Thesis defence_ Jeremy Badler

In order to overcome processing delays, the brain is able to generate predictive movements. However, for such movements to be effective the brain must also anticipate the future state of the world based on relatively simple rules. Two such heuristics were examined: the use of regular time intervals to predict periodic events, and the tendency of interacting objects to behave in a causally plausible way. Eye movements of human observers were recorded while viewing stimuli that either moved after a predictable time delay, or that moved in concordance with or in violation of the rules of causality. For some tasks, manual responses were recorded as well. The following effects were observed: for the time experiment, when a subject had to respond with both an eye movement and a hand  movement, the responses occurred later than when either modality was tested in isolation. It is evidence that simultaneous, separate motor responses do not have free access to a common timing system; each required response drains attentional resources, even if there is but a single interval to be timed. For the causality experiments, subjects consistently made predictive eye movements in the causally-plausible direction of target motion, even though the target moved implausibly 50% of the time. It shows that causality is indeed a heuristic that can serve as a basis for predictive movements. For the same stimulus, the     perception of causality was sensitive to specific physical parameters and varied according to the individual subject and prior training, while the predictive movements were less sensitive to individual stimulus characteristics. Overall, heuristics appear to be based on information integrated over multiple stimuli, designed to produce an optimized response that minimizes average error.

More details on his researchwill be exposed on Wednesday 7 December at 5 pm at Auditoire de Visscher