Multisensory integration, perception and ecological validity.

Studies of multimodal integration have relied to a large extent on conflict situations, in which two sensory modalities receive incongruent data concerning one aspect of the source. Exposure to such situations produces immediate crossmodal biases as well as longer lasting aftereffects, revealing recalibrations of data-to-percept matches. In the natural environment, such phenomena might be adaptive, by reducing the perturbing effects of factors like noise or growth-induced changes in receptor organs, and by enriching the percept. However, experimental results generalize to real life only when they reflect automatic perceptual processes, and not response strategies adopted to satisfy the particular demands of laboratory tasks. Here, we focus on this issue and review ways of addressing it that have been developed recently.