Train Red Review (FYER, PLUS)

I’ve used the Train Red muscle oxygen sensor on and off throughout the year in preparation for this review. December 2023 marked a nice update to the smartphone app and prompted me to finally put pen to paper.

TL;DR – This Pro Level Muscle Oxygen sensor whose gold-standard sensor seems to offer good data, a nice app and some integration with Garmin devices. It is aimed at Pro and Lab usage plus committed self-training athletes in the gym and on the road.

Train.Red – What this review is and is not trying to achieve

I don’t have access to a lab to validate accuracy. Nonetheless, I’ll compare it to my other muscle oxygen technologies (Moxy, NNOXX, Humon Hex) to highlight some distinctions.

I aim to strike a balance for readers at various levels, ranging from those unfamiliar with SMO2 to individuals considering the purchase of 12 units for their lab, possibly with extensive experience using Moxy over the years.

For this review, I intend to make it accessible to the average athlete, steering clear of overly technical descriptions of the human body.

What is Muscle Oxygen (SmO2) – Tissue Saturation Index?

The hint is in the name. Muscle oxygen, denoted as SmO2, differs significantly from SpO2, which measures blood oxygen. For performance athletes, the primary focus should be on SmO2, while SpO2 serves more as a wellness indicator for aspects such as sleep, altitude, illness, or jet lag.

Our primary concern lies in the percentage of oxygenated haemoglobin in the muscle, both in absolute terms and in terms of its trends—whether it’s rising, stable, or falling. We’ll delve into the maximums, minimums, and ranges of muscle oxygen to provide a comprehensive understanding.

• In English: Muscle oxygen is oxygen in the MUSCLE ie SmO2. TSI is TISSUE Saturation Index ie the oxygen in muscle PLUS other stuff.
• Oxygenated Hemoglobin Concentration = O2Hb
• Deoxygenated Hemoglobin Concentration = HHb
• Calc: Total Hemoglobin Concentration = tHb = O2Hb + HHb
• Calc: SmO2 is O2Hb as a percentage of total [ 100*O2Hb/(O2Hb + HHb) ]
• (De-oxy) Oxymyoglobin & (HMb) O2Mb cannot be distinguished by NIRS.

It’s also worth noting that blood volume and flow will change in the muscles during exercise.

A Brief History: Muscle Oxygen Sensors in Consumer Sports

While PORTAMON (Artinis) stands as the gold standard for laboratory use, several alternatives are vying to be its successor and cater to a broader audience.

• Moxy Monitor (2013-Present)
• BSX Insight (2014-2018)
• Humon Hex (2018-2020)
• Train.Red (2021-present), uses Artinis Sensor
• NNOXX (2023-present)
• Graspor, Repace, Portamon and Q-Lac are other products

Putting technical issues aside, muscle oxygen sensors solely aimed at the consumer space, such as the Humon HEX, seem to struggle to find a viable market.

Moxy, on the other hand, has consistently targeted labs, researchers, and professional or committed athletes. It is a scientifically validated and widely used product, albeit with low-resolution data and limitations in muscle penetration. Its measurements employ a Monte Carlo technique to address technical challenges.

In my view, Humon Hex boasted an excellent app and Garmin-based ecosystem, along with claims of scientific validation. However, these factors were insufficient to maintain a stable commercial position.

Train.Red, in contrast, focuses on the lab and professional market with what appears to be a solidly grounded product.

NNOXX enters the scene as a more affordable newcomer, also measuring Nitric Oxide in the muscle—a crucial regulator.

How does a Muscle Oxygen Sensor work

Train.Red – What’s In The Box, Compatability and Setting Up

The sensor is packaged with three branded velcro straps of varying lengths, neatly arranged in a compact carry case. Additionally, you may receive patches to place over the sensor for light exclusion, although snug lycra shorts can serve the same purpose.

Charging is facilitated with a standard USB-C cable.

