Bug fixes on the Bilateral Coordination Test. Implemented the screen split — left-side dots constrained to the left half of the screen, right-side dots to the right half. Built a comprehensive results screen with timeline visualization, cadence analysis, and fatigue index. Set the test duration to 60 seconds (rejecting 90 seconds).
Screen split decision: The original prototype had dots spawning anywhere on the screen — which meant users sometimes crossed their hands to reach dots, confounding the bilateral measurement.
Splitting the screen so left dots stay left and right dots stay right wasn't a UX improvement — it was a measurement validity fix. If the right hand can collect left-side dots, the symmetry metric is no longer measuring hand assignment; it's measuring whoever was faster. The screen split forces each hand to stay in its domain. This is the kind of constraint that looks arbitrary in a UI but is load-bearing for the test's meaning.
Results screen scope: Built a comprehensive results screen with a timeline of dot collection events, cadence analysis (dots per 10-second interval), and fatigue index (performance drop from first third to last third).
AI suggested 90 seconds to give users more time and capture more data points. I set it to 60 seconds. Reason: 90 seconds introduces fatigue as a confounding variable — the fatigue index at 90s would be measuring endurance, not coordination. At 60 seconds, fatigue is a secondary factor, not the primary driver of score differences. The test should measure coordination, and the duration should be calibrated to that, not to data volume.
Every design decision in an assessment has a measurement consequence. Screen layout, test duration, what counts as "correct" — these aren't just UX choices. They determine what the test is actually measuring. Getting these right requires thinking about validity first, then usability.
The results screen also showed what users need from an assessment vs. what they expect. Expected: a score. Need: context for the score — trend, comparison, breakdown. A number without context is just noise.
M3 is now complete with a fully functional Bilateral Coordination Test that produces measurements with clinical meaning — per-hand scores, symmetry ratio, cadence analysis over 60 seconds. The test uses the camera (everyday device) to measure something real about bilateral motor coordination. The quarter question is being answered.
M4 begins. Connect the homepage to Supabase backend. Build the remaining tests (Reaction Time, Rhythm Sync, Fine Motor). Add authentication and persistent data storage so results can be tracked over time.