Data Acquisition

Electroencephalographic recordings of the seizure group and control group were reviewed. Total duration of EEG monitoring in the seizure group was 1,213 h and in control group was 461 h. Each raw EEG data record was notch-filtered at 60 Hz, a bandpass filter (1–70 Hz) was applied and reviewed in a longitudinal bipolar montage using CURRY 8 software (Compumedics Neuroscan, Charlotte, NC, USA). LPDs and seizures were initially marked by our investigators (S.F.) and (M.S.) and confirmed by board-certified neurophysiologist (A.S.). In both the groups, LPDs were marked manually for a period of 60 s every 10 min consistently for 1-h duration (i.e., marking was done starting at 1st, 11th, 21st, 31st, 41st, and 51st min of an hour, each for a duration of 60 s). This process was repeated every 8 h until the recording ended. Each record was abstracted into a set of variables for each of these 60-s periods based on the ACNS nomenclature “modifiers” that characterized specific electrographic features of the LPDs. These include frequency, prevalence, sharpness, absolute amplitude, relative amplitude, polarity, number of phases, and plus feature (shown in Table 1). These measures were scored semi-quantitatively for each 60-s period and for a discharge that looked prevalent in terms of morphology and amplitude in that period. A total of 692, 60-s periods constituting 11.5 h in the seizure group and in 328, 60-s periods constituting 5.5 h in the control group were scored. Seizures were marked throughout the recording for patients in seizure group and the total number of seizures in each 8-h EEG epoch was documented. Medical records of all the individuals in the study population were reviewed to collect the following variables: age, etiology, gender, history of seizure disorder and recent episode of seizure preceding monitoring.

Electrographic features of LPDs that were abstracted for each epoch.

LPD, lateralized periodic discharges.