When initiating each usage, the smartphone app prompts for Bluetooth pairing with the strap. Conveniently, pairing with Wahoo, Stages, or Garmin is achievable using ANT+, leveraging the built-in muscle oxygen recording and display capabilities of these products.

For Garmin watches and Edge bike computers, Train.Red offers compatibility through CIQ Data fields. It’s advised to disable sensor pairing and instead connect directly to the data field. To accommodate multiple sensors concurrently, several data fields are available. Manual assignment of an ANT+ ID to each field is recommended to prevent unintended pairing with other sensors.

Furthermore, Train.Red can seamlessly integrate with the VO2 Master app.

TRAIN.RED FYER SPECIFICATIONS

• • Size: 1.2 x 4.4x 5.9 (cm) / .5 x 1.7 x 2.3 (in)
    
  • Weight: 20 grams
    
  • Measure-depth: up to 20mm in your muscle
    
  • Light source & detector distance: up to 35mm
    
  • Sweat-proof: Yes
    
  • Battery: up to 6 hours during active use, 2 hours to full charge. When charging the device is reset.
  • Speed: 10Hz
    
  • Connectivity: BLE and ANT+
    
  • Materials: high-quality biocompatible ABS and TPE

TRAIN.RED PLUS SPECIFICATIONS

Measure more frequently and deeper. Also has an accelerometer to assess external work.

• • Size: 1.5 x 4.4 x 6.8 (cm) / 0.6 x 1.7 x 2.7 (in)
    
  • Measure-depth: up to 25mm in your muscle
    
  • Light source & detector distance: up to 40mm
    
  • Sweat-proof: Yes
    
  • Battery: up to 6 hours during active use, 2 hours to full charge. When charging the device is reset.
  • Speed: 100Hz
    
  • Connectivity: BLE and ANT+
    
  • IMU: 9-axis inertial measurement unit

Using Train.Red

The device is a one-size-fits-all model that swiftly and dependably pairs.

With no buttons to manage, the device turns on and off automatically, and a reset is achieved by placing it on charge.

• LED colours:
  • The green LED is solid while charging and off when fully charged
  • When the blue LED is blinking the sensor is on and ready to connect
  • The white pulsing light indicates the device is connected and working
  • No light: Train.Red is off

Train.Red has its own Garmin Data Fields. However all other muscle oxygen sensor also have their data fields and all seem to work interchangeably, presumably because they all use the same ANT+ Profile for SmO2.

Practically, Train.Red is versatile and can be used on any muscle group, with the only limitation being the maximum size of the thigh due to the largest strap length available. Different muscle groups will need the sensor to be fastened on the body with one of the adhesive patches

To ensure optimal usage, consider the following key factors:

1. Keep Train.Red in a fixed position during the workout to prevent subtle positional variations that could yield different results.
2. Consistently place Train.Red in the same position across multiple workouts for more accurate comparisons.
3. Prevent external light interference with the sensor by wearing tight lycra shorts and avoiding close proximity to other sensors.
4. Utilize the app to calibrate before a workout. While this step is required when the app records a workout, it’s important to note that calibration through a Garmin data field may not be possible.

Train.Red App & Ecosystem

In summary, Train.Red’s smartphone app falls into the category of RECORD and REPORT applications. One segment of the app is dedicated to recording workouts, providing a live chart displaying changing muscle states and SmO2 ranges. Users can even capture a photo of their efforts overlaid with SmO2 metrics. The REPORT section of the app offers similar information for all workouts recorded within the app.

Reviewing apps can be challenging due to frequent updates. This review is based on the December 2023 release, which brings substantial, long-term changes. In its current state, the app has a visually appealing interface, and the relatively straightforward graphs gain enhanced significance through the addition of colour-coded sections corresponding to muscle states and profiles.

It’s important to note that only workouts recorded within the app are displayed. Currently, there is no functionality to sync workouts between Garmin, Wahoo, or VO2master.

