Your patient with fibromyalgia sits down and tells you she has already tried walking, that it made things worse, that she is afraid to try again. You know exercise is first line treatment. The guideline says moderate intensity. She will not manage it, and you know that before you prescribe.
That deadlock is a daily one, and it is exactly what this trial brings new information about.
What was done
The study enrolled 116 adults with fibromyalgia, mean age 47.1 ± 11.9 years, of whom 103 (88.8%) were women. They were followed in Boston between May 2019 and July 2024.
All of them did 16 in-person sessions over four weeks, always with treadmill exercise. Two things varied between groups, combined in every possible way:
Walking intensity. Half the participants walked at 60% to 70% of estimated maximum heart rate, which is genuine aerobic exercise. The other half walked light, keeping heart rate only 5% above resting, well below the aerobic threshold. Same treadmill, same 30 minutes, same supervision.
Brain stimulation. Half received active tDCS during exercise, half received sham, the placebo version of the procedure.
The main outcome was not pain
Here is the detail that changes how the whole paper reads: the primary outcomes were laboratory measures, not symptoms. Pain, fatigue, sleep and quality of life came in as secondary.
The main measure is called CPM, conditioned pain modulation, and it is worth understanding what it is, because without that the result makes no sense.
The procedure goes like this. A controlled painful stimulus is applied to the forearm and the patient rates the pain. The same stimulus is then repeated, but now while the other hand is immersed in water at 10 or 12 degrees, which hurts a good deal. In a nervous system that is working well, this second pain recruits the descending inhibition pathways and the forearm stimulus starts to hurt less. CPM is that difference. In practice it measures how much pain brake the person still has.
In fibromyalgia that brake tends to be weak or absent, and it is precisely the target tDCS aims to restore.
The second measure, temporal summation, is the mirror of the first. 15 identical heat stimuli are applied in sequence and the increase in pain from the first to the last is measured. The more it grows, the more the system is amplifying the signal instead of filtering it.
What they found
In physiology, tDCS worked. Active stimulation improved CPM compared with sham, with a medium to large effect size (Cohen's d 0.66), and this effect did not depend on exercise intensity. On temporal summation there was no relevant change.
In symptoms, nobody stood out. All four groups improved moderately to largely, and none improved more than another:
| Outcome | Mean change in the whole sample | Effect size | Difference between groups |
|---|---|---|---|
| Pain intensity (BPI) | not reported | d = 0.64 | none |
| Pain interference (BPI) | not reported | d = 0.92 | none |
| Disease impact (FIQR) | 11.7 points | d = 0.67 | none |
| Fatigue (PROMIS) | 5.7 points | d = 0.60 | none |
| Depression (BDI) | 3.3 points | d = 0.54 | none |
| Sleep (PSQI) | 1.9 points | d = 0.50 | none |
| Quality of life | 4.4 points | d = 0.23 | none |
tDCS changed physiology but did not improve the symptom. A mediation analysis confirmed the dissociation: the improvement in CPM did not explain the clinical improvement in any model tested.
And the two kinds of walking produced the same result on everything.
Three things that change the conversation with the patient
Intensity was not what made the difference
This is the finding with immediate effect in the clinic. Light walking, kept just above resting, produced the same improvement as walking at 60% to 70% of maximum heart rate, in pain, in fatigue, in sleep and in quality of life.
For anyone treating fibromyalgia, that removes an obstacle. Exercise intolerance, post-exertional malaise and fear of getting worse are real barriers, and the traditional recommendation of moderate to vigorous intensity collides with them every day. The study suggests the barrier can be negotiated at no therapeutic cost, because what seems to count is the regularity of the movement, not how much it raises the heart rate.
This is not a licence to prescribe anything at all. The light group was supervised, with a defined dose and a frequency sustained for four weeks. The reasonable reading is "start where the patient can and make sure she keeps going", not "intensity is irrelevant".
The body of evidence still points to moderate intensity aerobic exercise as the reference, and the paper itself acknowledges this when it says that exercise guidelines for fibromyalgia emphasise moderate to vigorous intensity, and that its result challenges that recommendation. The authors also cite a 2025 meta-analysis in which the effect on pain grows with intensity: 0.82 for light exercise against 2.77 for vigorous aerobic. The same caveat applies as to the trial itself, however: those are comparisons across different studies, not a randomised contest inside the same one.
And there is a limit that goes unnoticed. The trial measured symptoms, not fitness. There was no assessment of cardiorespiratory capacity at all, and the authors list this among the limitations. So what can be said is that light walking matched moderate walking on symptoms, not that it delivers the same conditioning, nor the cardiovascular and metabolic gains that exercise prescription also pursues and that matter in a population that tends to be sedentary.
