In Part I, we followed the path from raw electrical activity to a clean list of R-R intervals. We filtered noise, identified each heartbeat, and validated that our detected peaks matched the heart’s true rhythm.
By the end, we had a trustworthy timeline of beats for our HRV analysis. With our signal cleaned and verified, we can now start generating HRV metrics and doing some analysis!
Companion Notebook can be found on Kaggle: Signal to Meaning (Part II): R-R Intervals to Physiological Insight
Low HRV, High Readiness — What’s Going On?
I bought my Oura Ring the moment I learned it tracked Heart Rate Variability (HRV). That was almost seven years ago. Over that time, I’ve watched my HRV gradually decrease — exactly as most research papers say it will with age[3]. And with almost a decades worth of data its clear that I am not an outlier.
HRV Declines with Age: Every major HRV measure — SDNN, RMSSD, pNN50, LF, HF, and LF/HF — shows a consistent downward trend across the lifespan[3].
But recently, I noticed a few things that I did not quite understand. My HRV had dipped below my normal and for a prolonged period of time. For reference your “normal” baseline is personal. It can’t be compared to anyone else’s number. But my Readiness scores were decent to good.
HRV is personal: Large differences exist between individuals, so meaningful interpretation comes from tracking your own trends rather than comparing to others[1] .
Additionally, Oura HRV Balance is a comparison of your 2-week HRV trend to your 3 month average. I started wondering if I had ignored or not paid attention long enough that my new 3 month trend was lower than it should be and not a reflection of healthy “balance” and “readiness”? With these observations, I found myself seeking a deeper understanding and wanting more insight into my HRV being provided to me by my Oura Ring. That also spark a call to action for HRV improvement!
RMSSD: It is Not the Only HRV Metric
When you open the Oura app and look under Readiness, the HRV value you see is calculated from your overnight RMSSD — the root mean square of successive differences between heartbeats while you sleep[1]. Or the average amount your heartbeat changes from one beat to the next while you sleep.
That single number is the foundation for Oura’s HRV metrics:
| Metric | What It Represents | How Oura Uses It |
|---|---|---|
| RMSSD (ms) | Beat-to-beat variability — parasympathetic activity | The core HRV number displayed in your app |
| Average HRV | Overnight mean of RMSSD | Shown in your “Readiness” tab each morning |
| HRV Balance | Trend of HRV relative to your personal baseline | Used to adjust your Readiness score |
Remember, HRV is a measure of the balance between your “fight-or-flight” (sympathetic) and “rest-and-recover” (parasympathetic) branches of the autonomic nervous system — the part working quietly behind the scenes to keep you in balance. A higher HRV usually signals a flexible, well-recovered system with strong vagal influence, while a lower HRV can point to stress, fatigue, or reduced adaptability. With HRV, balance means adaptability, not equality. It means the ability for your nervous system to shift smoothly between stress and recovery.
While RMSSD is a reliable window into parasympathetic (vagal) activity, it’s only one piece of the HRV story. It doesn’t reflect the broader autonomic balance between the body’s “fight-or-flight” (sympathetic) and “rest-and-recover” (parasympathetic) systems. In practice, two people can show identical RMSSD values yet have very different physiological profiles. One might have a well-regulated nervous system with high overall variability, while the other could be showing early signs of sympathetic dominance or reduced system complexity—patterns that RMSSD alone won’t capture[1],[2].
So while RMSSD provides a convenient overnight snapshot, it’s a simplified view of a much richer signal. To understand why HRV changes — not just how much — we have to look beyond a single number and explore the rest of the variability spectrum.
Beyond RMSSD: The Rest of the HRV Story
Oura doesn’t currently display other HRV indices such as SDNN, LF/HF, or entropy-based measures — at least not in the app interface.
That doesn’t mean they aren’t calculated internally; they just aren’t part of the user-facing experience.
For most users, that one overnight RMSSD trend is enough to show whether recovery is improving or declining. But from a physiological perspective, RMSSD is only the tip of the HRV iceberg.
RMSSD tells us how much your heart rate varies from one beat to the next — but HRV extends far beyond those rapid, short-term changes. The complete picture of variability includes slower rhythms, broader oscillations, and more complex patterns that together describe how your body manages stress and recovery across time[1],[2].
When I looked at my own data through Kubios HRV, it felt like I was stepping into a new dimension of information. Suddenly, I wasn’t just seeing a single number; I was seeing an entire system in motion.
Time-domain metrics such as SDNN, RMSSD, and pNN50 describe how much variation exists over different time scales. RMSSD focuses on fast, vagal-driven fluctuations, while SDNN captures total variability — including both sympathetic and parasympathetic influences — across longer windows[1].
