On brain stimulation in epilepsy

Abstract

Brain stimulation has, for many decades, been considered as a potential solution for the unmet needs of the many people living with drug-resistant epilepsy. Clinically, there are several different approaches in use, including vagus nerve stimulation, deep brain stimulation of the thalamus, and responsive neurostimulation. Across populations of patients, all deliver reductions in seizure load and sudden unexpected death in epilepsy risk, yet do so variably, and the improvements seem incremental rather than transformative. In contrast, within the field of experimental neuroscience, the transformational impact of optogenetic stimulation is evident; by providing a means to control subsets of neurons in isolation, it has revolutionized our ability to dissect out the functional relations within neuronal microcircuits. It is worth asking, therefore, how preclinical optogenetics research could advance clinical practice in epilepsy? Here, we review the state of the clinical field, and the recent progress in preclinical animal research. We report various breakthrough results, including the development of new models of seizure initiation, its use for seizure prediction, and for fast, closed-loop control of pathological brain rhythms, and what these experiments tell us about epileptic pathophysiology. Finally, we consider how these preclinical research advances may be translated into clinical practice.

Keywords: brain-machine interface; epilepsy; feedback control; neuromodulation; optogenetics; seizure.

Figure 1

‘Critical slowing’ predicts the transition into seizure activity. (A) Schematic representations of the energy of a dynamic system (adapted from Scheffer et al.). Perturbations may transiently knock the system from its stable position at the trough of a ‘basin of attraction’, but the system rapidly returns to that state (B). For shallower basins [A(ii)], the force towards the valley is weaker and so the recovery from a perturbation is slower [B(ii)]. The transition to a different state requires a perturbation of sufficient energy to overcome the energy barrier between the states (double-headed arrows in A); this is obviously easier to achieve for shallower basins, which are thus said to lack resilience. (C) The black traces show 12 consecutive responses to an optogenetic stimulation of pyramidal cells, followed by two further responses immediately after, showing the same critical slowing in the response, immediately prior to the onset of seizure-like activity, in a mouse brain slice. (D) The timing of the critical slowing transformation showed a highly significant correlation with the onset of seizures, in a range of acute pharmacological models of ictogenesis. Notably, this plot had a unitary gradient (indicative of a precise temporal correspondence of the two changes) and an intercept of ∼400 s, which is the average advance warning ahead of the seizure onset provided by this biomarker.

Figure 2

Chronic models of epilepsy are likely to yield weaker correlations between critical slowing and seizure onset. (A) Schematic representation of an acute model of ictogenesis, in which a pharmacological manipulation creates an ‘ictogenic ramp’, driving the system monotonically towards seizures. Note the change in response prior to the seizure onset (arbitrary time scale). (B) Similar schematic of a chronic model of epilepsy, in which multiple factors interact, to create a variable seizure risk. These are presumed to summate, although the summation is unlikely to be linear, to create a multifactorial risk, which may be mapped less precisely than for the simple acute models. For instance, there may be times when the critical slowing transition occurs (change in responses marked ‘False warning’), but if the system then retreats from seizure threshold without a seizure actually occurring, the perturbations would give a false-positive warning. Consequently, chronic models may yield less precise correlations between critical slowing and seizure predictability than the acute models (Fig. 1D).