Using Train.Red with Wahoo & Golden Cheetah

Train.Red broadcasts SmO2% using the standard ANT+ profile for muscle oxygen, providing a basic but crucial functionality. While you can use the inbuilt muscle oxygen metrics to display SmO2% on your Wahoo device, it’s important to note that the data saved in the FIT file and app is limited to just that specific metric.

Using Train.Red with Garmin EDGE/WATCH and Connect

Similar to the previous example, if you pair your Edge or Watch to Train.Red, you’ll only be able to display and record the transmitted data, specifically SmO2%.

Garmin provides an additional option where you can use one or more Train.Red data fields to record data from multiple sensors. These data fields show the sensor placement, the SmO2 measure, and a line chart of recent readings. You can assign the ANT+ ID of each sensor to its corresponding data field. However, as of now, only SmO2 data is available through this method, but it’s anticipated that this might change over time.

I want to now go through a several differetn workouts recorded with different devices, hopefully to give you a kind of story that will progressively introduce some of the aspects of training with muscle oxygen as well as illustrating aspects of the technologies I’ve used.

Train.Red vs Moxy in an easy aerobic Ramp

Observing the graph, it’s apparent that as the ramp progresses, my SmO2/TSI levels start to plateau after around 18 minutes. This stabilization point may signify a sustainable level where oxygen utilization and demand reach a balanced state

You might now be wondering why the Moxy and Train.Red values are completely different despite seeming to follow a similar pattern.

Moxy vs. Train.Red Technical differences

Moxy asserts that its product provides readings of absolute SmO2, while Train.Red measures something similar termed TSI (relative concentration change of Tissue Oxygenation). However, Train.Red reads values from different tissues. It emits light at a wavelength of 850nm, which penetrates more deeply than shorter wavelengths. In comparison, Moxy employs wavelengths of 680nm, 720nm, 760nm, and 800nm, where the 760nm wavelength penetrates only a few millimeters. Train.Red claims to incorporate technical design factors related to transmitter/receiver spacing, enabling penetration of up to 30mm.

Both sensors utilize complex mathematical models to derive readings. Moxy incorporates an adipose tissue layer correction based on a Monte Carlo model/look-up table, resulting in lower values.

The data presented has been averaged over per-second intervals. Other charts may display fuzzier data when plotting more frequent raw data. Moxy reports sample frequencies of 0.5, 1, or 2 Hz but samples at over 1000Hz. In comparison, Train.Red has a reporting frequency of 10-25Hz (100Hz in the PLUS model).

Train.Red Used In a Warmup

One use for Train.Red and muscle oxygen data is to ensure you have warmed up properly. The second of these two charts shows you the power/HR profile for reference of the British cycling warmup I typically use. . The question to consider is whether this warm-up routine is effective.

In the first image, it’s noticeable that the warm-up initially reduces my Muscle Oxygen (Moxy left VL, Train.Red right VL), and at the end, it should ideally be higher than when I started. However, this expected increase didn’t occur in this instance, contrary to the usual outcome with this warm-up routine. Normally, I would anticipate Moxy to be around 70% and Train.Red in the mid-60s. In the subsequent workout, my SmO2/TSI did eventually reach the expected levels, suggesting that I might not have been fully warmed up during by this protocol.

On a different day, following the same warm-up protocol, the next chart displays only Train.Red SmO2/TSI data using its app. I’ve overlaid a green line representing total Hemoglobin (tHb = O2Hb + HHb, oxy plus deoxy). The warm-up appears to have the desired effect, increasing haemoglobin levels to an initial plateau where short sprints were performed. Following the sprints and additional rest, the plateau reached an even higher level, indicating a positive outcome.

