SPECIAL REPORT Speeded responses to audiovisual signal changes result from bimodal integration ERICH SCHRÖGER and ANDREAS WIDMANN Institut für Allgemeine Psychologie der Universität Leipzig, Germany Abstract Integration of auditory and visual information was studied in humans detecting targets ~i.e., location changes of the auditory, the visual, or both parts of a repetitively presented audiovisual stimulus!. Behavioral results suggest that the time advantage to bimodal compared with unimodal targets was due to combined rather than separate processing of the auditory and the visual target information. Event-related brain potential results revealed strong audiovisual inter- actions supporting interactive and not independent coactivation models. The time course of this interaction suggests that the audiovisual integration occurred after low-level, sensory processing but well before the execution of the motor response. Descriptors: Audiovisual integration, Redundant target effect, Crossmodal interaction Crossmodal interactions have been investigated extensively using behavioral and physiological measures ~for overviews see e.g., Stein & Meredith, 1993; Welch & Warren, 1986!. However, there are few studies using a combined approach to uncover the inter- action between different modalities in humans. The present study aimed at clarifying audiovisual integration processes occurring in a so-called redundant target paradigm ~ Miller, 1982!. The bimodal redundant target effect consists of speeded button-press responses in the case of a bimodal ~audiovisual! target compared with the responses to the corresponding unimodal targets ~only auditory or only visual!. Race models ~ Raab, 1962! postulate that the motor response to bimodal targets is triggered by the faster of two sep- arate processes, one detecting the auditory and the other detecting the visual part of the target. As a consequence of this separate activation, reaction times are shorter to bimodal than to unimodal target signals. However, analyses of the probability distributions of the reaction times usually reveal that the speeded motor response to redundant targets cannot be explained by race models ~ Miller, 1986, 1991!. Although separate activation can often be excluded as an ex- planation for the redundant target effect, it is unclear, however, which kind of combination of the auditory and the visual signal information takes place. Miller ~1991! proposed that the combina- tion could be either a summation of independently produced acti- vations or an interaction of the processes involved in detecting the redundant targets ~independent vs. interactive coactivation!. That is, with independent coactivation, the target information from both modalities is processed separately from each other; only the output of these processes is accumulated, and this combined activation underlies the triggering of the motor response. With interactive coactivation, the processing of target information of one modality is affected by the presence or absence of target information of the other modality. Provided that the redundant target effect is due to an interaction of auditory and visual information, the question arises: at what level of processing does this interaction occur? The finding of an increase in response force with bimodal as compared with uni- modal target signals is compatible with the hypothesis of coacti- vation at late, low-level motor stages ~Giray & Ulrich, 1993!. However, the recording of the lateralized readiness potential ~ LRP; an electrophysiological indicator of hand-specific response prep- aration; e.g., Coles, 1989! in a unimodal ~visual! version of the redundant target paradigm ~ Mordkoff, Miller, & Roch, 1996!, and in a related event-related potential ~ ERP! experiment in which an accessory auditory stimulus was completely task-irrelevant ~ Hack- ley & Valle-Inclán, 1998!, produced no evidence for the motor coactivation hypothesis. On the contrary, Miniussi, Girelli, and Marzi ~1998! found evidence for the hypothesis that the visual redundant target effect occurs at a sensory level, more specifically, in the extrastriate visual cortex. Moreover, recent electroencepha- lographic and magnetoencephalographic studies ~Giard & Peron- net, 1997; Sams et al., 1991! found evidence for audiovisual integration at early, sensory-specific levels. In the study of Giard and Peronnet ~1997!, subjects had to perform a two-alternative- This research was supported by the Deutsche Forschungsgemeinschaft ~ DFG!. We thank H. Leuthold, H.J. Müller, and R. Ulrich for helpful comments on this manuscript. Address reprint requests to: Erich Schröger, Institut für Allgemeine Psychologie, Universität Leipzig, Seeburgstr. 14-20, 04103 Leipzig. Ger- many. E-mail: schroger@rz.uni-leipzig.de. Psychophysiology, 35 ~1998!, 755–759. Cambridge University Press. Printed in the USA. Copyright © 1998 Society for Psychophysiological Research 755