Repetition priming and hyperpriming in semantic dementia

Evidence from neurologically normal subjects suggests that repetition priming (RP) is independent of semantic processing. Therefore, we may expect patients with a selective deficit to conceptual knowledge to exhibit RP for words regardless of the integrity of their semantic representations. We tested six patients with semantic dementia (SD) on a lexical decision task that incorporated four different lags between first (baseline) and second (primed) presentation of repeated words. The patients exhibited significant RP that was greater for words that were categorised as semantically ‘degraded’ than for words categorised as ‘known.’ This RP advantage for semantically degraded words declined as lag increased. The patients also demonstrated hyperpriming, and a significant correlation was identified between baseline response time and RP in SD but not in controls. These findings indicate that level of semantic knowledge about a word influences both baseline lexical decision performance and RP of that word. The observed hyperpriming can be parsimoniously explained by a cognitive slowing account.