Modulation of early auditory processing by visual information: Prediction or bimodal integration?

What happens if a visual cue misleads auditory expectations? Previous studies revealed an early visuo-auditory incongruency effect, so-called incongruency response (IR) of the auditory event-related brain potential (ERP), occurring 100 ms after onset of the sound being incongruent to the preceding visual cue. So far, this effect has been ascribed to reflect the mismatch between auditory sensory expectation activated by visual predictive information and the actual sensory input. Thus, an IR should be confined to an asynchronous presentation of visual cue and sound. Alternatively, one could argue that frequently presented congruent visual-cue-sound combinations are integrated into a bimodal representation whereby violation of the visual-auditory relationship results in a bimodal feature mismatch (the IR should be obtained with asynchronous and with synchronous presentation). In an asynchronous condition, an either high-pitched or low-pitched sound was preceded by a visual note symbol presented above or below a fixation cross (90% congruent; 10% incongruent), while in a synchronous condition, both were presented simultaneously. High-pitched and low-pitched sounds were presented with different probabilities (83% vs. 17%) to form a strong association between bimodal stimuli. In both conditions, tones with pitch incongruent with the location of the note symbols elicited incongruency effects in the N2 and P3 ERPs; however, the IR was only elicited in the asynchronous condition. This finding supports the sensorial prediction error hypothesis stating that the amplitude of the auditory ERP 100 ms after sound onset is enhanced in response to unexpected compared with expected but otherwise identical sounds.