Deficits of semantic control produce absent or reverse frequency effects in comprehension: evidence from neuropsychology and dual task methodology.

Patients with multimodal semantic impairment following stroke (referred to here as 'semantic aphasia' or SA) fail to show the standard effects of frequency in comprehension tasks. Instead, they show absent or even reverse frequency effects: i.e., better understanding of less common words. In addition, SA is associated with poor regulatory control of semantic processing and executive deficits. We used a synonym judgement task to investigate the possibility that the normal processing advantage for high frequency (HF) words fails to emerge in these patients because HF items place greater demands on executive control. In the first part of this study, SA patients showed better performance on more imageable as opposed to abstract items, but minimal or reverse frequency effects in the same task, and these negative effects of word frequency on comprehension were related to the degree of executive impairment. Ratings from healthy subjects indicated that it was easier to establish potential semantic associations between probe and distracter words for HF trials, suggesting that reverse frequency effects might reflect a failure to suppress spurious associations between HF probes and distracters. In a subsequent experiment, the aphasic patients' performance improved when HF probes and targets were presented alongside low frequency distracters, supporting this hypothesis. An additional study with healthy participants used dual task methodology to examine the impact of divided attention on synonym judgement. Although frequently encountered words were processed more efficiently overall, the secondary task selectively disrupted performance for high but not low frequency trials. Taken together, these results show that positive effects of frequency are counteracted in SA by increases in semantic control requirements for HF words.