Adaptive control of epileptiform excitability in an in vitro model of limbic seizures

Abstract

Deep brain stimulation (DBS) is a promising tool for treating drug-resistant epileptic patients. Currently, the most common approach is fixed-frequency stimulation (periodic pacing) by means of stimulating devices that operate under open-loop control. However, a drawback of this DBS strategy is the impossibility of tailoring a personalized treatment, which also limits the optimization of the stimulating apparatus. Here, we propose a novel DBS methodology based on a closed-loop control strategy, developed by exploiting statistical machine learning techniques, in which stimulation parameters are adapted to the current neural activity thus allowing for seizure suppression that is fine-tuned on the individual scale (adaptive stimulation). By means of field potential recording from adult rat hippocampus–entorhinal cortex (EC) slices treated with the convulsant drug 4-aminopyridine we determined the effectiveness of this approach compared to low-frequency periodic pacing, and found that the closed-loop stimulation strategy: (i) has similar efficacy as low-frequency periodic pacing in suppressing ictal-like events but (ii) is more efficient than periodic pacing in that it requires less electrical pulses. We also provide evidence that the closed-loop stimulation strategy can alternatively be employed to tune the frequency of a periodic pacing strategy. Our findings indicate that the adaptive stimulation strategy may represent a novel, promising approach to DBS for individually-tailored epilepsy treatment.

Artificial Intelligence (AI) in Rare Diseases: Is the Future Brighter?

“…AI technologies can help assess patients' experience through the analysis of patient reported outcomes and increase patient recruitment and engagement through social media [100,102]. Furthermore, it can be used to monitor patient adherence in CTs [17].…”

“…Finally, in reinforcement leaning, the algorithm's aim is to find the most suitable action in order to maximize a reward, which, in turn, depends on the action (e.g., dynamic treatment policies [16] and epilepsy prevention [17]).…”

Future of Seizure Prediction and Intervention

“…There are three approaches to closed-loop intervention, as illustrated in the figure 3: 1) A controller that slowly adapts stimulation parameters over time to maximize therapy (e.g. a gradient descent controller) (Panuccio et al 2013), 2) On-demand therapy (Armstrong et al 2013; Heck et al 2014; Krook-Magnuson et al 2013), 3) Physiologically adapting closed-loop neuromodulation (Little et al 2013; Montaseri et al 2013; Wilson et al 2011). …”

“…This algorithm would establish a functional relationship between the parameters and the symptoms, and iteratively adjust the parameters of the therapy to improve the symptoms. Adaptive reinforcement learning algorithms are well-suited for problems with noisy, nonlinear, and nonstationary signals (Panuccio et al 2013; Prokhorov and Wunsch 1997). This approach requires the algorithm to “learn” from its mistakes in order to descend upon a local solution.…”

“…It is often more acceptable to implement a closed loop algorithm that essentially does the same, but more formally, as a clinician or the patient might do given a user interface. Optimization approaches like this have been tested in animal models with some success (Panuccio et al 2013), but has yet to be used in a clinical trial. The first step towards obtaining Food and Drug Administration (FDA) approval for a closed-loop device is in an advisory role to the patient or clinician, who makes the parameter changes.…”