For about fifteen years the standard explanation of mental fatigue was ego depletion: willpower is a finite resource, like a fuel tank, and using it on one task leaves less for the next. It was intuitive, it was everywhere in popular productivity writing, and it produced the glucose advice — drink something sugary to top up your self-control.
Then it largely failed to replicate. A large multi-lab preregistered replication found an effect close to zero, and subsequent meta-analyses correcting for publication bias have found little left. The glucose claim in particular did not survive.
This matters practically, because the fuel-tank model implies one set of fixes and the models that replaced it imply another.
What seems to be going on instead
Motivation and opportunity cost. Kurzban and colleagues proposed that the feeling of mental effort is not a fuel gauge but a signal: your brain is computing that there is something more valuable you could be doing, and effort-aversion is the output. On this account fatigue is not depletion but a shifting cost-benefit calculation — which explains a very familiar experience. You are too exhausted to keep working, and then a friend calls with something interesting and you are fine. No tank refilled. The value of the alternative changed.
Attention and boredom. A lot of what gets called fatigue is the difficulty of sustaining attention on something unrewarding. The task has stopped supplying anything, and staying on it becomes effortful in exactly the way attention research would predict.
There may be a genuine metabolic component for sustained hard work. More recent work found that a full day of demanding cognitive control was associated with accumulation of glutamate in the lateral prefrontal cortex, and with shifts toward low-effort, immediately-rewarding choices. This is a single line of evidence rather than a settled finding, and it is much narrower than the old willpower model: it concerns a full day of genuinely demanding work, not deciding what to have for lunch.
The honest summary: fatigue is real, the fuel-tank explanation was wrong, and the replacement is mostly about motivation, attention and the accumulated cost of sustained control.
What this changes
Stop trying to "top up" willpower. Sugar does not restore self-control. If you are tired, you are tired.
Stop treating fatigue as a character failure. If effort-aversion is a signal about value, then losing the ability to continue is information, not weakness. Frequently the correct response is to change the task rather than to push through it.
Take variety more seriously than rest. Switching to a different kind of work often restores capacity better than a break does, because the signal was about this task, not about a depleted store. Rote work in a trough is genuinely productive — see how to improve focus.
Decision volume still matters, for a different reason. The strong "decision fatigue" claims did not hold up either — see decision fatigue, where the famous hungry-judges study is probably a scheduling artefact. But a process full of small choices still gets abandoned, because each choice is a moment to do something else instead. That is why a capture step with five decisions in it fails while a one-gesture one survives.
The things that reliably do restore you
Unglamorous and well supported:
- Sleep. By a wide margin the largest lever, and the one most often traded away. See sleep and memory.
- Real breaks, not screen breaks. Scrolling makes the same demands as the work. What restores is undemanding and involuntary — a walk, being outside. See attention restoration theory.
- Removing the load you are carrying. Every unfinished commitment held in your head costs something continuously — see the Zeigarnik effect and cognitive load. Writing them down is one of the few interventions that produces an immediate, noticeable drop in the feeling of being fried.
- Finishing something. Closure genuinely reduces the residue that makes the next task harder — see attention residue.
Distinguishing fatigue from three things it resembles
Boredom. Feels similar, resolves instantly with an interesting task. If a switch fixes it, it was not fatigue.
Avoidance. Tiredness that appears reliably in front of one specific task is usually procrastination — mood repair wearing a convincing disguise.
Burnout or illness. Persistent exhaustion across everything, including things you want to do, that sleep does not touch. That is not a productivity problem and no technique on this page addresses it. It is worth raising with a doctor, particularly if it has lasted weeks.
That third distinction matters. A great deal of writing about mental fatigue quietly encourages people to optimise their way through something that needs medical or organisational attention.
The practical version
Work in the shape of your day rather than against it. Change the kind of work rather than pushing through the same kind. Take breaks that are not screens. Write down what you are carrying so you stop carrying it. Sleep.
And when you cannot continue, treat that as data about the task rather than a verdict on you.
More in deep work, multitasking, and memory and focus.