One third of people with epilepsy (PwE) live with drug-resistant epilepsy (DRE), a condition in which seizures persist despite adequate medication. Epilepsy is now understood as a network disorder that disrupts large-scale neuronal coordination, and alterations in the scale-free properties and fractal organization of brain dynamics may emerge as signatures of pathological brain functions. This study investigates whether the temporal sequence of EEG microstates (the brief, quasi-stable topographies that reflect ongoing functional configurations) shows altered fractal properties in DRE compared with drug-responsive epilepsy (nDRE) and whether these properties change after therapeutic neuromodulation. We tested the following hypotheses: 1) fractal features of brain dynamics are different between people with DRE and people with drug-responsive epilepsy (nDRE), thus qualifying them as a potential diagnostic biomarker; 2) fractal features of brain dynamics can be modulated in people with DRE, thus qualifying them as a potential response biomarker.

Resting-state EEG (eyes closed) was examined in 60 DRE and 60 nDRE patients and in a subgroup of 10 DRE patients recorded before and after Vagus Nerve Stimulation (VNS). The study was approved by the Ethics Committee ‘Lazio Area 2’ (number 99.24CET2 CBM, April 11, 2024). Microstate sequences were extracted, and their temporal structure was characterized using scale-free metrics, including the Higuchi Fractal Dimension and Hurst exponent, which quantify short-time complexity properties and long-range dependencies. While conventional microstate metrics and transition patterns did not differentiate DRE from nDRE, fractal analyses revealed significantly higher scale-free disruption in DRE, indicating less self-similar and more erratic state trajectories. Notably, similar fractal shifts were observed after VNS and accompanied clinical improvement, suggesting that fractal markers are sensitive to therapeutic modulation. These findings indicate that the scale-free structure of brain-state sequences may capture fundamental alterations in cortical dynamics associated with DRE. Identifying fractal biomarkers for DRE could support earlier diagnosis, reduce disease burden, and advance precision approaches tailored to individual network dysfunction.