Testing multisensory interactions with event-related potentials. How to avoid artifacts of modality shifts.
Poster Presentation

Matthias Gondan
Department of Experimental Psychology, University of Regensburg

Mark W. Greenlee
Department of Experimental Psychology, University of Regensburg

     Abstract ID Number: 97
     Full text: Not available
     Last modified: March 21, 2005

Abstract
A frequent approach to study interactions of the auditory and the visual system is to measure event-related potentials (ERPs) to auditory, visual, and auditory-visual stimuli (A, V, AV). If AV–(A+V) differs from zero, it is concluded that the senses interact at a specific processing stage. Usually, stimuli are presented in randomized order, and modality shifting effects (MSEs) have been ignored up to now: For example, the auditory N1 is increased if the auditory stimulus follows a visual stimulus (crossmodal stimulus), whereas it is decreased if the auditory stimulus follows another auditory stimulus (ipsimodal stimulus). Bimodal stimuli are not affected by modality shifts, because at least one stimulus component is always identical to the preceding stimulus. This fact might lead to an apparent multisensory interaction in the result of AV–(A+V), which manifests as a modulation of unisensory ERP components (e.g. a decrease of the auditory N1).
We tested the influence of MSEs on auditory-visual interactions by comparing the results of AV–(A+V) using the entire stimulus set (conventional analysis) and using only ipsimodal stimuli. Significant MSEs were observed. Unexpectedly, the results of the new analysis largely converge with those of the conventional analysis.