REF-1259 · Gouw et al., 2022 · International Journal of Environmental Research and Public Health · Paper · PubMed · Free full text

Can my Tymewear find my VT1 and VT2?

Roughly at VT1, low at VT2: in 15 recreationally active adults the Tyme Wear shirt put VT2 0.8-1.3 km/h and 5-8 bpm below gas analysis; limits of agreement reached 35 bpm. Strap, cycling untested.

B-limited, mixed evidenceMeasured on recreational people

Gouw et al. asked whether the Tyme Wear smart shirt, which estimates minute ventilation from chest-wall stretch and uses it to locate the first and second ventilatory thresholds, is reliable and valid against laboratory gas exchange.

Nineteen recreationally active adults did two treadmill ramp tests 7-10 days apart, following the maker's app-guided protocol (walk 3 min at 3.0 mph, then run from 3.3 mph, +0.3 mph every 30 s to exhaustion), wearing the shirt while a Parvo Medics TrueOne 2400 measured gas exchange; two exercise physiologists located VT1 and VT2 by eye from the ventilatory equivalents.

Fifteen were analysed. The shirt's thresholds repeated well from test to test (ICC 0.91-0.98 for time and speed, 0.78-0.84 for heart rate), similar to gas exchange (0.90-0.96).

Against gas exchange the shirt put VT1 on average 0.2-0.3 mph slower and about 4 bpm lower (significant only for heart rate in trial 1), and VT2 0.5-0.8 mph slower and 5-8 bpm lower (significant in both tests), with wide individual limits of agreement (VT2 speed up to −1.5 to +3.0 mph, VT2 heart rate −18.7 to +35.1 bpm). The four fittest participants were under-read most.

The authors conclude the shirt is reliable but less valid than the lab, adequate for setting zones in most recreational users, with caution for fitter people.

Our reading: a small, one-generation test and the only check of a Tymewear threshold in this base; VT1 from the shirt sat near the lab value on average, VT2 was systematically low and individually scattered, and nothing here speaks to the strap, to cycling or to ventilation itself.

Protocol

Design: two-visit test-retest validation, 7-10 days apart; High Altitude Performance Lab, Western Colorado University, Gunnison (2,348 m); IRB HRC2017-02-03-R77, approved 5 February 2018.
Who: n = 19 recruited, 15 analysed (4 excluded for equipment error or data loss); 8 men, 7 women; age 27.7 ± 8.4 y, 67.8 ± 9.5 kg, VO2max 48.4 ± 10.5 mL/kg/min (men 52.8 ± 12.4, women 43.4 ± 5.1), peak heart rate 184.6 ± 8.9 bpm; all met VO2max criteria. Recreationally active, 18-65 y, no cardiopulmonary conditions.
Device: Tyme Wear smart shirt (Tyme Wear, Boston, MA) fitted to the maker's sizing guide, with the Tyme Wear pod attached to the shirt's hub; the shirt measures chest stretch (as a proxy for tidal volume) and rate of stretch (breathing rate) to estimate ventilation; the thresholds (S-PRED) came from the maker's algorithm, which the authors did not know.
Criterion (TRUE): Parvo Medics TrueOne 2400 metabolic cart; heart rate by Polar chest strap; VT1 and VT2 located by visual inspection of 15-s averages by two trained exercise physiologists — VT1 at the rise in VE/VO2 without a rise in VE/VCO2, VT2 at the rise in both.
Test: the maker's app-guided ramp on a motorised treadmill — 1 min standing, 3 min walking at 3.0 mph, running from 3.3 mph with +0.3 mph every 30 s to volitional exhaustion, 1 min standing after. Three full breaths recorded by both systems before the test. Pre-test questionnaire to repeat conditions; diet, sleep, prior training and time of day not controlled.
Analysis: paired t-tests; CV, typical error and ICC for test-retest reliability (Excel); Bland-Altman bias ± 1.96 SD for agreement (GraphPad Prism). The paper labels the Bland-Altman limits "95% CI"; they are bias ± 1.96 SD, i.e. limits of agreement.
Funding and interests: no external funding; the authors declare no conflict of interest. The test protocol was "developed by Tyme Wear".

Key Results

Agreement, gas exchange (TRUE) minus shirt (S-PRED); positive = shirt lower (Bland-Altman, Results 3.4; km/h our conversion at 1 mph = 1.609 km/h):