Better physiology is not the same as a better patient
tDCS did exactly what theory predicted, restoring part of the descending inhibition. And that turned into no clinical gain at all over six weeks.
It is an uncomfortable and useful reminder. A surrogate outcome, however elegant on paper, is not a clinical outcome. When someone presents a technology showing that it normalises pain modulation, the next question has to be whether the patient reported feeling better.
The authors themselves raise three possible explanations, and none of them can be dismissed. Fibromyalgia involves far more than one inhibitory circuit. The effect of exercise may have created a ceiling that hid any additional gain. And four weeks may be too short for a neurophysiological change to translate into symptoms.
Who tends to respond better
Older age appeared consistently as a predictor of improvement, in pain intensity, in pain interference, in disease impact and in sleep. Higher body mass index, on the contrary, was associated with a smaller gain in CPM.
These are exploratory analyses, without statistical power planned for them. They serve as a hypothesis, not as a criterion for selecting who to treat.
What the study does not allow you to conclude
There is an important limit in the design: all four groups received supervised exercise. There is no untreated arm. That means the 11.7 point improvement in FIQR and the effect sizes between 0.5 and 0.9 cannot safely be attributed to exercise. Inside them are also the natural history of the symptom, regression to the mean and the effect of being closely followed for four weeks.
The authors face the objection. They point out that the BPI effect sizes exceed what is usually attributed to placebo in fibromyalgia, around 0.53, and are comparable to drugs such as pregabalin and duloxetine. The argument is legitimate but indirect: comparing across different studies does not replace a control group inside the same one.
Declared limitations
The authors are transparent. Four weeks of intervention is short. The clinical outcomes were secondary and without statistical power of their own, which raises the risk of type II error. There was no measure of cardiorespiratory fitness, which weakens the check that the aerobic arm really delivered an aerobic load. And 33 participants had missing data on the primary outcomes, handled by multiple imputation.
The sample is also heavily selected: around 2,000 people went through screening for 116 to be included. That protects internal validity and charges the price in external validity.
The paper already has replies, and they are divided
A few weeks after publication, the same journal carried two letters about this trial. They agree about tDCS and disagree on exactly the point that matters in the clinic.
Stéphane Perrey (Lancet Reg Health Am 2026;55:101379) reads the result the way this review has read it so far. He argues that low intensity activity, when supervised, achieves comparable improvement, and that guidelines should move away from "intensity mandates" towards individualised prescriptions that match patient preference.
André Pontes-Silva (Lancet Reg Health Am 2026;55:101384) disagrees, and brings the strongest argument against the conclusion about intensity: time-matched protocols are not dose-matched. Thirty minutes at 60% to 70% of maximum heart rate impose a different stimulus from thirty minutes below 40%, and the trial neither reported nor controlled total work, accumulated mechanical load or energy expenditure across groups. For him, the integrated hypothesis of the study is left unsupported.
The consequence is subtle and worth following. Without quantifying the dose, the observed equivalence allows two very different readings: either the effect genuinely does not depend on modality, or the two groups converged on a similar dose without anyone measuring it. The second scenario fits a limitation the authors themselves admit, the absence of any measure of cardiorespiratory fitness. If the aerobic group did not in fact deliver the intended load, the comparison did not test what it claimed to test.
On tDCS the two letters converge. Perrey argues against routinely adopting protocols that add stimulation to exercise, and proposes reserving neuromodulation for situations where clinical benefit has been demonstrated.
For anyone reading with practice in mind, the lesson of the split is direct. A few weeks after publication, specialists are still arguing about whether the trial does or does not license the claim about intensity. That is reason to hold the finding with caution, and one more argument for treating light walking as a point of entry, not as a destination.
What this leaves for practice
Nothing here changes the place of exercise as first line treatment in fibromyalgia, nor does it remove moderate intensity aerobic exercise from the position of target. What changes is what to do when that target is not reachable today.
If your patient does not tolerate moderate intensity, this trial supports starting light and insisting on consistency, rather than chasing a heart rate band she will not sustain. Adherence that exists is worth more than an ideal prescription that never happens.
What it does not license is abandoning progression. Light walking as a point of entry, with a plan to raise intensity as tolerance allows, is a reading that fits this evidence. Light walking as the end point, for everyone, is not.
On tDCS, the honest reading is that the technique does what it promises at the physiological level. As an add-on to an exercise programme that is already well run, over six weeks, it added no measurable clinical benefit.
To follow your patient's response, the two instruments used here are available on the platform: the Brief Pain Inventory, which separates intensity from interference, and that separation is exactly where the effect appeared largest, and the Visual Analogue Scale.