The consensus model of sleep and memory is cooperative. Sleep consolidates memories. Disrupted sleep impairs learning. The relationship is sequential: experience first, then sleep processes the experience into durable form. Decades of research in mammals, birds, and insects support this architecture.
Huang, Sigrist, and colleagues, publishing in PLOS Biology in 2025, broke it. Drosophila lacking the Insomniac protein — a Cullin-3 ubiquitin ligase adaptor — sleep drastically less and have robustly enhanced Pavlovian olfactory memory. Not preserved memory despite poor sleep. Enhanced memory. The flies that barely sleep remember better than the ones that sleep normally.
The mechanism is a shared signaling pathway. Insomniac loss elevates PKA (Protein Kinase A) activity in the mushroom body — the fly brain's learning center. Elevated PKA drives both phenotypes: it suppresses sleep and it enhances memory. A genetic screen confirmed the link: mildly reducing PKA rescued sleep and shortened lifespan but further increased memory and mushroom body overgrowth. The pathway runs in one direction. More PKA means less sleep and more memory. Less PKA means more sleep and less memory. They are competing outputs of the same signal.
This is not the standard tradeoff story — the idea that limited resources force a choice between competing demands. Sleep and memory are not competing for energy, time, or neural substrate. They are opposing readouts of the same molecular switch. The switch doesn't allocate between them. It simply has two consequences that point in opposite directions. There is no allocation to optimize. The tradeoff is structural.
The cooperative model assumed sleep and memory were in series: wake encodes, sleep consolidates. The PKA finding shows they can be in parallel and oppositional: the same signal that enhances encoding suppresses the state that was supposed to consolidate it. The flies don't need sleep to consolidate because the elevated PKA provides enough synaptic strengthening to substitute for whatever consolidation sleep would have provided — at the cost of dying sooner.
The structural lesson: when two functions appear cooperative because removing one impairs the other, the relationship might still be competitive at the mechanistic level. The usual experiment — disrupt sleep, measure memory decline — conflates the pathway with the consequence. Insomniac mutants survive because their memory doesn't depend on the sleep they lost. It depends on the signal that took sleep away.