Project proposal · neuroscience
From a question
to a brain map.
A conversational atlas that keeps the evidence in view.
In design. The current viewer works; the question-to-map service is a separate project.

Start with a question
Keep the path visible
- Translate the question into an editable study query.
- Compute an evidence-derived map with a versioned method.
- Explore the overlay, contributing studies and uncertainty.
Working memory
Begin with a defined set of tasks and studies, then compare an established meta-analysis with the retrieved evidence. A statistical association map and a probability map need different labels and interpretations.
Comparing concepts
Show each map and its evidence separately before comparing overlap. A plausible-looking overlay is not enough to establish that two processes share a mechanism.
Context matters
Age, task design and contrast definitions belong in the query. When the evidence is too sparse, the system should say so.
The first version would use a common reference space. Overlaying a result on an individual scan requires a validated spatial transformation; it would not reveal that person’s thoughts or provide a diagnosis.