Hopefully, that makes sense so far. Looking at the overlaid Muscle State information on the same data (below), it’s important to realise that there aren’t defined SmO2 zones akin to HR zones. In simplistic terms, 55% SmO2 doesn’t have a specific meaning, as the oxygenation could be either increasing or decreasing at a variety of rates. Instead, the concept of Muscle States is employed, as depicted in the chart with its visually informative shading.

Explanation: Train.Red Muscle States & What They Mean

• BLUE: Recovery – the muscle is recovering. Not as simple as a SmO2 increase, but a combination of all the parameters.
• PURPLE: Load – similar to an occlusion where blood flow is restricted by your muscles growing and compressing small capillaries
• GREEN/YELLOW/RED: EASY – MEDIUM – HARD effort – Increasing rates of SmO2 consumption

TRAIN.RED Used In A Ramp Test vs. Moxy

This is a straightforward ramp. As you might expect heart rate rises with power and SmO2 falls with power. Both recovered quickly after the effort stopped, interestingly, two and a half minutes after I finished, SmO2 rose to my normal level of around 68% (77% with Moxy).

You can perhaps partly see that at about 10:00, the rate of decline of  SmO2 (Moxy) lessened, whereas the rate of decline of TSI/SmO2 (Train.Red) appears to fall consistently in line with the increased load.

To demonstrate improvement you would want to ramp to higher power levels before failure. However, another thing that marks improvement is to watch when the flattening of the SmO2/TSI curve occurs, this is called the 2nd SmO2 break point. The longer you can sustain an effort after this point reflects how well your aerobic system contributes to your performance. This point would also broadly correspond to a VT or Lactate Threshold.

The following chart better shows this for someone fitter than me!

TRAIN.RED Desaturation Intervals vs. HR

Here is another warmup (as above) followed by 6 intervals of a minute or so. I did a complete recovery in between. From memory, the power level was quite a bit over FTP; hard, but nothing crazy. You can see that my SmO2/TSI declined to about 58% each time and that my HR was similar each time. So using those conventional metrics you might assume each effort was the same. Clearly not! as the recovery in SmO2 is different each time. The first recovery is to 67.5% and the fifth to 65.5%…funnily enough it shows that the more intervals you perform the harder they get!

That’s perhaps not the most enlightening thing I’ve ever shared, however, consider if rather than using time for recovery you used SmO2 or Muscle state.

ie I could have performed the effort based on the same power/duration but waited until I recovered to 66.5% or some other desired level. This would potentially make each rep equally as difficult.

If you’re familiar with power-based training using Xert, you may have engaged in similar types of intervals with recoveries of varying durations based on the modelled recovery from the preceding intensity and duration. The principle I’m discussing here is akin to that, but with Train.Red/Muscle Oxygen, the key distinction is that you can MEASURE the recovery rather than relying on a model.

TRAIN.RED to build muscle more efficiently

A broad principle for your work lifting in the gym might be that endurance athletes should do higher numbers of reps and lift lighter weights. Whereas those looking to build muscle mass would do fewer amounts of heavier weights. If you were doing weights to build more powerful muscles you might also want to wait until you have fully recovered from one set before embarking on the next.

Using the same principles outlined in the previous section we can see from this chart how monitoring Muscle Re-Oxygenation can be used to tell us when to start the next rep depending on what we want to achieve.

Check out my calf workout in the following images. The second image is zoomed in, showing two of the troughs. You might think the line is a little bit jagged. Those are individual muscle contractions…you can see each one!

Train.Red Smart Recovery and De-Saturation Intervals

I restarted this workout after a good warmup as my SmO2 levels jumped to 80% which is unusually high for me at the start of a workout, I thought maybe the calibration was off and that a restart would fix that with recalibration. It didn’t. I’m assuming the sensor placement was right over a chunky blood vessel? 🙂

This test was designed to let me do a lot of 30-second intervals at 120% FTP (thirty reps, I think) but I wouldn’t start the next interval until my reoxygenation rate had just started to flatten off. What I hoped would happen would be that the recovery time would lengthen between intervals…but it didn’t, at least not noticeably. The only clear change is about three-quarters of the way through where my muscle oxygen range drops.

