Authors :
Presenting Author: Dhinakaran Chinappen, PhD, MBA, MEng – Massachusetts General Hospital, Harvard Medical School, Kennedy Krieger Institute, Johns Hopkins University
Wen Shi, MD, MS – Massachusetts General Hospital and Harvard Medical School
Kate Isaac, BS – Massachusetts General Hospital / Harvard Medical School
Katherine Walsh, BS, PhD Student – Massashusetts General Hospital/Harvard Medical School
Peter Hadar, MD, MS – Massachusetts General Hospital; Harvard Medical School
Hunki Kwon, PhD – Massachusetts General Hospital, Harvard Medical School, Kennedy Krieger Institute, Johns Hopkins University
Mark Richardson, MD, PhD – Massachusetts General Hospital
Uri Eden, PhD – Boston University
Xue Han, PhD – Boston University
Mark Kramer, PhD – Boston University
Catherine Chu, MD, MSC – Massachusetts General Hospital, Harvard Medical School, Kennedy Krieger Institute, Johns Hopkins University
Rationale:
Neuromodulation is a promising, minimally invasive treatment option to reduce pathologic epileptic activity in patients with drug refractory epilepsy. However, the parameters that optimally disrupt epileptic activity are poorly understood. This study evaluated the impact of different stimulation frequencies to reduce epileptic activity in the epileptogenic zone of patients with drug resistant epilepsy (DRE).
Methods:
We prospectively recruited patients with DRE undergoing intracranial EEG monitoring (n = 11). The epileptogenic zone (EZ) was estimated as the channel with the highest spike ripple rate during a baseline recording. Each patient received 180 minutes of open-loop neurostimulation in the estimated EZ (Cerestim device, Blackrock) during which 10s trials consisting of 100ms 0.5–1 mA, 100μs pulse width burst stimulations every 900ms followed by 5 second breaks. During each 10s trial one of five stimulation frequencies (1, 20, 40, 140, 200 Hz) was delivered in a randomized order. Spikes and spike ripples were automatically detected using validated detectors and the rates of each event were computed during each of the 10s trials for each frequency, normalized to baseline activity. The relationship between stimulation frequency and event rate was evaluated using mixed-effects linear regression models. Results:
Spikes rates were significantly reduced relative to baseline rates across all frequencies (1 Hz:
p = 0.005; 20–200 Hz:
p < 0.001); where larger effects sizes were observed for higher frequencies (1 Hz: β = −9.06; 20 Hz: β = −11.29; 40 Hz: β = −16.34; 140 Hz: β = −15.82; 200 Hz: β = −16.22). Spike ripple rates were significantly reduced relative to baseline rates at higher but not lower frequencies (40 Hz:
p = 0.028; 140 Hz:
p = 0.006; 200 Hz:
p = 0.004) where larger effect sizes were again observed at higher frequencies (40 Hz: β = −1.07; 140 Hz: β = −1.36; and 200 Hz: β = −1.42). Maximal reductions in spike and spike ripple rates occurred in the first 400ms after stimulation and diminished by 800ms. While higher frequencies (140–200 Hz) showed the greatest efficacy at the group level, different frequencies had maximal efficacy across individuals.
Conclusions:
In conclusion, electrical stimulation reduced epileptiform activity at all stimulation frequencies tested but higher frequencies had larger effect sizes at the group level. Given the variability observed across subjects, personalized parameter testing may be required. Future work is required to evaluate whether suppression of epileptiform biomarkers predicts improved seizure control. Funding:
NIH NINDS R01NS119483