task interaction in the left inferior frontal gyrus, left dorsolateral prefrontal cortex and the supplementary motor area at voxel and regional levels of analysis.

Attenuation of frontal activation during cognitive task performance in schizophrenia does not represent a fixed deficit in frontal function, but may depend on the specific cognitive demands of the experimental task employed.