Embedding digital chronotherapy into medical devices - A canine case study in controlling status epilepticus through multi-scale rhythmic brain stimulation

Mayela Zamora, Sebastian Meller, Filip Kajin, James J. Sermon, Robert Toth, Moaad Benjaber, Derk-Jan Dijk, Rafal Bogacz, Gregory A. Worrell, Antonio Valentín, Benoit Duchet, Holger A. Volk, Timothy DenisonarXiv 2021 · arXiv

Abstract

This case study details the successful application of "digital chronotherapy" via a deep brain stimulator in a canine patient with drug-resistant epilepsy. By synchronizing stimulation with circadian rhythms and utilizing low-frequency entrainment, researchers eliminated status epilepticus (SE) events and reduced reliance on rescue medications.

Introduction

Epilepsy is characterized by seizures that often follow predictable temporal patterns, including circadian (24-hour) and infradian (multi-day) cycles. Conventional Deep Brain Stimulation (DBS) devices typically provide "tonic" high-frequency stimulation (>100Hz) to suppress seizure propagation. However, this fixed approach ignores biological clocks and can lead to side effects or failure in preventing "cluster seizures" from escalating into Status Epilepticus (SE)—a life-threatening medical emergency. This study investigates a device capable of sensing biological rhythms and employing low-frequency stimulation to entrain healthy brain activity.

Methods and System Design

The study utilized the Picostim–DyNeuMo research system, a cranially-mounted stimulator implanted into the centromedian nucleus of the thalamus (CMN) of a 60kg canine.

Overview of the Picostim–DyNeuMo system, including the multi-scale control scheme (A), implantable hardware (B), and cranial placement (C, D).

Multi-Scale Control Algorithm

The device implemented a three-layered stimulation strategy:

  1. Circadian Basal Rate: A low-intensity 13Hz signal delivered during active daylight hours.
  2. Protective Sleep Mode: Activated by an onboard accelerometer during periods of inactivity (>4 minutes), increasing stimulation amplitude during high-risk sleep.
  3. HF Boost Mode: A high-frequency (130Hz) burst triggered by the owner (via a tap on the dog's head detected by the accelerometer) to abort active seizures.

Arnold Tongue Analysis and Entrainment

Rather than arbitrary frequency selection, the team employed a mathematical framework (Arnold Tongues) to identify frequencies for reliable entrainment. The goal was to synchronize the brain's natural 13Hz rhythm while avoiding the 2Hz oscillations associated with pathological seizure activity.

Power Spectral Density (PSD) analysis identifying the target frequency for entrainment (13Hz) and the pathological frequency to avoid (2Hz).

Key Insights

  • Entrainment vs. Suppression: Unlike traditional DBS which "jams" signals, this study suggests that "guiding" the brain into healthy rhythms (entrainment) using lower frequencies is more effective and better tolerated.
  • Feedforward Control: By utilizing a clock and accelerometer to anticipate seizure risk (e.g., during sleep), the device proactively adjusts therapy rather than reacting to onset.
  • Veterinary Medicine as a Translational Bridge: Treating veterinary patients with naturally occurring diseases provides a vital proof-of-concept for complex neurotechnologies prior to human clinical trials.

Results

Over a seven-month post-implant period, the following outcomes were observed:

  • Elimination of Status Epilepticus: Zero SE events occurred, compared to frequent events pre-implant.
  • Medication Reduction: The patient transitioned to phenobarbital monotherapy with reduced dosage, eliminating several anti-seizure and rescue drugs.
  • Boost Mode Efficacy: The owner-triggered "boost" mode successfully interrupted ongoing seizures in 64% of attempts.
  • Quality of Life: Caregivers reported significantly reduced anxiety and improved quality of life for the animal.

Limitations and Future Work

As an N=1 case study, results may not generalize across all epilepsy types. The study relied on manual seizure diaries and lacked a blinded, sham-controlled phase due to the experimental nature of the treatment. Future iterations may integrate automated "closed-loop" sensing to adapt to longer infradian cycles and further refine the "digital zeitgeber."

Summary updated

BibTeX
@misc{zamora2021embedding,
  author = {Mayela Zamora and Sebastian Meller and Filip Kajin and James J. Sermon and Robert Toth and Moaad Benjaber and Derk-Jan Dijk and Rafal Bogacz and Gregory A. Worrell and Antonio Valentín and Benoit Duchet and Holger A. Volk and Timothy Denison},
  title = {{Embedding digital chronotherapy into medical devices - A canine case study in controlling status epilepticus through multi-scale rhythmic brain stimulation}},
  year = {2021},
  eprint = {2107.02993},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2107.02993},
}
Figure