Then there are the frequency-domain measures — HF, LF, and VLF — which separate variability into rhythm bands. The HF band reflects parasympathetic (vagal) influence tied to breathing. LF represents a mix of sympathetic and parasympathetic modulation, often linked to baroreflex control — the body’s built-in feedback system that keeps blood pressure stable. When blood pressure rises, stretch sensors (baroreceptors) in the arteries signal the heart to slow down; when it falls, they prompt it to speed up. This rhythmic push and pull creates a natural oscillation in heart rate, showing up as power in the LF band. And VLF, which appears only in longer recordings, captures slower oscillations driven by hormonal and thermoregulatory cycles[1],[2].
| Band | Range (Hz) | Dominant Influence | Typical Duration Needed |
|---|---|---|---|
| HF (High Frequency) | 0.15 – 0.40 | Vagal tone and respiration | 1–5 min |
| LF (Low Frequency) | 0.04 – 0.15 | Baroreflex + mixed autonomic input | ≥ 5 min |
| VLF (Very Low Frequency) | < 0.04 | Hormonal and thermoregulatory rhythms | ≥ 24 h |
These frequency components reveal the architecture of your autonomic regulation — not just how much variability you have, but where that variability lives.
Finally, nonlinear and complexity measures add another layer. Metrics like entropy, Poincaré (SD1/SD2), and DFA α₁ describe how structured or adaptive your heart rhythm patterns are over time. They don’t just measure how much variability exists, but how that variability behaves — whether it’s orderly, chaotic, or somewhere in between.
Together, these metrics build a multidimensional view of your physiology. RMSSD alone gives a snapshot of vagal activity, but the rest of the HRV landscape — time, frequency, and nonlinear — reveals how your nervous system actually coordinates recovery, stress, and adaptability [1],[2].
Analyzing Our Data
Across time-domain, nonlinear, and frequency-domain views, the story is consistent:
- Short-term variability (RMSSD, SD1, HF) remains suppressed
- Long-term variability (SDNN, SD2) increases with observation length
- Regulatory structure remains orderly and predictable
- Flexibility does not increase with time
The autonomic system is stable and well-controlled, but operating in a sustained low-adaptability mode— consistent with alertness, fatigue, or cognitive load rather than deep recovery. However, without age, baseline history, posture, or situational context:
- These metrics describe physiological state, not health or fitness
- Absolute values should not be labeled “good” or “bad”
Final Summary
- RMSSD answers: How relaxed or recovered is the system right now?
- Nonlinear metrics answer: How adaptable and resilient is the system?
- Frequency metrics answer: How autonomic control is being expressed rhythmically
Together, they show that HRV is not just about more variability, but about how variability is structured and controlled.
Companion Notebook can be found on Kaggle: Signal to Meaning (Part II): R-R Intervals to Physiological Insight
Detailed Analysis, Charts and Graphs below...
Time Domain Metrics
Key observations:
- MeanNN remains stable across all windows
- RMSSD stays consistently low
- SDNN increases as window length increases
- pNN50 appears only in the longest window
Interpretation
- Heart rate itself is steady; changes reflect variability, not rate
- Short-term parasympathetic modulation remains limited
- Increasing SDNN reflects longer observation time, not improved recovery
- Beat-to-beat flexibility does not scale with window length
Takeaway
Overall variability increases with time, but short-term adaptability remains constrained, indicating a stable but low-flexibility autonomic state.
Nonlinear Metrics
Key Observations:
- SD1 remains relatively flat across windows
- SD2 increases substantially with longer windows
- SD1/SD2 steadily declines
- DFA α₁ stays low (~0.6)
- DFA α₂ remains near ~1.0
Interpretation
- Long-range variability expands with time
- Short-term adaptability is capped
- Control strategy favors predictability over responsiveness
- Long-range regulation remains orderly and stable
- This pattern reflects constraint, not instability.
Takeaway
- Nonlinear metrics show a well-regulated but low-adaptability system—stable, orderly, and operating with limited short-term flexibility.
Frequency Domain Metrics
Key observations:
- Total Power declines sharply with longer windows
- HF power drops faster than LF power
- LF/HF rises steadily across windows
- No frequency band increases in absolute power
Interpretation
- The autonomic system becomes quieter overall
- Parasympathetic oscillations diminish disproportionately
- Rising LF/HF reflects loss of vagal modulation, not sympathetic escalation
- Control becomes more centralized and constrained
- This is not a stress-escalation pattern.
Takeaway
- Frequency-domain metrics confirm a low-spectral-richness, constrained autonomic state driven by reduced parasympathetic modulation rather than increasing sympathetic activity.
References:
[1] F. Shaffer, R. McCraty, and C. L. Zerr, “An overview of heart rate variability metrics and norms,” Frontiers in Public Health, vol. 5, p. 258, 2017.
[Online]. Available: https://doi.org/10.3389/fpubh.2017.00258
[2] T. Pham, Z. J. Lau, S. H. A. Chen, and D. Makowski, “Heart rate variability in psychology: A review of HRV indices and an analysis tutorial,” Sensors, vol. 21, no. 12, p. 3998, 2021.
[Online]. Available: https://doi.org/10.3390/s21123998
[3] I. Antelmi, R. S. de Paula, A. R. Shinzato, C. A. Peres, A. J. Mansur, and C. J. Grupi, “Influence of age, gender, body mass index, and functional capacity on heart rate variability in healthy subjects,” American Journal of Cardiology, vol. 93, no. 3, pp. 381–385, 2004.
[Online]. Available: https://doi.org/10.1016/j.amjcard.2003.09.065