Long-term use needs periodic recalibration; a one-shot calibration does not last

What it is

The rest take or wizard task at unboxing only covers electrode geometry, skin state, and physiological set point as they were then. Overnight, a season, a medication change, a fitness change, and that template expires. A long-lived product has to make recalibration a periodic event, not assume the user will remain on day one.

Why it happens

Expiry has three clocks. Device: every re-don is a different placement, and the first calibration is bound to old geometry. Skin: stratum corneum, hair, and seasonal humidity change impedance. Body: training, illness, and hormonal cycles change the autonomic set point. BCI classifiers add a fourth: user strategy and cortical plasticity. A one-shot freeze-frame nails all four to install day. Unsupervised adaptation can delay expiry and can also absorb unlabeled error states. The cost of periodic calibration is interruption; the cost of not calibrating is silent failure. Both belong in the product lifetime model.

Studying it

Repeat the same task across days and weeks, and plot accuracy or error against days since last calibration. Factors: scheduled recalibration or not, duration, whether idle rest fragments can steal a background take. Outcomes: days to an unusable threshold, compliance with completing recalibration. Treating “willing to do the first wizard” as “willing to do it weekly” overstates long-run availability. Field logs see abandonment after the third time better than the lab.

Where it stops holding

A product that only shows a raw waveform and never emits a state label can skip calibration; the user reads it. Implant revision cycles are months to years and should not inherit a consumer watch’s weekly cadence. Growing children need a shorter period. Forced too-frequent calibration becomes an uninstall reason; period should follow drift speed, not a global once-a-week.

Applying it

• Give each physiological feature a calibration lifetime; when it expires, degrade output and request a short recalibration.
• Nest recalibration in an existing habit (thirty seconds after morning donning), not a separate laboratory-style wizard appointment.
• Allow skip, but skipping must lower confidence or stop adaptation—do not silently keep the expired template.
• Verify over at least two weeks of real wear, comparing error curves and completion for “first calibration only” versus “recalibrate on lifetime.”

Related

• Same group: C9.10.1 Resting physiological baselines differ across people; absolute values cannot be compared across users · C9.10.2 After calibration sets an individual baseline, affect, fatigue, and time drift the baseline itself · C9.10.4 Unrecognized baseline drift is misread as a state change
• Adjacent: C9.02 Brain-Computer Interfaces · C8.08 Eye-Tracking Calibration
• Search: recalibration · calibration half-life · session-to-session BCI