There is a particular kind of tiredness that follows a few hours of sustained concentration — reading something difficult, holding a demanding conversation, tracking a complicated problem. It resembles physical tiredness closely enough that people describe it in the same words, and it is routinely treated as a lesser thing. A 2025 meta-analysis is a good reason to stop treating it that way.

The Assumption

Physical fatigue has an obvious account: muscles have done work, substrates are depleted, waste products accumulate. Something material happened.

Mental fatigue has no equivalent everyday story, and in its absence one gets supplied — that it is really about motivation, interest or discipline. The sentence ‘you’re not tired, you’re bored’ does a lot of work in ordinary life.

That assumption is testable in a specific way: if mental fatigue is a state rather than an attitude, it should have a consistent neural signature, and that signature should appear across quite different circumstances rather than only where someone is unmotivated.

How the Question Was Asked

A 2025 meta-analysis used activation likelihood estimation — a method that asks where results converge across independent neuroimaging studies rather than reanalysing raw data.

Its inclusion criterion was deliberately broad and is the most interesting design choice in the paper: any medical condition in which patients showed significantly higher levels of fatigue than healthy controls.

That is a transdiagnostic approach. Instead of asking what fatigue looks like in one illness, it asks whether there is something common to fatigue wherever it appears.

A search across three major databases identified 46 eligible studies totalling 2,603 individuals (Brain Communications, 2025).

Activation likelihood estimation is worth one line of explanation, because its logic determines what the result means. Each study contributes the coordinates of the peaks it reported; the method asks where those peaks cluster more tightly than chance would produce. It is therefore a map of where independent teams keep finding something, which is a different and more robust claim than any single study makes.

What Was Found

A widespread cortical–subcortical network involving frontal, limbic, basal ganglia and parietal structures, organised into three main clusters: a frontal–striatal–limbic cluster, a frontal–cingulate cluster, and a parietal cluster.

The authors note these are regions implicated in the cognitive, emotional and somatosensory symptoms associated with mental fatigue — which matches how people describe the experience. It is not only difficulty thinking; it comes with flatness of mood and a distinctly bodily quality.

The transdiagnostic framing is what makes this worth knowing. A network that appears in fatigue across many different conditions is more likely to be about fatigue than about any one illness — and considerably harder to explain as a lack of interest.

Why Concentration Is Expensive

The regions involved are informative about why this happens at all. Frontal and cingulate areas are central to maintaining a goal against competing demands; parietal regions to allocating attention; striatal and limbic structures to motivation and to how effort is valued.

Sustained concentration is not one operation running for hours. It is continuous re-selection — repeatedly suppressing what is more immediately appealing in favour of what was decided earlier. That is exactly the machinery the fatigue network implicates.

There is a second demand running alongside the first. Sustained attention requires not only holding the goal but monitoring whether attention has drifted — a background check that itself consumes resources. Part of what makes a long stretch of concentration costly is that the system is doing the task and supervising itself doing the task at the same time.

It also explains a familiar asymmetry. Concentrating on something absorbing is far less tiring than concentrating on something that is not, because the amount of re-selection required differs. The state and the effort are related, and neither reduces to interest alone.

The related question of what sustained demand does to decision quality is examined in what the decision-fatigue literature does and does not establish.

The Limitation the Authors Name

Quality analysis indicated a moderate risk of bias in the majority of the included studies. That is the authors’ own assessment, and it constrains the whole result.

It does not mean the finding is wrong. It means the studies feeding the meta-analysis had methodological weaknesses common in this literature — small samples, varied fatigue measures, heterogeneous patient groups — and that a converging network built from them inherits those weaknesses.

Heterogeneous patient groups deserve particular mention here, because they cut both ways. Pooling fatigue across many conditions is what makes a transdiagnostic claim possible at all — and it also means the studies differ in nearly everything except the symptom, which is a source of noise as well as of generality.

Coordinate-based meta-analysis also converges on what was published, and published peaks reflect what researchers looked for and what reached print.

The honest summary is that there appears to be a consistent network associated with fatigue across conditions, identified from a literature of moderate quality, and that this is a starting point rather than a settled map.

What It Does Not Establish

It does not identify a cause. A network associated with fatigue could be producing it, responding to it, or reflecting something both share. Cross-sectional imaging comparisons cannot separate those.

It does not diagnose anything. There is no scan for fatigue, no threshold, and these are group-level convergences with wide individual variation.

It says nothing about treatment. Identifying regions involved in a state is not the same as knowing what changes it.

And persistent exhaustion needs a proper assessment. Fatigue accompanies a very long list of medical conditions, some of them straightforwardly treatable. A neuroscientific account of why concentration is expensive is not a reason to skip finding out why it has become unusually so.

What it does support is worth having: the tiredness that follows sustained concentration is a state with a measurable neural signature that recurs across conditions — not a failure of will, and not a lesser category of tired.