I basically ran out of pre-programmed interval steps, so stopped! (I got a bit bored as well) The session was perhaps ‘comfortably’ hard with my HR only just creeping into Z3 despite Z6ish power efforts – I don’t think I’d made the workout hard enough! If only Train.Red could help with that. Oh…wait a minute. What’s this?

You can see the SMART REST analysis shows the REST ZONES. What I was aiming for was REST ZONE 1 (“Stamina” for endurance) but clearly spent way too much time recovering beyond that in the R2 and R3 zones. Grrr. If only Train.Red could have helped with that whilst I was exercising rather than after the fact. Oh…wait a minute…you guessed it

The REST and GO message on the live workout dashboard gives you the instruction you need to end the rest period at just the right time. I think the GO message is determined by the classifications you provide at the start of the workout.

The rest zones and smart rest feature are pretty cool…I’d like to have them on my Garmin.

Identify Your ‘Threshold’

One of the neat features of the old Humon Hex product was that it included a feature that purported to determine your LTHR (LT2). It always seemed to me to be “broadly” right. But what if you want a more rigorously tested approach? To that end, I spoke with Bruce Rogers (NFRMC, Florida) who co-authored this recent study that found how to better detect the RCP/MMSS (maximal metabolic steady state) from HHb break points and dfa a1.

Pacing Endurance Efforts

A couple of years ago in the Pyrenees, I used a different muscle oxygen sensor to manage my efforts during several days of relatively challenging riding. The consistently hot weather, combined with the effects of altitude, presented additional challenges. Caffeine affected the reliability of my heart rate data, and the interpretation of power data was influenced by altitude. In this context, SmO2 was useful for gaining a better understanding of when I was approaching my SmO2 2nd breakpoint. My approach was more practical than scientific at that time but it seemed to work.

Here’s an example from one of Train.Red’s employees in the Rotterdam marathon, where they utilized various gadgets. The SmO2 data became notably useful around 60-90 minutes into the race, indicating a settled and hopefully sustainable level that was maintained through to the end. Post-run data revealed that the athlete spent most of the time in roughly equal proportions in the medium and hard states, a pattern potentially suitable for the marathon distance. If the athlete was pacing based on heart rate, you can appreciate the dilemma they would have faced given that HR rose progressively through the race…what instead was a target HR?; SmO2 adds another potentially valuable data point.

I wouldn’t entirely entrust my race pacing solely to Muscle Oxygen level or threshold in the same manner as I might rely on power, speed, or heart rate. Instead, I view it as an additional data point that adds value. It provides supplementary insights and may maintain its usefulness and reliability for an extended duration compared to some other data points. A variety of data sources can contribute to a more comprehensive understanding of your performance and aid in making informed decisions during a race.

Placement & Wearing

Ensuring consistent placement and excluding external light interference is crucial for accurate readings with Train.Red. Wearing dark lycra shorts is highly recommended for stability during running or cycling. While Train.Red provides a robust Velcro strap, and the addition of lycra prevents undesired movement. Disposable patches are also available for this purpose, suitable for lab experiments yet potentially costly for everyday athletic use. It’s worth noting that if you intend to place the sensor on the torso, the largest strap will not be sufficient.

In discussions with Jem Arnold @UBC during the preparation of this review, he highlighted that day-to-day variability for one athlete could be 5-10% SmO2 with another technology. This implies a notable degree of uncertainty around repeatability, making it challenging to rely on a given SmO2 level as a pacing target. It’s unclear if the same variability applies to Train.Red – thank you, Jem!, he authored a paper on Muscle Oxygen, albeit with competing technology.

