Single-Sensor SmO2 and Wearable NIRS for Coaches
A practical field guide for threshold detection, trace interpretation, and hybrid-racing application, updated with 2025-2026 findings.
Executive summary
• Wearable NIRS is most useful as a local-muscle dashboard, not as a universal threshold oracle.
• The strongest threshold evidence still favors the later breakpoint (BP2 / second-threshold region), not the first threshold.
• Recent rower and single-leg cycling studies reinforce the same coaching rule: strong relationships do not make local and systemic thresholds interchangeable.
• Recovery kinetics, repeatability, and within-athlete trend tracking may be the most practical wins for coaches.
Wearable near-infrared spectroscopy has matured enough to deserve a place in coaching, but only when the coach respects what the sensor actually measures. The strongest current evidence supports SmO2 as a local-muscle dashboard for within-athlete monitoring, especially around second-breakpoint behavior, work-to-recovery kinetics, and repeated standardized testing. The weakest evidence remains first-threshold detection and any claim that local NIRS values are interchangeable with blood lactate or ventilatory thresholds. [1,3,5,6]
Use the quick sheet on the floor. The deeper evidence review starts below.
| Question | Best current answer | Coach move |
| What does SmO2 measure? | A local balance between oxygen delivery and oxygen use in the tissue under the sensor, not a whole-body metabolic state. | Treat it as a local dashboard, not a blood lactate surrogate. |
| Which threshold is most defensible? | The later breakpoint, often called BP2 or the second muscle oxygenation threshold region. First-threshold detection is still less consistent. | Anchor coaching first around the higher-intensity breakpoint. |
| Can it replace lactate or gas-exchange testing? | No. Strong correlations can coexist with wide limits of agreement and failed equivalence tests. | Use NIRS alongside pace, power, heart rate, RPE, and session context. |
| What is the best coaching use case? | Repeatable within-athlete monitoring on the same muscle, same side, same setup, and same protocol. | Standardize hard and compare athletes mostly to themselves. |
| What matters outside thresholding? | The rate of decline, the floor reached, and the speed of reoxygenation between efforts. | Save kinetics, not just one percentage. |
| Biggest mistakes | Too-short stages, poor contact, light leakage, motion artifact, adipose-related attenuation, and cross-athlete absolute comparisons. | Fix setup before interpreting physiology. |
What wearable NIRS and SmO2 actually measure
SmO2 is the percentage of oxygenated heme signal relative to total heme signal within the sampled tissue, typically reflecting oxygenated hemoglobin and myoglobin inside the muscle under the sensor. That makes it a local muscular response, not a direct measure of systemic lactate accumulation or whole-body ventilatory state. The practical consequence is simple: the coach should care more about trace shape, slope changes, and repeatability than about chasing one universal SmO2 number. [3,4,5]
The more honest coaching language is this: SmO2 tells you whether the monitored muscle is keeping up with local oxygen demand or falling behind. When work rises and oxygen use increasingly outpaces local delivery, the trace trends downward. When recovery is adequate, the trace re-oxygenates upward between efforts. That is why NIRS can add value even when it does not perfectly match a laboratory threshold. [3,4,5]
Where the evidence is strongest in 2026
The 2023 systematic review and meta-analysis remains the clearest starting point for threshold claims. It found that agreement between first-threshold measures was only moderate, with a pooled ICC of 0.53, whereas second-threshold comparisons reached a pooled ICC of 0.80. In other words, the evidence hierarchy is still clear: BP2 is more defensible than BP1. [1]
The 2024 updated systematic review reinforced the broader practical value of muscle oximetry. Across 191 studies and 3435 participants, the review concluded that NIRS can be a meaningful marker of skeletal muscle oxidative capacity and may become one of the primary monitoring tools used alongside heart rate or mechanical power. The same paper also emphasized that instrumental limitations remain and that more longitudinal studies are needed. [3]
Recent studies mostly strengthen that hierarchy rather than overturning it. Coaches should therefore stop asking whether wearable NIRS has solved threshold testing and instead ask a better question: where does this tool improve day-to-day decision-making when used with discipline? The best current answer is around repeatable second-breakpoint behavior, recovery kinetics, and local strain during standardized sessions. [1,2,3,5,6]
What is new from 2025 to 2026?