ThresholdMeasureTrial 1: bias ± SD (limits of agreement)Trial 2: bias ± SD (limits of agreement)
VT1Time point0.3 ± 0.9 min (−1.4 to 2.0)0.4 ± 1.4 min (−2.2 to 3.1)
VT1Treadmill speed0.2 ± 0.5 mph (−0.8 to 1.2); ≈ 0.3 km/h (−1.3 to 1.9)0.3 ± 0.8 mph (−1.3 to 1.8); ≈ 0.5 km/h (−2.1 to 2.9)
VT1Heart rate4.0 ± 7.2 bpm (−10.1 to 18.1)4.3 ± 11.1 bpm (−17.6 to 26.1)
VT2Time point0.82 ± 1.46 min (−2.0 to 3.7)1.2 ± 1.8 min (−2.4 to 4.8)
VT2Treadmill speed0.5 ± 0.9 mph (−1.2 to 2.2); ≈ 0.8 km/h (−1.9 to 3.5)0.8 ± 1.1 mph (−1.5 to 3.0); ≈ 1.3 km/h (−2.4 to 4.8)
VT2Heart rate5.3 ± 8.6 bpm (−11.7 to 22.2)8.2 ± 13.7 bpm (−18.7 to 35.1)

Means (Table 4), TRUE vs shirt: VT1 speed 4.6 vs 4.4 mph in both trials, heart rate 130.8 vs 126.8 bpm (trial 1, significant) and 129.7 vs 125.4 (trial 2); VT2 speed 7.1 vs 6.6 and 7.1 vs 6.4 mph, heart rate 163.5 vs 158.2 and 162.3 vs 154.0 bpm, time 10.5 vs 9.1 and 10.5 vs 9.3 min (all VT2 differences significant).

Reliability (Tables 2-3; CV, ICC): shirt VT1 time 9.1%, 0.91; heart rate 3.9%, 0.78; speed 6.9%, 0.91; VT2 time 4.6%, 0.98; heart rate 2.8%, 0.84; speed 4.3%, 0.97. Gas exchange VT1 8.0%, 0.95; 3.2%, 0.90; 6.8%, 0.95; VT2 5.3%, 0.96; 2.4%, 0.93; 4.9%, 0.96.
As a share of peak: VT1 at 71 vs 69% (trial 1, significant) and 70 vs 68% of peak heart rate; VT2 at 88 vs 86% and 88 vs 83% (both significant); VT2 at 74 vs 69% and 75 vs 66% of maximal speed (both significant); VT1 speed share not different.
The shirt's spread was narrower than the lab's: VT2 heart rate 158.2 ± 8.9 vs 163.5 ± 13.6 bpm and VT2 speed share 69 ± 4 vs 74 ± 7% (trial 1); the highest VT2 the shirt reported was 79-76% of maximal speed against 88-89% by gas exchange (Table 6).
Individuals: in trial 2 the shirt's VT2 speed was 10.3-27.3% below the lab's in eight participants and its VT2 heart rate 7.21-14.7% below in seven; four participants with higher fitness (VO2max above 60 mL/kg/min in men, above 50 in women) had the largest under-reads at both thresholds in both trials. The authors estimate the shirt's thresholds fall in a usable training range for about 83% (VT1) and 57% (VT2) of this cohort by speed.

Mechanisms

Measured: the shirt's thresholds were repeatable but sat lower than gas-exchange thresholds, more at VT2 than at VT1, and most in the fittest participants; its range of thresholds was narrower than the lab's.
Authors' interpretation: error can enter at the sensor (stretch and rate of stretch as proxies for tidal volume and breathing rate; breathing rate should be easy, tidal volume harder because of shirt fit and sensitivity to chest expansion) or in the threshold algorithm, which may not suit fitter users whose thresholds come at higher intensities; without ventilation in litres per minute from the shirt they could not tell which.They add that unit-to-unit variation between shirts is unknown.
Inference: a narrower spread of thresholds than the criterion, with the highest thresholds pulled down most, is the pattern of an algorithm drawing estimates toward the middle of the group; it predicts that an athlete with a high VT2 relative to maximum — typical of trained endurance athletes — will be under-read. The study had only four such people.
Inference: gas-exchange VT2 is itself located by eye and carried a test-retest CV of 4.9% for speed here; the criterion has its own error (REF-169).

Practical Application

Can my Tymewear find my VT1 and VT2? Roughly at VT1, low at VT2: in 15 recreationally active adults the Tyme Wear shirt put VT2 0.8-1.3 km/h and 5-8 bpm below gas analysis; limits of agreement reached 35 bpm. Strap, cycling untested.
Treat a breathing-sensor VT2 as a floor until checked (B-, 15 recreational runners, one device generation): the shirt's VT2 was on average below the lab's and up to 27% slower in individuals; zones built on it would put threshold work too easy, most for fitter athletes (the authors: an elite athlete "may need more precision"). A field or lab test settles it (SYNTH-53).
VT1 from the shirt was close on average but not for every individual: about 0.3-0.5 km/h and 4 bpm below the lab on average, with limits of agreement of −18 to +26 bpm (lab minus shirt). For keeping easy work easy, the base's default — comfortably below the first threshold, not pinned to it (SYNTH-51) — absorbs an error of this size better than a VT1 used as a target.
What this does not cover (be exact): the study tested where the shirt placed the thresholds (time, speed, heart rate) on a treadmill ramp; it did not test the current VitalPro strap, cycling, the ventilation number itself, breathing rate, or a ventilation value at VT1/VT2 used as a live zone boundary during a ride.Using the strap's ventilation reading as a within-session signal rests on other evidence: ventilation at VT1 held after 2 h of cycling while power at VT1 fell (REF-180), and breathing rate, not minute ventilation, tracked perceived effort (REF-1083); not from this paper.
Compare like with like: the shirt's own thresholds repeated about as well as the lab's (ICC 0.91-0.98 for speed and time), so a change in the same device's threshold between two identical tests is more informative than its absolute value against a lab (inference from the reliability data; training changes were not studied).

