Claim: In post-viral syndromes, short-term instability (variance) of nocturnal RMSSD across a 7-day rolling window is a stronger predictor of next-day cognitive fatigue and headache severity than absolute RMSSD level.

Rationale: Mean HRV can remain near-normal while autonomic control becomes unstable. That instability may reflect intermittent vagal withdrawal and cytokine-linked neuroimmune oscillations, which are more likely to precede symptom spikes.

Test design:

• Cohort: adults with persistent post-viral symptoms, 8-week wearable follow-up
• Exposure: rolling 7-day RMSSD coefficient of variation (CV)
• Outcomes: next-day symptom score (fatigue, brain fog, headache), optional CRP/IL-6 subset
• Model: mixed-effects regression comparing predictive value of RMSSD-CV vs mean RMSSD

Falsification criteria:

1. RMSSD-CV does not improve out-of-sample prediction versus mean RMSSD (ΔAUC <= 0.02).
2. Association disappears after controlling for sleep duration/fragmentation and activity load.
3. No temporal lead (variance fails to predict next-day symptoms).

Discussion question: Would adding overnight respiratory rate variability improve specificity for neuroinflammatory (vs psychosocial) flare prediction?