Postersession 3
Poster #: 81
Topic: MMN across modalities
Friday, Sep 11, 2015
15:30-17:00
1st floor
Prediction of vision from invisible stimuli
1Institut of Psychology, University of Leipzig, Leipzig, Germany
2South-West University, Coffs Harbour, Australia
3Institute of Psychology, University of Leipzig, Leipzig, Germany
4Institute for Psychology, University of Leipzig, Leipzig, Germany
5Southern Cross University, Coffs Harbour, Australia
urte.roeber@uni-leipzig.de
The human brain establishes predictive models encoding regularities in sensory input. For example, if we are stopped in a car at a traffic light and the indicator light of the car in front of us is blinking regularly, we form the prediction that it will continue to exist and to blink in the same way. Accordingly, we are not distracted by each blink of the indicator light, and we are able to attend to something else, such as a pedestrian crossing the road. However, when a prediction is violated (e.g., the indicator skips a blink), the predictive model has to be updated. An essential component of predictive models for visual information processing is that predictions are made even when objects are not consciously experienced (proto-objects). We review studies showing that the mismatch negativity (MMN; a well-established brain signature of prediction and prediction-error) can be elicited by prediction-violating stimuli that are invisible from binocular rivalry suppression. The MMN is essentially identical to that when the identical stimulus is visible during episodes of binocular rivalry dominance. This suggests that predictive models for visual information processing are established, tested, and updated similarly for objects (visible) and for proto-objects (invisible).