RESEARCH / AN INTERDISCIPLINARY APPROACH
Build the system.
Question the result.
Software engineering makes ideas testable. Psychology keeps the human context visible. Computational neuroscience offers models to challenge. Experimental AI connects these questions without removing their differences.
Observe.
Build.
Evaluate.
Systems people can inspect
Make tool behavior, permission boundaries and workflow state understandable.
PevraiStructure around human work
Explore how assignments and contextual feedback support a clear practitioner–client workflow.
Psychology platformModels with explicit assumptions
Test what neural activity contributes to a constrained task and what a readout actually learns.
Connectome researchTHE RESEARCH DISCIPLINE
Evidence has more than one level.
State the question
Define the behavior being investigated and the assumptions that make the model or workflow possible.
Separate the tests
A unit test verifies code behavior. A task evaluation measures performance. Neither automatically validates a clinical or biological claim.
Use meaningful controls
For neural experiments, compare learning and non-learning conditions, held-out inputs and appropriate graph or signal interventions.
Keep the limits visible
Report uncertainty, failed conditions and the distinction between a development goal and an observed capability.