Brain and blood cells in young adults with major depressive disorder produce more ATP at rest than healthy controls — but have reduced capacity to increase production under demand. Published in a 2026 study from the University of Minnesota (Cullen et al.), the finding was the first to detect these bioenergetic patterns simultaneously in both brain tissue (via phosphorus-31 MRS imaging) and peripheral blood cells in the same depressed patients.
The pattern is paradoxical only at first glance. The mitochondria in depressed patients are running harder at baseline. They're compensating for something — an efficiency deficit, a signaling demand, an ongoing energetic cost of the depressive state itself. But running at higher baseline leaves less headroom. When cognitive or physical demand increases, the system can't ramp up proportionally. The generator is already near capacity when the lights are off.
This maps onto the phenomenology of depression in a way that neurochemical models don't. The fatigue of depression isn't laziness, low motivation, or suppressed drive in the psychological sense. It's a literal energy production constraint at the cellular level. The brain cells are working harder than normal and have less capacity for more. The subjective experience of exhaustion reflects an objective bioenergetic state.
The clinical implication is diagnostic: measuring mitochondrial function could detect depression through a biomarker rather than a symptom checklist. Blood-based ATP measurement is simpler than brain imaging. If the blood pattern reliably tracks the brain pattern — which this study suggests it does — then depression could be diagnosed like diabetes: a metabolic measurement, not a subjective report.
What this doesn't explain is causation. Does mitochondrial dysfunction cause depression, or does depression cause mitochondrial dysfunction? The compensatory over-production at rest suggests the mitochondria are responding to a demand, not creating one. But the ceiling on ramp-up capacity could be either cause or consequence. The system is running hot. Whether it overheated first or was put on a fire first is, for now, unknown.