NCAA Division I female rowers: useful, but not interchangeable
A 2025 study in NCAA Division I female rowers sharpened the threshold discussion in a very coach-relevant way. SmO2BP1 significantly underestimated LT1 and showed poor-to-moderate agreement. SmO2BP2 did not differ significantly from LT2 and showed moderate-to-good agreement, but equivalence testing still failed. That matters because it demonstrates the real coaching truth: a local muscular breakpoint can sit near LT2 without being statistically interchangeable with LT2 for every athlete. [5]
The same rower study also adds something practical beyond breakpoint chasing. Mean SmO2 re-saturation changed longitudinally in the subgroup tested repeatedly, including a significant reduction from November to February. That suggests that recovery kinetics may become one of the most useful ways to follow how an athlete is adapting over time. [5]
Single-leg cycling: strong relationships, loose agreement
A 2025 single-leg cycling study reached a similar conclusion from a different angle. In the locomotor vastus lateralis, local NIRS and EMG breakpoints correlated strongly with the respiratory compensation point, with correlations around r = 0.90 and ICC values of 0.92 to 0.94. However, the mean absolute differences were still about 8.8 to 10.2 W, and the limits of agreement were large enough that the authors cautioned against using local and systemic breakpoints interchangeably. [6]
This is the sentence coaches should remember: correlation is not interchangeability. A convincing relationship on paper does not mean the athlete's local breakpoint is exactly LT2 in a way precise enough for universal zone prescription. [5,6]
Wearable validation is improving
A 2025 validation paper compared the NNOXX One continuous-wave wearable with a bench top frequency-domain NIRS reference during incremental cycling. Individual correlations ranged from 0.69 to 0.88, with a mean of 0.79. Most subjects showed RMSD under 5%, and Bland-Altman analysis placed 95% of paired measurements between about -8.1% and 7.6% SmO2. That is promising for real-time field tracking. [7]
It is also worth reading the paper honestly. The study included NNOXX employees, was funded by NNOXX, and states that company personnel contributed to data analysis. The results are useful and encouraging, but independent replication would make the case much stronger. [7]
Site selection is getting smarter
A 2026 study in 52 world-class Nordic skiers showed that BP1 was too often undetectable or inaccurately detected for practical use. BP2, however, was detected in the vastus lateralis, biceps femoris, and biceps brachii far more successfully than in the triceps brachii, with no meaningful male-female difference. That is an important upgrade to older one-site thinking: the best sensor site depends on the sport task and the limiting muscles, not on a universal placement rule. [9]
Upper-body strength work remains a caution zone
Not every exercise setting is equally friendly to wearable SmO2. In a 2026 study of recreationally trained men during upper-body strength training, MOXY-derived SmO2 showed low inter-session reliability, particularly during and immediately after sets. That does not make NIRS useless in the gym, but it does mean coaches should be much more cautious when trying to turn upper-body set-to-set SmO2 into precise programming decisions. [10]
Local versus systemic thresholds: the interpretation rule that saves coaches
A recurring coaching mistake is to treat every SmO2 breakpoint as a substitute for blood lactate threshold, ventilatory threshold, or respiratory compensation point. The recent literature does not support that habit. What it supports is a more careful statement: wearable NIRS often identifies a repeatable local muscular breakpoint region that can complement systemic threshold testing, especially around the higher-intensity boundary between heavy and severe exercise. [1,5,6]
That distinction is not just academic. It changes how you talk to athletes. 'Your right vastus lateralis shows a repeatable later breakpoint near this pace' is an honest interpretation. 'This is exactly your LT2' usually is not. Used with that level of honesty, NIRS becomes more useful, not less useful, because it stops pretending to be something it is not. [5,6]
How to run a usable single-sensor field test
The first decision is not the protocol. It is the muscle. For most endurance running, cycling, rowing, and many hybrid-racing applications, the vastus lateralis remains the most practical single-site choice because it is superficial, commonly studied, and highly relevant to locomotion. At the same time, recent evidence from Nordic skiers shows clearly that the best site should follow the demands of the sport rather than a fixed dogma. [4,9,11]
