# Proof Narrative: The human brain accounts for 2% of body weight but uses 20% of the body's oxygen at rest.

## Verdict

**Verdict: PROVED**

This is one of neuroscience's most cited statistics — and it checks out. Two independent authoritative sources confirm both figures precisely.

## What was claimed?

The human brain is disproportionately expensive to run. Despite being a small fraction of total body mass, it demands a far larger share of the body's oxygen supply just to keep the lights on. This claim captures that disproportion with two specific numbers: 2% of body weight, 20% of resting oxygen. It's the kind of fact that shows up in textbooks, science documentaries, and pop-neuroscience articles — often without a source. Worth checking.

## What did we find?

The evidence here is unusually clean. A landmark 2002 paper in the *Proceedings of the National Academy of Sciences* by Raichle and Gusnard states directly: "In the average adult human, the brain represents about 2% of the body weight" and "the brain accounts for about 20% of the oxygen and, hence, calories consumed by the body." Both quotes are exact matches to the claim, verified against the live published text.

A second independent source — a chapter on cerebral metabolic regulation in *Basic Neurochemistry*, a reference textbook hosted on the NIH's NCBI Bookshelf — confirms both numbers separately: "the brain, which represents only about 2% of total body weight" and "accounts for 20% of the resting total body O₂ consumption." All four statements were verified against live sources; both sources are published on NIH-administered domains.

The two sources agree exactly: 2.0% on weight, 20.0% on oxygen, with zero discrepancy between them.

An independent numerical cross-check further supports the oxygen figure. The brain's normal cerebral metabolic rate of oxygen is about 3.5 mL per 100 grams per minute. For a typical adult brain of roughly 1,400 grams, that works out to around 49 mL of oxygen per minute. The body's total resting oxygen consumption is roughly 250 mL per minute. Dividing: 49 ÷ 250 = 19.6% — which rounds directly to 20%.

No credible source was found disputing either number. The brain's weight fraction varies slightly with body size and sex (roughly 1.6–2.5% across the population), but every source in the literature describes this as "about 2%." The oxygen figure has similarly broad consensus.

## What should you keep in mind?

The "at rest" qualifier in the claim is doing real work. During demanding cognitive tasks, local brain blood flow can surge 30–50% in specific regions, which is why neuroimaging works. But whole-brain oxygen consumption rises only about 1–5% above resting baseline — the brain is already running near full capacity metabolically, even when you're doing nothing. The claim is accurate specifically for the basal, resting state, which is what both cited sources measure.

The 2% figure also reflects an average. Brains vary in size, bodies vary more, and the ratio shifts across age, sex, and body composition. In leaner individuals, the fraction is higher; in heavier individuals, lower. "About 2%" is a population average, not a fixed biological constant.

One minor note: the NCBI Bookshelf oxygen citation matched the verification search at 60% word coverage rather than a full exact match, likely due to Unicode formatting differences in the original text ("O₂" vs "O2"). This doesn't affect the conclusion — the PNAS source independently and fully confirms the same 20% figure.

## How was this verified?

This claim was checked by extracting the specific numerical figures from live, primary sources and comparing them against the claim's stated thresholds, then performing independent cross-checks between sources and adversarial searches for counter-evidence. For the full breakdown of sources, quotes, and logic, see [the structured proof report](proof.md). For step-by-step verification details including computation traces and citation status, see [the full verification audit](proof_audit.md). To reproduce the verification yourself, run [re-run the proof yourself](proof.py).