Late cross-modal biases in neural spatial representations revealed by EEG decoding

What the study found

Early auditory and visual spatial representations stayed independent. A shared spatial representation emerged at about 200 ms, after which each modality was pulled toward the other, auditory first and visual later. Behaviorally, people integrated the cues with slight visual over-weighting.

Limitations

Simple clicks and flashes at horizontal offsets; online ahead of print.

Neurodesign interpretation

The brain builds a 'where' from hearing and a 'where' from sight separately, then reconciles them about a fifth of a second later. Rooms that put those two answers far apart give it more to reconcile. Whether that costs anything measurable in a real space is open.

Possible design implications

Hypothesis: spaces where sound and its visible source disagree (announcements from hidden speakers, reflections that relocate a voice) ask the brain to reconcile two separate estimates. Test before stating.

Why it stays in the corpus

Retain as the anchor for a sensory spatial agreement concept: where a sound appears to come from versus where its visual source is.

Study design

EEG with cross-generalised inverted encoding models

Population

Adults of both sexes (sample size not extracted from full text)

Setting

Laboratory audiovisual localisation task

Field

Multisensory perception

Mechanism

Audiovisual spatial integration, ventriloquist effect, recurrent feedback

Spatial variable

Spatial agreement between sound source and visual source

Keywords

multisensory, audiovisual, ventriloquist effect, spatial localisation, EEG decoding, cross-modal