The second decision is standardization. Use the same muscle, the same side, the same sensor orientation, the same attachment method, the same warm-up, and the same stage design whenever you want to compare tests over time. Exact landmark conventions vary across studies, so consistency is usually more important than pretending there is one sacred coordinate. Prepare the skin, secure the sensor firmly, and shield it from ambient light. [2,4]
The third decision is stage structure. The most usable recent protocols cluster around 3 to 5 minute steps with short fixed recoveries. Elite running work has used 3-minute stages. A cycling reliability study used 5-minute stages with 1-minute passive recovery and averaged the final 60 seconds of each stage. The rowing paper used a discontinuous 7 x 4-minute step test with 1-minute rests, calculated recovery slopes from the first 30 seconds of recovery, and used the final 30 seconds of each work step for end-stage SmO2. That family of protocols offers a practical balance between signal stability and coaching usability. [2,5,12]
How to read the trace in practice
A usable trace usually shows a stable baseline, then a relatively shallow decline through easier stages, followed by a clearer change in slope as intensity rises. Later in the test, the line often falls more steeply, compresses toward a lower floor, or both. Between stages or repetitions, the rate of reoxygenation can be just as informative as the desaturation itself. [5,8]
This is one place where older coaching habits need updating. The signal's rate of change may be more informative than the raw percentage. In elite male soccer players, SmO2 rate correlated very strongly with blood lactate across multiple sites, and the rate-derived breakpoints were not significantly different from LT1 or LT2. A separate 2025 cycling paper argued that coaches should analyze the whole signal rather than simplifying it to one breakpoint. Together, those studies support a practical rule: read the pattern, not just the number. [8,12]
When the trace is flat, wildly jagged, or oddly random, think setup problem before physiology. Loose contact, light leakage, sweat-related slip, motion artifact, poor site selection, and stages that are too short to settle are still the most common reasons coaches get fooled by NIRS. [2,4]
HYROX and hybrid-racing application
The first peer-reviewed HYROX physiology paper makes one thing clear: HYROX is endurance-dominant. Faster completion times were associated with higher VO2max, greater endurance training volume, and lower body fat percentage. The authors concluded that HYROX places greater emphasis on endurance capacity than on maximal strength when compared with other forms of high-intensity functional fitness. [11]
That matters for sensor selection. A single quadriceps sensor can be genuinely useful in HYROX because so much of the race depends on compromised running, sled push and pull drive, sandbag lunges, and wall-ball fatigue. The sensor can show whether the monitored locomotor muscle is recovering between stations, whether local strain is accumulating faster than pace suggests, and whether the athlete is leaving the sustainable domain earlier than expected. [5,11]
The limitation is representation, not usefulness. A quadriceps sensor will tell you less about SkiErg-specific upper-body demand, grip-limited carries, or trunk fatigue. In some sessions the athlete may feel wrecked while the monitored quadriceps trace still looks relatively calm. That does not mean the device failed. It means the measured muscle is not the main limiter for that task. The 2026 upper-body strength reliability paper is a good reminder not to force false precision in upper-body work that is not yet measured reliably enough. [9,10,11]
Failure modes and data-quality red flags
The 2025 critical review is especially useful here because it states the problems coaches most often learn the hard way. Subcutaneous adipose tissue thickness attenuates the optical signal. Source-detector separation influences effective depth. Skin preparation, exact site selection, and standardization are still major methodological gaps in the field. Those issues do not make NIRS unusable, but they do mean that setup quality controls the value of the data more than most coaches think. [4]
Bottom line for coaches
Wearable NIRS is best understood as a local-muscle performance dashboard. Its strongest current coaching value is in repeatable within-athlete monitoring, especially around the second-breakpoint region and around recovery kinetics after demanding work. The best recent studies do not support treating SmO2 as a universal threshold oracle, and they do not support using local breakpoints interchangeably with lactate or ventilatory thresholds. They do support careful use for pacing, load management, session design, and athlete education when the setup is standardized and the interpretation stays honest. [1,2,3,5,6]
