Authors :
Presenting Author: Yash Vakilna, MS – The University of Texas Health Science Center
Ramya Manjunatha, MD – University of Alabama Medical Center
Ajay Deep Kachhvah, PhD – The University of Texas Health Science Center at Houston
Chaitanya Ganne, MD, PhD – The University of Texas Health Science Center at Houston
Jay Gavvala, MD – UT Health Houston
Samden Lhatoo, MD, FRCP – University of Texas Health Science Center at Houston
John Mosher, PhD – The University of Texas Health Science Center
Sandipan Pati, MD – UT Health
Rationale: Closed-loop stimulation using the NeuroPace Responsive Neurostimulator System (RNS) system has emerged as a promising treatment option for patients with drug-refractory epilepsies. Most therapeutic stimulation paradigms in current devices aim to decrease seizure frequency but do not provide tailored strategies to terminate seizures, primarily due to a lack of mechanistic understanding of the cortical effect of electrical stimulation. Unlike the commonly explored stimulation parameters of frequency, pulse width, and current intensity, we aim to study the cortical effects of burst duration parameters via electrical source imaging using 128-channel high-density EEG on patients with RNS implants. Gaining a deeper comprehension of the impact of stimulation parameters would offer a more refined approach to devising personalized stimulation protocols for individual patients.
Methods: 7 adult patients with drug-resistant generalized epilepsy were implanted with the RNS. In this institutional review board (IRB) approved the study, the brain activity of these patients was recorded using a 128-channel EEG system (Brain Vision ActiCHamp Plus). 20 trials consisting of 5 s before and 5 s after the RNS unilateral stimulation were recorded for each stimulation parameter setting. Two sets of stimulation parameters were used in this study:
1. amplitude: 2.5 mA, frequency: 100 pulses per second, pulse width: 120 us, burst duration: 100 ms;
2. amplitude: 2.5 mA, frequency: 100 pulses per second, pulse width: 120 us, burst duration: 5000 ms.
Brainstorm toolbox was used to develop the data-processing pipeline (Fig 1a) to quantify the stimulation-induced change in mean band-specific power and synchrony in the Salience and Ascending Reticular Activating System (ARAS). The location of the RNS implantation was confirmed to be Centromedian sub-thalamic nuclei using CT-MRI co-registration, after performing patient-specific anatomically accurate segmentation of thalamus-optimized multi-atlas segmentation (THOMAS) (Fig 1b).
Results: 1. Thalamic RNS long-bursting stimulation led to a statistically significant decrease in source power (p-values < 0.05), when compared to conventional stimulation, in the ARAS network, but not in the Salience network (Fig 2b).
- δ (1-4 Hz): 2/10 ROI
- θ (4-8 Hz): 2/10 ROI
- α (8-12 Hz): 2/10 ROI
- β (13-30 Hz): 6/10 ROI
- Low γ (30-50 Hz): 4/10 ROI
- High γ (50-70 Hz): 4/ 10 ROI
2. A statistically significant increase in connectivity strength across both ARAS and Salience network was observed (p-value < 0.05) indicating the long-bursting stimulation led to an increase in connectivity (Fig 2c)
- The change in connectivity was observed in β (13-30 Hz) and γ (50-70 Hz)
Conclusions: Thalamic long-bursting stimulation shows significantly different cortical effects compared to stimulation with a conventional 100ms burst duration on the ARAS networks indicating that long-bursting stimulation leads to changes in the vigilance state of the patient.
Further understanding of the cortical network effects of different stimulation parameters may lead to the development of patient-specific tailored stimulation protocol which may improve the overall efficacy of stimulation therapies.
Funding: None.