Other Things you can try or should know

• Breathing – this might sound stupid! If you focus on breathing then oxygenation can increase. It does for me. So long as you are not in an athletically stressed state, breathing in with two breaths and out with one long breath impacts stroke volume, NO, HRV and SmO2. Who would have thought breathing was good for you? 😉 The point here is that you might benefit from actively moving away from your body’s default learnt breathing behaviour, or at least consider some alternatives. With Train.Red you can see within a few 10s of seconds if it makes any difference.
• Getting off the saddle – Other things equal, you should find that short periods of riding out of the saddle boosts SmO2. Perhaps blood flow is improved or perhaps the muscle you are measuring is being partially relieved by other muscle groups taking the strain in your new position.
• Calibration with Train.Red involves an app prompt, but there’s an alternative calibration method involving an occlusion. By using a band to temporarily restrict blood flow to the muscle where the sensor is placed, you aim for SmO2 to decrease to a minimum during the occlusion. Comparing these minimums with readings from other sensors or workouts on different days allows for better comparisons.
• Physiological occlusions – some cyclists’ physiology adversely affects blood flow in certain positions as, say, the hip bones could restrict flow in certain riding positions eg TT. You can perform tests to check for this.
• Demographics – Moxy is reliable on young, fit men (Source: Jem Arnold). Less so with other demographics. I don’t know if the same holds for Train.Red.

What’s Missing

As an endurance athlete, I want better guidance from Train.Red in the Garmin and Wahoo environments. As a minimum, I want coloured Muscle State information in real time.

I have some of Moxy’s apparel which has a pocket and a shield to ensure the product is placed consistently from one day to the next AND that light is excluded. Similar apparel from Train.Red would be great and would add the opportunity for the brand to boost revenues…WHOOP apparel is similar.

Take Out

I felt a bit disheartened a couple of years ago when Humon Hex shut down. While I wasn’t a daily user of SmO2, I did use it occasionally, and it still sits on my desk as I write this. Hex had a solid ecosystem for endurance athletes, covering both the smartphone app and Garmin data fields. Train.Red has reignited my interest in SmO2 sensors, particularly for warm-ups and optimizing rest periods between intervals. I’ll continue using it periodically for endurance runs and rides. However…

However, Train.Red needs to enhance its rather basic Garmin data field. The smartphone app, while generally good, could also see improvements for in-workout usage. Considering it’s a relatively new company, they deserve some leeway for another year or so. The product is certainly usable as of today, and I have faith in the sensor’s quality. Although it might involve a bit of “blind faith” as I haven’t found a way to conclusively prove the data’s accuracy. I rely on trust in others and the fact that the sensor comes from the same people behind the PORTAMON Gold Standard Lab product (Artinis).

Now, let’s talk about the repeatability of the data. Wearing it in the same place from one workout to the next isn’t as straightforward as you might think. Even if you manage to do that, there’s likely to be variability from one day to the next. While there’s a calibration mechanism on the app, the challenge arises on race day when I’m not carrying a phone. Additionally, the gym crowd faces the choice of repositioning the sensor multiple times in one workout on various muscle groups or investing in multiple sensors.

This review was published in the lead-up to Christmas 2023. Despite Train.Red’s relatively high price, the company sold out of both tiers of sensors. I suspect a few sports labs may have acquired some, and it seems the CrossFit/HYROX community has also enthusiastically embraced this technology, perhaps much more so than endurance athletes like me.

As a sports data geek, I’m naturally inclined to want to use this kind of tech. However, I’m not entirely convinced that there are many other endurance athletes with the same trait, especially given the hefty price tag of €674 each.

Train.Red Products

• Train.Red FYER: $749 $674 (x4 @ $2846) 10Hz, (Moxy I believe 2 or 4 Hz
• Train. Red PLUS: $1829 $1646 (x4 @$6950) 100 Hz, with 9 axes IMU, 6 channels.

10% Train.Red discount automatically applied at checkout with code THE5KRUNNER at this link https://train.red/discount/THE5KRUNNER