Objectives: These hypotheses were tested by investigating the time course of activation and selection of homonym meanings in single-word and sentence context using cross-modal priming. Children with high functioning autism (HFA; n=23), and chronological (CA; n=23) and verbal age (VA; n=23) matched controls participated.  

Methods: In Experiment 1 children heard homonym primes (BANK) and unrelated primes (CAKE) and named dominant (MONEY) or subordinate (RIVER) picture targets. Experiment 2 investigated sensitivity to biasing context in facilitating appropriate meanings; children heard neutral (Helen walked towards the bank) or biased (Harry fished from the bank) sentences and named appropriate picture targets (RIVER). To measure inappropriate activation, children heard inappropriately related (Helen fished from the bank) or unrelated control sentences (Harry fished from the stream) and named inappropriate targets (MONEY).

Results: Experiment 1 revealed priming effects for dominant targets at 250ms and 1000ms ISI but subordinate priming only at 250ms ISI. There were no Group interactions. In Experiment 2 the control groups showed appropriate facilitation at both ISIs; the HFA group only demonstrated appropriate facilitation at 250ms ISI. VA controls and the HFA group showed inappropriate facilitation at 250ms; only the HFA group showed inappropriate facilitation at 1000ms.

Conclusions: Children with HFA showed similar single-word semantic processing to controls. They showed sensitivity to context to activate appropriate meanings but this diminished over time, suggesting a difficulty with maintaining appropriate activation. Only the HFA group showed inappropriate meaning activation after 1000ms, suggestive of inefficient suppression. The extent to which these results can be attributed to weak central coherence and/or an inefficient suppression mechanism will be discussed.