Poltorak's new insomnia taxonomy uses physics of sleep-state transitions

Insomnia has long been classified by subjective complaint, yet similar symptoms can stem from vastly different underlying dysfunctions in sleep regulation, state transitions, and architectural sequencing. In a new arXiv preprint (2608.05462), physicist Alexander Poltorak proposes a multiaxial dynamical taxonomy that rethinks how insomnia is categorized. Rather than treating insomnia as a single syndrome, his framework specifies a patient profile by three coordinates: the dynamical operation that fails (e.g., stabilization, boundary transition, spatial recruitment), the sleep stage or boundary where failure occurs, and causal status. Modifiers like age, circadian phase, comorbidity, and night-to-night variability are treated as covariates, not mechanistic classes.

Poltorak adopts a local Landau-Ginzburg formalism from cortical and sleep-dynamics research to verbalize nested hypotheses. He notes that the relaxational form applies only to boundary-local dynamics under a gradient approximation; non-gradient escapes require action or quasipotential methods, while whole-night REM-NREM sequencing needs reactive or oscillatory dynamics. The strongest existing evidence is at sleep onset, where published data are consistent with bifurcation-like bistable dynamics, though a driven smooth transition from homeostatic-circadian ramps cannot be excluded. Other operation classes remain hypothesis-generating. The framework is explicitly a phenomenological organizing model, not a new diagnosis or treatment-selection system, but it is designed to improve clinical communication, patient stratification, and prediction. The 28-page paper includes 3 figures and is categorized under Biological Physics.