Network alterations in temporal lobe epilepsy during non-rapid eye movement sleep and wakefulness

•Low-frequency integration and segregation of IED-free resting state networks are higher in TLE patients than in healthy subjects.•Network changes are less visible in N2 and unrelated to epilepsy, likely because physiologic theta waves mask pathologic activity.•In absence of IEDs, it might be easier to detect epilepsy-related alterations during wakefulness than N2 sleep. Investigate sleep and temporal lobe epilepsy (TLE) effects on brain networks derived from electroencephalography (EEG). High-density EEG was recorded during non-rapid eye movement (NREM) sleep stage 2 (N2) and wakefulness in 23 patients and healthy controls (HC). Epochs without epileptic discharges were source-reconstructed in 72 brain regions and connectivity was estimated. We calculated network integration and segregation at global (global efficiency, GE; average clustering coefficient, avgCC) and hemispheric level. These were compared between groups across frequency bands and correlated with the individual proportion of wakefulness- or sleep-related seizures. At the global level, patients had higher delta GE, delta avgCC and theta avgCC than controls, irrespective of the vigilance state. During wakefulness, theta GE of patients was higher than controls and, for patients, theta GE during wakefulness was higher than during N2. Wake-to-sleep differences in TLE were notable only in the ipsilateral hemisphere. Only measures from wakefulness recordings correlated with the proportion of wakefulness- or sleep-related seizures. TLE network alterations are more prominent during wakefulness and at lower frequencies. Increased integration and segregation suggest a pathological ‘small world’ configuration with a possible inhibitory role. Network alterations in TLE occur and are easier to detect during wakefulness.