Background. While some cognitive accounts of disorganized speech, or thought disorder, in schizophrenia have emphasized failures in working memory/discourse planning or selective attention, we have suggested that thought disorder resides in the semantic system. In this study we assessed the effect of neuroleptic medication on thought disorder and semantic processing.

Methods. Seventeen patients with schizophrenia were assessed while receiving neuroleptic medications and in crossover fashion, placebo. A number of measures were obtained: clinically rated thought disorder (using the Thought, Language and Communication Scale); working memory (letter number span); lexical integrity (naming and receptive vocabulary); and, semantic priming of intracategorical word pairs.

Results. Semantic priming measures improved with neuroleptic medication, as did clinically rated thought disorder. No other measure changed significantly. Priming selectively covaried with changes in thought disorder.

Conclusion. Changes in spreading semantic activation, measured in a semantic priming paradigm and presumably brought about by neuroleptics' influence on dopaminergic neuromodulatory systems, might reflect changes in the biases of pre-existing associative networks that favour or increase the accessibility of representations related by shared features. This study also has implications for the architecture of normal language in that a dissociation between the lexical and semantic levels was observed, due to the selective compromise of tasks demanding semantic processing.