IRRATIONALITY LAB
WHAT?
How do people sample and weight information when they make decisions? We ask this question for both simple (e.g., perceptual) and complex (e.g., multiattribute, risky) decisions.
WHY?
Human decisions exhibit intriguing patterns –including biases and irrational preference reversals – that are barely understood in mechanistic terms. We aspire to do so at various levels (computational, algorithmic, neural) and to provide neurophysiologically grounded insights on human decision-making.
Decisions entail complex information and engage multiple cognitive and neural processes. Thus, understanding human decisions in information processing terms is elusive. This is why we study human decisions using elaborate experimental techniques (such as cognitive psychophysics, pharmacological manipulations & MEG) that allow us to trace the flow of decision-related quantities, and the transformations they undergo, as decisions take shape. We combine experimental data with computational modelling to unveil the general principles governing human decisions.
HOW?
FOR WHOM?
We aspire to make our curiosity-driven, blue-skies research societally relevant. We try to address big open questions in the cognitive & decision sciences at multiple levels: from the level of neurotransmitters, to large-scale brain networks, to cognitive mechanisms, to behaviour. Our insights can inform clinical practice, by shedding new light on how information processing & decision behaviour go awry in neuropsychiatric disorders; and help revolutionise applied behavioural science by offering mechanistic lenses to understand, predict and foster behavioural change.
VISION
How can we uncover what people truly value? Directly asking people to disclose what they like falls short because introspection is poor. Also, reading people's preferences from overt choices is undetermined, much like reading out the temperature of a room from the on/ off state of a thermostat.
We propose a new approach based on our observation that people sample pieces of information differently depending on the value they assign to them. In the next 5 years, in the context of an ERC Consolidator grant, we will use MEG to decode snapshots of people's preferences from subtle patterns in their information sampling. This will enables us to map out the mechanisms and long-term dynamics of preference formation, perseverance and change.