Four notes
Four project notes.
Links below lead to the corresponding project.
01 · Neural engineering · Experimental design
BCI calibration: interface effects
Sensors are not the only thing shaping a measurement. Instructions, feedback, pacing, fatigue, attention, and task effort can change the conditions under which data are collected.
That was the challenge in my BCI calibration work. We could make a repetitive task easier to stay with, but we could not quietly change what the experiment measured. The case study describes preliminary engagement and threshold results in three participants. The more-than-940-session count belongs to my broader RNEL calibration work.
Takeaway: Decide what must stay fixed before redesigning the task. Then improve the experience around that constraint.
Open the BCI calibration case study →
02 · Research software · Reasoning
ReasonOS: preserving rejected paths
If a system stores only the final answer, you lose the record of what changed and why. That matters when someone needs to review the reasoning later.
ReasonOS keeps accepted and rejected steps separate. A path you ruled out can stay in the record without being mistaken for the answer.
Takeaway: Keep the steps you ruled out, but label them clearly. The history should help a reviewer, not create another puzzle.
Open Vector EKG + ReasonOS →
03 · Motorsport safety · Human factors
Motorsport handoff: facts vs interpretation
A handoff gets muddy when observations, interpretation, and decision authority are blended together. The next person should be able to tell what was seen, what was inferred, and who made the call.
My motorsport neurotrauma draft separates crash mechanics, occupant-protection context, neurological observations, disposition, and the boundary between medical decisions and Race Control authority.
Takeaway: Record what was observed, what was inferred, what happened next, and who owns the next decision.
Open the neurotrauma toolkit →
04 · Interactive anatomy · Spatial interfaces
HoloAnatomy: spatial context
A labeled structure can still be hard to understand when its neighbors disappear. In anatomy, the relationships around a structure are often part of what makes it make sense.
HoloAnatomy lets you rotate anatomy models, separate structures, and inspect their neighbors. It is an interactive learning exhibit; I have not tested whether it improves learning outcomes.
Takeaway: Let people pull structures apart without losing the context around them, then move straight into identification and recall.
Explore HoloAnatomy →