Piai, V. 1, 2 & Roelofs, A. 1
1 Donders Institute for Brain, Cognition and Behaviour. Radboud University Nijmegen. Nijmegen, Netherlands.
2 International Max Planck Research School for Language Sciences. Nijmegen, Netherlands.
Picture-word interference is a widely employed experimental paradigm to investigate naming performance: Speakers name pictures while trying to ignore superimposed distractor words. The distractor can be congruent to the picture (pictured cat, word cat), categorically related (pictured cat, word dog), or unrelated (pictured cat, word pen). Categorically related distractors slow down picture naming relative to unrelated distractors, an effect called semantic interference. Furthermore, categorically related distractors slow down picture naming relative to congruent distractors. This finding is analogous to findings in the colour-word Stroop task, in which colour-naming responses in the categorically related condition (blue printed in red ink) are slower relative to the congruent condition (blue printed in blue ink). The locus of semantic interference and Stroop-like effects in naming performance has recently become a topic of much debate. Whereas some researchers argue for a pre-lexical locus of semantic interference and a lexical locus of Stroop-like effects (e.g., Dell‘Acqua et al., 2007), others localize both effects at the lexical selection stage (e.g., Roelofs, 2003). We investigated the time course of semantic and Stroop-like interference effects in overt picture naming by means of event-related potentials (ERP) and time-frequency analyses. Moreover, we employed cluster-based permutation tests for the statistical analyses (Maris & Oostenveld, 2007). Naming latencies showed the expected semantic and Stroop-like interference effects. The ERP waveforms for congruent stimuli started diverging statistically from categorically related stimuli around 250 ms. Deflections for the categorically related condition were more negative-going than for the congruent condition (the Stroop-like effect). The time-frequency analysis revealed power increase in the beta band (12-30 Hz) for categorically related relative to unrelated stimuli roughly between 250 and 370 ms (the semantic effect). The common time window and absolute timing of these effects suggests that both semantic interference and Stroop-like effects emerged during lexical selection.