Vagal nerve stimulation in patients with drug-resistant epilepsy acts differently on the electroencephalographic aperiodic components of responder patients compared with non-responders

Rationale The mechanisms through which Vagal Nerve Stimulation (VNS) acts on seizure frequency and quality of life in the treatment of epilepsy are currently unknown: no study has yet focused on the action of VNS on the aperiodic component of the EEG, which is a power spectrum 1/f-like signal related to biological functions1 (aging and cognitive processes) and processes of neuronal excitation and inhibition in different diseases.2 This study aims to investigate the effect of VNS on the Exponent and Offset aperiodic parameters during resting state on 64ch EEG tracks in patients affected by drug-resistant epilepsy. Methods Among the patients referred to our Centre, 10 patients with drug-resistant epilepsy, not eligible for epilepsy surgery and without psychiatric comorbidities, were selected. Each patient underwent a 64-ch EEG recording in resting state before and one year after VNS implantation; clinical response was computed by McHugh's Class.3 On each trace, a neurologist experienced in electroencephalography selected 20 epochs of 8 seconds (sampling rate at 1024 Hz), free of artifacts and interictal abnormalities. The aperiodic parameters were computed over the entire scalp and in individual channels on each extracted epoch. The differences in these parameters before and after VNS were studied with a U test, in responders (McHugh Class I-II) and non-responders (McHugh Class III-IV) separately. Results Exponent and offset parameters in the global scalp analysis showed a decrease in the responders and an increase after VNS in the non-responders. (Table 1) The analysis of individual channels showed similar behaviors. Responders exhibited higher exponent values before VNS in the frontal and temporal channels whereas non-responders after VNS in the frontal, temporal, parietal and central channels. In the responders' group, the offset parameter was higher before VNS in frontal, temporal, occipital and central channels, and in non-responders' group after VNS in all channels. (Table 2) Conclusions Despite the need of an analysis on a larger sample and the limits of a scalp-derived signal, our results show that VNS acts on aperiodic EEG components differently in responders and non-responders, suggesting a possible reduction in neuronal excitability in responders. Further studies should be carried out to identify which patients will exhibit a good response to VNS, to treat only those most likely to benefit it.