Why Stimulants Help Brain Fog but Not PEM
This is the sixth in a series on the energy biology linking ADHD, autism, and ME/CFS. This article explains the striking dissociation between stimulant benefit for cognitive symptoms and their null effect on post-exertional malaise — and what it reveals about an untested treatment combination. The metabolic reserve model posits that ADHD involves a baseline brain energy deficit (Zametkin et al. 1990).
1 The data
A 2025 PNAS survey of 3,925 ME/CFS patients assessed treatment outcomes across multiple domains (Eckey et al. 2025):
- Stimulants improved brain fog in 77.1%
- Stimulants improved fatigue perception in 71.7%
- Stimulants had a net -1.5% effect on PEM
These three numbers appear contradictory. How can something improve fatigue by 72% while simultaneously having zero effect on the defining symptom of the disease?
The metabolic reserve model offers a coherent explanation.
2 What stimulants actually do
Stimulants — methylphenidate, amphetamines, modafinil, solriamfetol — increase dopamine and norepinephrine availability at the synapse. This makes neural signalling more efficient. Each thought, each decision, each act of sustained attention requires fewer firing cycles and therefore less metabolic energy.
In engineering terms: stimulants reduce energy demand per cognitive unit without increasing energy supply.
The subjective experience: thinking becomes easier. The constant effort required just to maintain attention — which is the baseline state in both ADHD and ME/CFS brain fog — eases. The 77% and 71.7% numbers reflect this real improvement in cognitive efficiency.
3 Why PEM doesn’t improve
Post-exertional malaise is not triggered by how hard you feel you’re working. It is triggered by how much total energy you actually spend — relative to your production capacity.
PEM occurs when total metabolic demand exceeds mitochondrial maximum production for long enough to initiate oxidative damage (Walitt et al. 2024). Stimulants do not increase mitochondrial capacity. They do not raise the ceiling. They reduce the cost of one specific category of demand (cognition) without changing the ceiling that determines PEM.
And here is the trap: when cognition becomes easier, you do more.
You feel better. You take on that extra meeting. You catch up on emails. You think “maybe today I can manage a normal afternoon.” The stimulant hasn’t given you more energy. It’s made existing energy cheaper to spend on thinking. So you spend it — plus more, because the improved cognition makes you feel capable.
The result: total energy expenditure increases. The cognitive portion is cheaper per unit, but the number of units increases. Net effect on total metabolic demand: approximately zero. Net effect on PEM: approximately zero.
Under the metabolic reserve model, the -1.5% is not a pharmacological failure — it is a behavioural one. An alternative interpretation: stimulants simply do not affect the PEM mechanism at all, regardless of pacing behaviour, because PEM has a fundamentally different pathophysiology from cognitive fatigue. Distinguishing these interpretations requires a trial combining stimulants with enforced pacing (see below).
4 The untested prediction
The metabolic reserve model makes a specific, testable prediction:
Stimulant-treated ME/CFS patients who maintain strict activity pacing should show LOWER PEM frequency than untreated patients.
The logic: if the stimulant reduces the metabolic cost of cognitive function by (say) 20%, and the patient maintains the same activity level, then 20% of their cognitive energy budget is freed up — added back to their reserve. More reserve = more buffer before PEM threshold is crossed.
Stimulant-treated patients who increase their activity should show HIGHER PEM frequency.
The logic: they feel better, do more, exceed the envelope. The stimulant enabled them to overshoot without realising it.
The net -1.5% in the Eckey PNAS survey is likely the average of these two opposing effects. Some patients paced and benefited. Some patients did more and crashed. The average: approximately zero.
No trial has tested this directly. No study has combined stimulant prescription with enforced pacing protocols. The stimulants + strict pacing combination is the untested intervention that the model predicts should work.
5 A secondary benefit: mitochondrial protection?
There is a separate line of evidence suggesting methylphenidate may have direct mitochondrial effects (Almutairi et al. 2024):
- Increases Parkin expression (a mitophagy regulator — helps clear damaged mitochondria)
- Reduces reactive oxygen species production
- Maintains mitochondrial membrane potential
If confirmed, this means methylphenidate may not just reduce cognitive energy demand — it may also slightly protect the mitochondria that produce that energy. A drug that does both would be particularly valuable in the metabolic reserve framework. However, this evidence is primarily preclinical and has not been validated in ME/CFS patients.
6 The Blockmans RCT
The most rigorous trial of methylphenidate in ME/CFS was a double-blind crossover RCT (n = 60, 4 weeks per arm) (Blockmans et al. 2006):
- Fatigue scores fell significantly on methylphenidate vs placebo
- Concentration improved significantly
- Quality of life and physical functioning did not improve significantly
- In a separate long-term observational follow-up by the same group (n = 149 surveyed of 194 prescribed): 65% had stopped methylphenidate; among those continuing, 48% reported ≥50% fatigue improvement, 62% reported ≥50% concentration improvement
The high dropout rate (65% stopped) is itself informative. Consistent with the model: patients who stopped may have been those who increased activity, experienced worse PEM, and attributed the worsening to the medication. Patients who continued may have been those who maintained pacing and experienced the cognitive benefit without PEM cost.
7 Practical implications
For ME/CFS patients using stimulants:
The benefit is “same amount costs me less” — not “I can do more.” Maintain the same activity level you had before starting the stimulant.
Use external timers, not internal feeling, to decide when to stop. The stimulant suppresses the subjective warning signal. By the time you feel cognitively fatigued on a stimulant, you have already exceeded your envelope.
Track PEM at 24h and 48h after cognitive sessions. If PEM is worsening despite improved cognition during sessions, you are spending the savings rather than banking them.
Consider pacing-first, stimulant-second. Establish a stable, paced baseline first. Then add the stimulant. Monitor whether PEM frequency changes. If it decreases: the combination is working as predicted. If it increases: you’ve increased activity without realising it.
Cardiac caution in POTS patients. Methylphenidate and amphetamines are sympathomimetic. In the 30–40% of ME/CFS patients with comorbid POTS, they can worsen tachycardia. Specific concerns: (a) stimulants + midodrine may cause hypertensive spikes; (b) stimulants + beta-blockers create opposing cardiovascular drives; (c) stimulants + fludrocortisone increase overall sympathetic load. If starting stimulants in a POTS patient, use lowest effective dose, monitor resting and orthostatic heart rate, and co-manage with the clinician prescribing the POTS medications.
Severity matters. The above pacing advice assumes the patient can implement structured cognitive work sessions. Severe and very severe ME/CFS patients — those who are bedbound or housebound — may not be able to self-manage timers, tracking, or activity decisions without caregiver support. Adapt the principles to the patient’s functional level.
Note: These are pacing principles derived from the metabolic reserve model, not prescribing guidance. All medication decisions — including stimulant initiation, dosing, and cardiac monitoring — require a qualified clinician.
Part 6 of a series on the energy biology linking ADHD, autism, and ME/CFS.