Abstracts

Rationale: Published, controlled, studies suggest that neurofeedback training in patients with epilepsy can be an effective means of reducing seizure frequency, even in patients with drug-refractory seizures (Sterman 2000, Tan, Thornby et al. 2009). Sensorimotor rhythm augmentation, the original neurofeedback protocol for seizure reduction in patients with epilepsy, has been shown to be most effective in patients taking fewer anti-seizure drugs, those with better Wonderlic scores prior to training, and those who performed worse on tests with strong motor components at baseline (Lantz and Sterman 1988).  However, there is very little data to predict which patients will have better seizure reduction with training using newer neurofeedback technologies, such as LORETA z-score training. EEG with digital analysis is routinely done during the neurofeedback planning process for these newer techniques. This case series will explore the association between the presence of interictal epileptiform discharges on baseline EEG and seizure frequency change reported by patients during LORETA z-score neurofeedback training. Methods: The records for all consecutive patients with medically-refractory epilepsy seen in the Neurofeedback Clinic at a single academic medical center between November 1, 2013 and May 31, 2017 were retrospectively reviewed. All patients were either not candidates for epilepsy surgery (based on consensus decision of the center’s faculty) or had refused to consider surgery for personal reasons. All patients had at least 10 sessions of neurofeedback training using an individualized LORETA z-score training protocol (1-2 sessions per week for 20-30 minutes per session). Data on patient demographics, epilepsy history, baseline EEG findings and baseline seizure frequency were abstracted and analyzed. Patient-reported seizure frequencies during training were averaged over the duration of training for each patient and outcomes were divided into three categories: (1) a decrease from baseline of greater than 50%; (2) a decrease from baseline of less than 50%; or (3) an increase from baseline of any magnitude. Results: 511 total training sessions were reviewed (n=16, the data from 3 patients were excluded due to missing seizure counts). Mean number of sessions per patient was 31.8 +/- 5.3. Mean patient age was 34.8 +/- 2.9 years with mean duration of epilepsy prior to training of 16 +/- 3 years. Thirteen of 16 patients had focal onset epilepsy. Ten of 16 patients had a visuomorphically abnormal baseline EEG (contained either slowing or epileptiform discharges). Six of 16 patients had epileptiform discharges on baseline EEG. Eight patients reported a decrease in seizure frequency of greater than 50% from baseline. Four patients reported a decrease in seizure frequency of less than 50% from baseline. Four patients reported an increase in monthly seizure frequency, compared to baseline. There was no significant association between seizure frequency outcome class and the presence of epileptiform discharges on baseline EEG. Conclusions: There was no significant association between the presence of interictal epileptiform discharges on baseline EEG and the degree of improvement in seizure frequency with neurofeedback training, although this case series is quite small. Larger studies are needed to more definitively assess the power of this, or other, EEG findings to predict neurofeedback training outcomes in patients with drug-refractory seizures who are not, for either medical or personal reasons, candidates for surgical intervention. Funding: None