TMS combined with EEG in genetic generalized epilepsy: A phase II diagnostic accuracy study

Objectives: (A) To develop a TMS–EEG stimulation and data analysis protocol in genetic generalized epilepsy (GGE). (B) To investigate the diagnostic accuracy of TMS–EEG in GGE. Methods: Pilot experiments resulted in the development and optimization of a paired-pulse TMS–EEG protocol at rest, during hyperventilation (HV), and post-HV combined with multi-level data analysis. This protocol was applied in 11 controls (C) and 25 GGE patients (P), further dichotomized into responders to antiepileptic drugs (R, n = 13) and non-responders (n-R, n = 12).Features (n = 57) extracted from TMS–EEG responses after multi-level analysis were given to a feature selection scheme and a Bayesian classifier, and the accuracy of assigning participants into the classes P-C and R-nR was computed. Results: On the basis of the optimal feature subset, the cross-validated accuracy of TMS–EEG for the classification P-C was 0.86 at rest, 0.81 during HV and 0.92 at post-HV, whereas for R-nR the corresponding figures are 0.80, 0.78 and 0.65, respectively. Applying a fusion approach on all conditions resulted in an accuracy of 0.84 for the classification P-C and 0.76 for the classification R-nR.

// Clinical Neurophysiology — 2017. – №128. – С. 367-381