Limitations

Small, one device generation, one mode. n = 15 analysed, treadmill running only, the smart shirt of its time with an unknown app and firmware version and an algorithm the authors could not see; unit-to-unit variation between shirts unknown (the authors' limitations). No study of the current strap form is in the base.
Attrition. 4 of 19 were excluded for equipment error or data loss; which system failed is not stated.
Altitude and protocol. Tested at 2,348 m with a fast ramp (+0.3 mph every 30 s; VT2 at about 10.5 min); thresholds at altitude and from short ramps may not transfer to sea-level training. Lifestyle before tests and time of day were not controlled.
Criterion by eye. Gas-exchange thresholds were located visually by two physiologists, with test-retest CVs of 2.4-8.0%.
Reporting that does not add up. The typical-error column is labelled in bpm for heart rate, but a typical error of 0.29-0.58 bpm cannot sit with CVs of 2.4-3.9% at heart rates near 130-160 bpm (that would be about 4-5 bpm, our arithmetic), so the unit is probably not bpm; every typical error in Tables 2-3, for time and workload too, is close to CV × mean ÷ the between-subject SD of Table 4 (our arithmetic), i.e. a standardised, unitless typical error, whatever the footnote's units say.The discussion gives heart-rate differences of "3.7 and 3.7–9.2 bpm", which match neither the Bland-Altman biases (4.0-4.3 and 5.3-8.2) nor Table 4. One VT2 speed bias is printed in "bpm". The limits of agreement are labelled "95% CI".
Interests. "This research received no external funding" and "The authors declare no conflict of interest." The test protocol "was developed by Tyme Wear", and the thresholds came from the maker's algorithm, which the authors say they did not know.
How this entry was checked and changed
Why this level

single-lab validation study, n = 15 analysed of 19 tested, read from the full text (open access, CC BY 4.0). The rubric places a validation study at B; held at B- because 4 of 19 participants were dropped for "equipment error or data loss" without saying which device failed (a shirt that fails in some people and is then excluded looks better than it is), and parts of the reliability and discussion numbers do not add up (see Limitations)

Full text (checked 2026-10-27)
19 subjects recruited; "Of the 19 subjects tested, four subjects' data were excluded due to equipment error or data loss during testing" (full text)
High Altitude Performance Lab in Gunnison, Colorado (2348 m); two GXTs separated by 7 to 10 days; protocol approved 5 February 2018 (full text)
"The protocol for the GXT was developed by Tyme Wear"; walk at 3.0 mph for three minutes, running from 3.3 mph, increased 0.3 mph every 30 s until volitional fatigue (full text)
Total (n =15): age 27.7 ± 8.4, body mass 67.8 ± 9.5 kg, VO2max 48.4 ± 10.5 mL⋅kg−1⋅min−1, peak heart rate 184.6 ± 8.9 bpm (full text)
S-PRED reliability: VT1 time ICC 0.91, HR 0.78, workload 0.91; VT2 time 0.98, HR 0.84, workload 0.97 (full text)
VT1 workload mean difference 0.2 ± 0.5 mph (95% CI, −0.8 to 1.2 mph) and 0.3 ± 0.8 mph (95% CI, −1.3 to 1.8 mph) (full text)
VT2 workload mean difference 0.5 ± 0.9 bpm (95% CI, −1.2 to 2.2 mph) and 0.8 ± 1.1 mph (95% CI, −1.5 to 3.0 mph) (full text)
VT1 heart rate 4.0 ± 7.2 bpm (95% CI, −10.1 to 18.1 bpm) and 4.3 ± 11.1 bpm (95% CI, −17.6 to 26.1 bpm); VT2 heart rate 5.3 ± 8.6 bpm (95% CI, −11.7 to 22.2 bpm) and 8.2 ± 13.7 bpm (95% CI, −18.7 to 35.1 bpm) (full text)
Four subjects with the greatest difference "had higher fitness levels (VO2max > 60 mL⋅kg−1⋅min−1 for males and > 50 mL⋅kg−1⋅min−1 for females)" (full text)
"Without the ability to compare S-PRED VE and TRUE VE in liters per minute, it is difficult to conduct an equivalent comparison" (full text)
"For heart rate, this translates to a difference of 3.7 and 3.7–9.2 bpm between TRUE and S-PRED VT1 and VT2" (full text)
"This research received no external funding." and "The authors declare no conflict of interest." (full text)