Recently it has been shown that the presentation of accessory auditory stimuli reduces reaction times of saccadic eye movements towards visual target stimuli. The magnitude of reduction depends on the spatio-temporal relationship between visual target and auditory non-target (Frens et al., 1995). Psychological and physiological experiments give evidence for convergence and common processing of visual and auditory sensory information in central processing stages (Hughes et al., 1994, Stein and Meredith, 1993).

In this complex system changes of stimulus intensities may influence processing time on different levels:

1. Changes in stimulus intensity could evoke changes in the arousal state of the organism: Non-target stimuli would act as a general warning signal independent of spatial stimulus arrangement.
2. It is known that peripheral processing time decreases with increasing stimulus intensity. Thus the effect of stimulus intensity may be mediated completely by changes in peripheral processing time. Colonius & Arndt (1998) developed a model of visual-auditory interaction proposing such a mechanism. In this model stimulus intensity influences reaction time by changing the probability of visual-auditory interaction. Thus, processing of the combined bimodal information does not dependent directly on stimulus intensity.
3. Stimulus intensity could modify the shared processes directly.

In the experiments reported here the intensity of the auditory stimulus is systematically changed in steps of 3 dB SPL while visual stimulus intensity is kept constant (10 cd/m²). Five levels of auditory stimulus intensity were applied in separate blocks. Six different spatial combinations of visual and auditory stimuli were tested in a step paradigm. The subjects' task was to redirect the eyes from the fixation point to the visual stimulus as fast and as accurate as possible and to ignore the auditory stimulus.

The data suggest that the auditory stimulus does not act as a mere warning signal. The amount of reduction of reaction time depends critically on the spatial stimulus arrangement. As expected, mean reaction times decreased with increasing stimulus intensity. The effect was significant for intensity differences of 9 -12 dB SPL. The model of Colonius & Arndt allowed to fit the mean reaction times and resulted in physiologically plausible estimated processing times for peripheral and central processing stages. A direct influence of stimulus intensity on central processing times is not necessary to account for the data.