Prediction of changes in the environment is a fundamental adaptive property of the brain. To do so, cognitive and neural mechanisms must activate in memory potential future stimuli on the basis of preceding ones. This basic function is involved in a large variety of cognitive activities ranging from semantic and syntactic processing during language comprehension to decision-making in cognitive economics. Dysfunctions of such prediction processes are also reported to generate psychiatric symptoms in early Alzheimer's disease and formal thought disorders in schizophrenia.

Researchers from Université Côte d’Azur want to gain a better understanding of the process of prediction through complementary experimental approaches: behavioral, eye-tracking, EEG experiments... (and even open-brain experiments in collaboration with neurosurgeons) in situations of language comprehension and/or decision-making.

Most of these experiments will be conducted thanks to the CoCoLab experimental platform, which in particular aims at collecting physiological data such as skin conductance or HRV as well as facial expression recognition or EEG. Such an impressive collection of data is useful for understanding the prediction process only if the tools for finding pertinent and relevant information exist. This is why new statistical methods are found to analyze data without averaging over trials, or modify the experiment on-line to access more relevant features and link neuronal models to behavioral data and EEG-recorded evoked potential. Moreover, these data and their analysis with respect to suitable underlying mathematical models help to provide more accurate foundations for economic behavior and therefore improve the predictive power of economic theory as well as social efficiency through appropriate policy design.