Głęboki sen selektywnie wzmacnia cechy sieciowe przedziału uważanego za źródło padaczki w padaczce ogniskowej
Slow-wave sleep selectively amplifies network signatures of the presumed epileptogenic zone in focal epilepsy
W skrócie
Badacze badali, jak różne fazy snu wpływają na aktywność mózgu u pacjentów z oporną na leki padaczką ogniskową. Odkryli, że podczas głębokie snu (NREM3) aktywność elektryczna w obszarze źródła padaczki znacznie się wzmacnia w porównaniu z resztą mózgu, co można wykorzystać do precyzyjnego zlokalizowania tego obszaru bez konieczności czekania na atak padaczki. Te odkrycia mogą pomóc lekarzom w lepszym planowaniu zabiegów chirurgicznych u pacjentów z padaczką.
Oryginalny abstract (angielski)
STUDY OBJECTIVES: Slow-wave sleep increases interictal epileptiform discharges (IEDs) and seizures in focal epilepsy, but how sleep stages reshape epileptogenic tissue remains unclear. We assessed whether sleep stages selectively modulate network properties of the presumed epileptogenic zone (pEZ) relative to irritative (IZ) and non-involved (NIZ) zones. METHODS: We studied 22 patients with drug-resistant focal epilepsy undergoing pre-surgical intracranial electroencephalography (iEEG). IED-free resting-state epochs spanned four vigilance stages: wakefulness, non-rapid eye movement (NREM) sleep stages 2 and 3, and rapid eye movement (REM) sleep. Spectral power, functional connectivity, graph-theoretic indices, and the 1/f exponent were quantified across seven frequency bands. Sleep stage and zone interactions were tested using permutation testing. pEZ electrodes were classified using elastic net logistic regression with leave-one-subject-out cross-validation. RESULTS: NREM3 selectively amplified within-pEZ gamma-band synchrony beyond surrounding zones (pEZ-NIZ: d=1.204, P=0.027). pEZ local efficiency and betweenness centrality rose in beta1 and gamma bands versus wakefulness. The aperiodic exponent steepened within the IZ versus NIZ during NREM3 and NREM2 (d =1.452 and 0.87, P < 0.001 and P=0.015). A three-feature classifier (NREM3 aperiodic exponent, beta2 connectivity, gamma local efficiency) localised the pEZ without ictal recordings (median AUC 0.859 pEZ vs. non-pEZ). CONCLUSIONS: Slow-wave sleep selectively unmasks pEZ network properties via amplified power, synchrony and hub topology. NREM3-derived multimodal features may complement conventional pre-surgical biomarkers. Prospective validation in larger cohorts is needed to determine whether these markers improve surgical planning and outcome prediction.