After 20 min of near-infrared light on the head (810 nm headset with a nasal probe) just before riding, 25-min time-trial power did not differ significantly from an inactive sham (283 vs 287 W, lower with the light, p = 0.105), and neither did heart rate, blood lactate or RPE in a preceding 15-min ride at LT1 + 5%, in a double-blind crossover of 18 trained male amateur road cyclists
Main caveat: The authors' own: no brain imaging (fNIRS or EEG), so whether light reached the prefrontal cortex is unknown; one application only.
Arrighi and colleagues tested whether transcranial photobiomodulation, proposed to raise prefrontal cortex oxygenation and alter central fatigue, improves cycling performance. Twenty male amateur road cyclists enrolled; two were excluded for not keeping the required days between sessions, leaving 18.
After a first visit for a lactate-threshold test and familiarisation, each rider came twice, in counterbalanced double-blind order, for 20 min of active PBM or sham with a Vielight Neuro Gamma headset (four 810 nm LEDs on the head and one in the right nostril, pulsed at 40 Hz).
Three minutes later they warmed up for 10 min, rode 15 min at 5% above their first lactate threshold, rested 3 min, then rode a 25-min self-paced time trial on their own bikes on an Elite Direto X-RT trainer, seeing only elapsed time. Mean time-trial power was 283 W after PBM and 287 W after sham (p = 0.105).
Heart rate, lactate, RPE and the power/HR and power/RPE ratios did not differ by condition and showed no condition × time interaction in either ride. The authors say the device's irradiance probably did not reach the cortex in a useful amount, and that higher doses, other wavelengths or repeated use would be needed to test the idea.
Participants
18 male trained amateur road cyclists analysed (20 enrolled), 29.5 ± 4.1 years, 71.8 ± 7.0 kg, 6 ± 2 years of practice.
LT1 234.2 ± 26.6 W (3.28 W/kg) and LT2 290.9 ± 39.8 W (4.06 W/kg), set at fixed lactate values of 2.0 and 4.0 mmol/L.
Protocol
Visits: 1, incremental lactate test (4-min stages, +30 W) and familiarisation with both rides; 2 and 3, PBM or sham in counterbalanced order, both within at most 2 weeks, same time of day (± 2 h), easy standardised ride the day before (under 3 h, below LT1), no caffeine on test days, same diet.
What was applied: a Vielight Neuro Gamma 3 (Vielight Inc., Toronto, Canada): four LEDs on the head at 810 nm, pulsed at 40 Hz, 1 cm² spots, three at 100 mW/cm² and one at 75 mW/cm², plus an intranasal LED at 25 mW/cm² in the right nostril; the configuration is designed to target the prefrontal cortex. 20 min. Energy delivered in joules per site or in total is not reported, nor whether the head LEDs touched the skin.
When: the 20-min application ended 3 min before a 10-min warm-up at 80% of LT1; the steady ride started 3 min after the warm-up, and the time trial 3 min after the steady ride.
Sham: the same device in the same position, not switched on. Whether riders could tell active from sham was not asked or reported.
Rides: 15 min at LT1 + 5% (246.7 ± 32.2 W), free cadence; 3 min seated rest; a 25-min time trial for the highest sustainable power, only elapsed time shown, no encouragement, water allowed; power averaged in 5-min splits; riders did not see their data until the study ended.
Measures: HR (Polar H10) and Borg 6-20 RPE every 5 min in both rides; earlobe lactate every 5 min in the steady ride.
Analysis: paired t-tests on whole-ride means; two-way repeated-measures ANOVA (condition × time); partial eta squared and Cohen's d; no confidence intervals. Sample size said to be based on previous studies, no calculation given.
Key Results
| Outcome | PBM | Sham | Test |
|---|---|---|---|
| Time-trial mean power | 283.22 ± 41.27 W | 287.06 ± 41.34 W | p = 0.105, d = 0.40 (lower with PBM, not significant) |
| Time-trial power over 5-min splits | — | — | condition p = 0.411; interaction p = 0.952 |
| Time-trial mean HR | 165.89 ± 10.36 bpm | 167.17 ± 10.12 bpm | p = 0.361, d = 0.22 |
| Time-trial HR over time | — | — | interaction p = 0.088 |
| Time-trial RPE | — | — | condition p = 0.333; interaction p = 0.681 |
| Time-trial power/HR, power/RPE | — | — | interaction p = 0.859 and p = 0.583 |
| Steady-ride mean HR | 147.06 ± 11.59 bpm | 147.33 ± 11.60 bpm | p = 0.793, d = 0.06; interaction p = 0.536 |
| Steady-ride lactate | — | — | condition p = 0.235; interaction p = 0.407 |
| Steady-ride RPE | — | — | condition p = 0.297; interaction p = 0.342 |
| Steady-ride power/HR, power/RPE | — | — | interaction p = 0.469 and p = 0.364 |
All measures changed over time in both conditions (time effects p < 0.001 to p = 0.009); none differed by condition. The figure caption notes a small non-significant tendency to higher power with PBM early in the time trial, converging later.
No side effects were reported.
Practical Application
In 18 trained cyclists, after 20 min of near-infrared light on the head before riding, 25-min time-trial power, heart rate, lactate and effort did not differ significantly from an inactive sham; mean power was 283 W after the light and 287 W after sham (p = 0.105). The base has no evidence that a head light device before a race or key session helps.
To "should I use a red-light helmet before racing?": say one double-blind crossover tested an 810 nm headset for 20 min before a 25-min time trial and found no significant difference; the authors suggest the device's irradiance was likely insufficient to reach the cortex, so stronger or repeated protocols are untested rather than ruled out.
Inference, not tested here: sleep and the warm-up stay the pre-race priorities (SYNTH-07); by the general rule of SYNTH-83 every element needs a reason from the base, and that guide's row on red and near-infrared light reports this trial (a head device, no significant time-trial difference from sham).
Limitations
The authors' own: no brain imaging (fNIRS or EEG), so whether light reached the prefrontal cortex is unknown; one application only; no measure of executive function, decision-making or pacing strategy.
18 riders; no confidence intervals and no sample-size calculation; the observed difference in time-trial power (d = 0.40, in favour of sham) was not significant, so a small effect in either direction cannot be excluded.
The dose in joules is not reported; the sham was the device switched off, and blinding was not checked.
No trial registration is reported. One reported value is internally inconsistent (time-trial HR condition effect F = 0.785 given with p = 0.142).
Funding: the funding statement says no financial support was received, while the acknowledgements credit Italian Ministry of Health core funding to IRCCS MultiMedica (an author affiliation). No commercial ties declared; how the device was obtained is not stated; one author is an editorial board member of the journal, and one is affiliated with a professional cycling team.
Male amateurs only.
How this entry was checked and changed
Why this level
a crossover of 18 (15 or more), read in full; the main outcome is time-trial power, a performance measure