The frontier is still moving. Concepts like the SmO2 plateau or critical oxygenation are interesting and may eventually sharpen coaching interpretation, but they are still emerging research frameworks rather than ready-made field metrics. For now, the winning approach is straightforward: use the same site, the same protocol, and the same decision rules repeatedly. Save stage averages, pace or power, heart rate, RPE, breakpoint region, floor behavior, and re-oxygenation speed. Let NIRS inform your judgment. Do not let it tempt you into false precision. [13]
Practical sample protocols
| Use case | Practical format | What to save |
| Education/demo | 3 to 4 progressive stages of about 2 minutes with 1 minute easy between | Show the concept, but do not set zones from this alone. |
| Threshold-style field test | 4 to 6 stages of 3 to 5 minutes with about 1 minute easy or passive recovery | Final 30 to 60 second stage average, slope change, pace or power, heart rate, RPE. |
| Recovery kinetics check | Repeated bouts with fixed work and fixed recovery | Resaturation slope, next-bout floor, and whether recovery is slowing. |
| Constant-load durability check | 1 to 2 steady bouts near goal pace or power | Whether SmO2 stabilizes, keeps drifting down, or recovers appropriately after the bout. |
Practical hybrid-racing use cases
| Session type | How a single SmO2 sensor can help | Best metric to watch |
| Progressive run or bike | Estimate the local later-breakpoint region | Slope change and stage average |
| Compromised run after a station | See whether the quads actually recover while pace stays controlled | Reoxygenation speed and next-bout floor |
| Repeat sled or lunge intervals | Detect accumulating local strain earlier than pace alone | Progressive drop in floor and slower recovery |
| Race rehearsal | Teach pacing and identify where the athlete leaves the sustainable domain | Breakpoint location relative to goal pace |
Red flags to respect
| Failure mode | What it often looks like | Better response |
| High adipose thickness at the site | Blunted or less responsive trace | Choose the best available site and manage expectations. |
| Loose contact or sweat slip | Sudden jumps or unstable signal | Re-secure the sensor and improve fixation. |
| Ambient light leakage | Noisy line, often worse outdoors or under bright lights | Add a darker cover and check the edges. |
| Too-short stages | No clear inflection or unstable stage averages | Lengthen stages to 3 to 5 minutes for real testing. |
| Cross-athlete comparison obsession | Claims like 'your 32% beats their 36%' | Compare each athlete against their own prior files. |
| Overgeneralizing from one muscle | False conclusions about whole-body state | Interpret within the function of the muscle being measured. |
Resources:
| Source | Why it matters |
| Sendra-Pérez et al., 2023, Scientific Reports [1] | Best overview of threshold reliability; shows second threshold is stronger than first. |
| Perrey et al., 2024, Sports Medicine [3] | Broadest updated review of practical applications in sports science. |
| Yogev et al., 2023, Frontiers in Sports and Active Living [2] | Useful reliability paper for repeated cycling testing. |
| Eserhaut et al., 2025, Frontiers in Physiology [5] | Excellent applied paper on rowers, threshold agreement, and longitudinal recovery kinetics. |
| Tilp et al., 2025, Frontiers in Physiology [6] | Strong reminder that correlation does not equal interchangeability. |
| Peikon et al., 2025, Bioengineering [7] | Promising wearable-validation paper; useful, but industry-linked. |
| Sendra-Pérez et al., 2025, European Journal of Applied Physiology [8] | Helpful for learning to interpret the whole signal, not just one breakpoint. |
| Forot et al., 2026, Experimental Physiology [9] | Sharp update on sport-specific site selection and the weakness of BP1. |
| Marcos-Frutos et al., 2026, Applied Sciences [10] | Important caution for upper-body strength monitoring. |
| Brandt et al., 2025, Frontiers in Physiology [11] | Best current peer-reviewed paper on HYROX demands and performance determinants. |
References
1. Sendra-Pérez C, Sánchez-Jiménez JL, Marzano-Felisatti JM, Encarnación-Martínez A, Salvador-Palmer R, Priego-Quesada JI. Reliability of threshold determination using portable muscle oxygenation monitors during exercise testing: a systematic review and meta-analysis. Scientific Reports. 2023;13:12649. doi:10.1038/s41598-023-39651-z.
