Decisions under changing evidence

Staying open to new evidence

A driver approaching a moving hazard must respond to new information without reacting to every fluctuation. I ask how the brain balances that need to update with the growing pressure to act. Evidence-accumulation and urgency-based models can make similar predictions when evidence is stable; reversals offer a way to tell them apart.

My doctoral project focuses on three connected questions:

• How does urgency change the influence of new and past evidence?
• Why do people differ in how long they wait before committing?
• When does new evidence stop changing a choice, and can brain state shift that point?

From behavior to brain dynamics

I combine a tokens decision task recorded with MEG and a face-morph decision task recorded with EEG. In the first, evidence arrives over time and can reverse; in the second, a face gradually becomes happy or sad, sometimes with informative evidence arriving late.

I fit computational models to behavior and compare their predictions with MRI-guided source estimates, alpha and beta rhythms, low-dimensional neural trajectories, and coupling between cortical regions. Time-resolved decoding helps track the information available in brain activity. A central distinction is whether evidence is still represented in the brain or still influences the eventual choice.

I built the EEG/MEG source-analysis pipeline and develop reusable analysis tools for this work in the CoCo Lab and Mila, supervised by Prof. Karim Jerbi.

What I am testing next

The project tests how frontal, parietal, and sensorimotor dynamics change around commitment. Planned analyses use change-point and hidden-state models to investigate that transition. An exploratory direction asks whether near-critical brain dynamics—and, in a planned study, caffeine—shift how long a decision remains open.