Our lab uses computational models to explain human learning, planning and how they are altered in psychopathology. Located in Bar-Ilan’s brain center, we’re investigating how trait anxiety biases planning content, meta-control, and efficiency. To do so we use online tasks, advanced virtual reality, eye-tracking, and MEG.
The brain’s internal model of how things work: causal knowledge about the self and the world, and what is worth pursuing. It is what lets us plan, imagine and learn without having to act first.
Transition priors about oneself: what one can do and what one’s actions bring about — how controllable outcomes are, how one copes, how one is seen by others.
Transition priors about the environment: expectations about how situations unfold — what follows what, how likely harm is, and how long it lasts.
What matters at the deepest level — safety, connection, competence, understanding. They set the rewards that every task is evaluated against.
The specific model built for the situation at hand. It inherits the priors and goals above it only to the extent that the task’s instructions and semantics are similar to them: similarity sets the weighting between a default representation, informed mostly by the instructions, and the prior. So a bias higher up — overstating threat or understating control, as in anxiety — reshapes most the tasks that most resemble it.