Poster C27, Friday, August 17, 10:30 am – 12:15 pm, Room 2000AB
Lexical access in reading vs. naming: MEG evidence from semantic priming and inhibition
Julien Dirani1, Liina Pylkkänen1,2;1New York University Abu Dhabi, 2New York University
INTRODUCTION. There is general consensus that lexical access in comprehension involves spreading activation of semantically related nodes. Support for this comes from lexical decision tasks, showing that reaction times are facilitated by semantically related primes. Interestingly, semantic priming has an inhibitory effect on object naming. This is demonstrated by the Picture-Word Interference paradigm and is taken as evidence for a conceptual search that involves a competition among activated concepts. To date no study has systematically compared these different priming effects and their neural correlates. It is thus unclear whether semantic access proceeds entirely differently in comprehension and production, or whether, for example, a facilitatory effect of semantic priming is observed for both in an early time window, followed by a late inhibition in naming. METHOD. Our aim was to contrast Object Naming to a task that is maximally similar but does not include a process of conceptual search. We therefore used Overt Reading as a comprehension task in comparison to Object Naming as a production task, thus manipulating the modality of the targets to be either a word or an image. Both tasks involved three levels of word-primes (semantically related, unrelated, and identical, e.g., the primes TABLE/DOG/CHAIR were followed by “chair” or a picture of chair), with primes and targets presented using four different stimulus onset asynchronies (SOAs; 150, 200, 250, 300ms). Magnetoencephalography (MEG) assessed the timing and localization of the priming effects. Specifically, for naming, the interference effect could happen at the early conceptual level, or later on at the lexical level or at the motor preparation level. RESULTS. The behavioral results replicated the previously identified semantic inhibitory effects in Object Naming, and showed a trending facilitatory effect in Overt Reading. We found no interaction of SOA and Prime Type. MEG data for Overt Reading showed that the activation pattern in the superior temporal gyrus (STG) at 165-375ms matched the behavioural results, with lowest amplitudes for the identity primes, followed by the semantically related primes, and then the unrelated primes. In contrast, in Object Naming, rather than following a stepwise effect of priming, the activation in the STG showed that the semantically related condition diverges from the other two at 300-500ms, suggesting a late interference effect. Crucially, this effect goes in opposite direction from the effect of the identity priming. Further, this interference pattern spreads to more posterior regions, in the inferior parietal lobe (IPL). CONCLUSION. The results present direct evidence supporting the conceptually driven, competitive process involved in object naming and the incremental, spreading activation in word comprehension. The opposing semantic priming effects in object naming and overt reading appear to partially co-localize within the STG. Specifically, the early facilitatory effect in reading is focused to the middle part of the STG as opposed to the later interference effect for naming, which covered the full STG and spread onto the IPL. This late interference effect presents evidence that the locus of semantic interference is not at the early conceptual level, but rather at the later lexical access level.
Topic Area: Meaning: Lexical Semantics