Online colloquium by Dr. Redmond O'Connell
When: 27th 05, 2020, 18:00 CET
Zoom link for the meeting: https://lmu-munich.zoom.us/j/91587030299
Meeting ID: 915-8703-0299
Bridging neural and computational viewpoints on decision making
Decades of mathematical psychology research have yielded a powerful set of computational models for explaining how perceptual decisions are forged in the brain. These 'sequential sampling' models come in many forms which, thus far, have been evaluated based on their ability to quantitatively account for behavioural data. However, as recent controversies highlight, any two model variants can make highly similar behavioural predictions despite invoking fundamentally different algorithmic elements. This is of particular concern in light of the rapid increase in the adoption of sequential sampling models in psychiatry and in functional neuroimaging research where the choice of model to employ in a given study may have a major bearing on the results. These challenges can be potentially overcome by also considering the ability of a model to capture key observable aspects of the biological implementation of the decision process: where two models may produce the same behavioural trends, they frequently make distinguishable predictions regarding the associated decision-related neural dynamics. It is only recently that it has become possible to explore these possibilities using human data owing to the discovery of non-invasively recorded brain signals that reflect the key processing levels underpinning decision formation. I will present recent work from our lab demonstrating how these signals can be used to directly inform the construction and validation of mathematical decision models.