2. Yogev A, Arnold J, Nelson H, Clarke DC, Guenette JA, Sporer BC, Koehle MS. Comparing the reliability of muscle oxygen saturation with common performance and physiological markers across cycling exercise intensity. Frontiers in Sports and Active Living. 2023;5:1143393. doi:10.3389/fspor.2023.1143393.
3. Perrey S, Quaresima V, Ferrari M. Muscle oximetry in sports science: an updated systematic review. Sports Medicine. 2024;54(4):975-996. doi:10.1007/s40279-023-01987-x.
4. Sendra-Pérez C, Encarnación-Martínez A, Priego-Quesada JI. Critical narrative review of the applications of near-infrared spectroscopy technology in sports science. Sensors. 2025;25(21):6798. doi:10.3390/s25216798.
5. Eserhaut DA, De Leo JM, Provost JA, Ackerman KE, Fry AC. Monitoring skeletal muscle oxygen saturation kinetics during graded exercise testing in NCAA division I female rowers. Frontiers in Physiology. 2025;16:1538465. doi:10.3389/fphys.2025.1538465.
6. Tilp M, Mosser N, Schappacher-Tilp G, Kruse A, Birnbaumer P, Tschakert G. The relationship and agreement between systemic and local breakpoints in locomotor and non-locomotor muscles during single-leg cycling. Frontiers in Physiology. 2025;16:1465344. doi:10.3389/fphys.2025.1465344.
7. Peikon E, Corso JL, Otic N, Kierul O, Franceschini MA, Robinson M. Continuous monitoring of muscle oxygenation in endurance athletes during incremental cycling: experimental validation of a wearable continuous-wave NIRS sensor using frequency-domain near-infrared spectroscopy. Bioengineering. 2025;12(11):1153. doi:10.3390/bioengineering12111153.
8. Sendra-Pérez C, Encarnación-Martínez A, Salvador-Palmer R, Murias JM, Priego-Quesada JI. Profiles of muscle-specific oxygenation responses and thresholds during graded cycling incremental test. European Journal of Applied Physiology. 2025;125:237-245. doi:10.1007/s00421-024-05593-1.
9. Forot J, Schorderet F, Burdet C, Brocherie F, Millet GP. Agreement between muscle saturation breakpoints and lactate thresholds: muscles comparison and sex difference in world-class Nordic skiers. Experimental Physiology. 2026. doi:10.1113/EP092216.
10. Marcos-Frutos D, Rojas-Cepero I, Martos-Arregui A, Rivero-Rodríguez J, García-Ramos A. Limited applicability of MOXY-derived muscle oxygenation for monitoring upper-body strength training. Applied Sciences. 2026;16(4):2033. doi:10.3390/app16042033.
11. Brandt T, Ebel C, Lebahn C, Schmidt A. Acute physiological responses and performance determinants in Hyrox© - a new running-focused high intensity functional fitness trend. Frontiers in Physiology. 2025;16:1519240. doi:10.3389/fphys.2025.1519240.
12. Batterson PM, Kirby BS, Hasselmann G, Feldmann A. Muscle oxygen saturation rates coincide with lactate-based exercise thresholds. European Journal of Applied Physiology. 2023;123:2249-2258. doi:10.1007/s00421-023-05238-9.
13. Kowalski T, Rębiś K, Malczewska-Lenczowska J, et al. Muscle oxygen saturation plateau: definition and verification in different oxygen availability conditions in locomotor and non-locomotor muscles. Experimental Physiology. 2026;111:464-473. doi:10.1113/EP092862.