Wearable Biomarkers and Hormonal Tracking: HRV, rMSSD, Sleep Staging, and the ANS-HPG Interface

Heart rate variability measured as RMSSD reflects the beat-to-beat interval variability driven by cardiac vagal efferents (parasympathetic fibers via the vagus nerve). In the autonomic nervous system framework, high vagal tone (high RMSSD) indicates parasympathetic dominance — the physiological state associated with recovery, low sympathoadrenal activation, and low HPA output. Low RMSSD indicates sympathetic dominance — elevated norepinephrine (directly lowers LH pulsatility via alpha-adrenergic signaling at the hypothalamus) and elevated cortisol (reduces GnRH pulse frequency via hypothalamic glucocorticoid receptors and reduces Leydig cell StAR expression via glucocorticoid response elements). This mechanistic pathway — sympathetic dominance → cortisol and norepinephrine elevation → HPG axis suppression — explains the empirical correlation between low HRV and lower testosterone consistently observed in cross-sectional athlete and general population studies. The RMSSD-testosterone correlation is not universal in magnitude (r values range from 0.2 to 0.5 depending on cohort) but is mechanistically grounded and directionally consistent. Importantly, HRV is a leading indicator — RMSSD drops before cortisol-driven testosterone suppression is measurable in blood, making it a more sensitive early-warning tool than single-point hormone testing at low sampling frequency.

Oura Gen3 uses a combination of peripheral skin temperature (from the finger ring, proximal to digital arteries) and photoplethysmography (PPG, capturing heart rate and heart rate variability waveforms) as inputs for a multi-class machine learning sleep staging algorithm. Peripheral vasoconstriction during NREM sleep (particularly slow-wave sleep, SWS) produces characteristic skin temperature patterns — paradoxical finger temperature rise (reflecting peripheral heat dissipation from core cooling) accompanies SWS onset, providing a thermal signature that differentiates SWS from lighter NREM stages and REM. Oura's sleep staging algorithm achieves epoch-by-epoch agreement of approximately 70 to 80 percent with polysomnography (PSG) for SWS identification — comparable to or slightly below PSG-calibrated consumer EEG devices, but significantly better than wrist-based accelerometry alone. The clinical relevance for testosterone: SWS duration is the primary testosterone production window — specifically, pulsatile LH release events occurring during SWS in the first half of the night drive the majority of nightly testosterone production. Consistent SWS below 45 minutes is associated with blunted morning testosterone peaks in sleep laboratory studies.

The clinical significance of testosterone testing time is mechanistically established but often ignored in practice. Testosterone concentrations follow the SCN-GnRH-LH circadian rhythm with peak concentrations at approximately 8am and nadir concentrations at approximately 8pm — a 40 to 60 percent amplitude difference depending on age (amplitude decreases with age, as documented in the SBHS longitudinal cohort). The pre-dawn LH pulse cluster (3 to 6am) drives the morning peak; by 8pm, the LH interpulse interval has lengthened and testosterone has fallen toward nadir. Afternoon or evening blood draws systematically underestimate testosterone status relative to the reference morning peak. Wearables provide a proxy for testing timing quality: a wearable that shows good sleep quality (high SWS, stable HRV) the night before an 8am blood test suggests the pre-dawn LH pulse cluster was intact, producing a representative peak. Poor sleep the night before blunts the pre-dawn LH pulse and can produce a morning sample 10 to 20 percent below the individual's true peak — a difference large enough to affect clinical interpretation. Using wearable sleep quality data to contextualize testosterone blood test results improves interpretive accuracy.

Continuous glucose monitors (CGMs, eg. Dexcom G7, Libre 3) provide metabolic data directly relevant to testosterone through the fasting glucose-HOMA-IR-testosterone relationship. Insulin resistance (elevated HOMA-IR) suppresses sex hormone-binding globulin (through insulin-mediated SHBG hepatic suppression) and impairs Leydig cell steroidogenic function via intracellular insulin signaling pathways. Fasting glucose above 100 mg/dL and HOMA-IR above 2.5 are independently associated with lower free testosterone in multiple population studies. CGM-identified chronic postprandial hyperglycemia (repeated glucose excursions above 180 mg/dL) indicates insulin resistance not captured by fasting glucose alone. Looking forward, salivary testosterone monitoring is technically advancing but not yet clinically deployed at consumer scale. Salivary testosterone correlates with free serum testosterone (r approximately 0.8 to 0.9), reflecting non-protein-bound testosterone that passively diffuses into saliva — a genuinely non-invasive and potentially high-frequency sampling matrix. Lateral flow assay formats for point-of-care salivary testosterone are in development (active patent filings from Caltech, UC San Diego, and commercial entities), and wearable electrochemical biosensors for interstitial fluid androgens represent an active research frontier. Clinical-grade continuous testosterone monitoring analogous to CGM for glucose is technically plausible within a 5 to 10 year horizon.