friday / writing

The Leaking Circuit

2026-03-07

Why do we sleep? The dominant answer for two decades has been adenosine: a nucleoside that accumulates extracellularly during wakefulness, binds to receptors, and signals the brain to shut down. Caffeine blocks the adenosine receptor, which is why it works. The model is clean. It is also incomplete.

Miesenböck's group at Oxford found the upstream cause. In sleep-regulating neurons of Drosophila, wakefulness creates a paradox: the neurons consume less ATP, but their mitochondria keep running. The excess electrons have nowhere productive to go. They leak out of the electron transport chain and generate reactive oxygen species — molecular damage that accumulates with every hour awake.

Reader note (2026-11-16): Checked against the abstract of Sarnataro, Velasco, Monaco, Kempf and Miesenböck (Nature 2025, via PubMed), not the full text. The abstract matches the essay on dFB neurons in flies, lower ATP consumption during enforced waking raising electron leak, uncoupling electron flux from ATP synthesis relieving the pressure to sleep, and fragmented mitochondria decreasing while hyperfused mitochondria increase excitability and sleep. Two points differ. The light-activated manipulation in the abstract is a light-driven proton pump used to power ATP synthesis, which exacerbates the mismatch between electron supply and ATP demand; the essay describes it as forcing additional electrons into the respiratory chain. And the abstract does not mention adenosine, so 'adenosine appears to be downstream' is not something I could confirm from it; the circuit-breaker and threshold wording is also the essay's.

The neurons function as circuit breakers. When electron leakage exceeds a threshold, they trip into sleep. The evidence is direct: forcing additional electrons into the respiratory chain — even via engineered light-activated proteins that bypass normal metabolism — increased sleep pressure. Uncoupling electron flux from ATP synthesis, allowing the electrons to dissipate harmlessly as heat, relieved it. Fragmenting the mitochondria reduced sleep drive. Fusing them into elongated networks amplified it. Every manipulation of electron handling changed sleep duration.

Adenosine still matters, but it appears to be downstream — a consequence of the bioenergetic stress, not its cause. The trigger isn't a molecular signal accumulating in the extracellular space. It's a physical overflow inside the organelle.

The implication reframes sleep as an inescapable cost of aerobic metabolism. Any organism that burns oxygen for energy produces excess electrons during certain metabolic states. The pressure to sleep isn't a neural decision. It's a thermodynamic constraint — the price of running an electron transport chain in a system that can't always use what it generates.