Tailoring Subthalamic Nucleus Deep Brain Stimulation for Parkinson's Disease Using Evoked Resonant Neural Activity.

Abstract

Deep brain stimulation (DBS) applied to the subthalamic nucleus (STN) can be a highly effective therapy for Parkinson's disease (PD); however, there are significant issues which limit its effectiveness, reliability, and tolerability. Inaccurate electrode implantation is common, which can reduce efficacy and cause side-effects due to the undesired activation of neighboring brain regions (Okun et al., 2005; Paek et al., 2013; Rolston et al., 2016). Consequently, in most centers patients are typically kept awake during implantation surgery so that clinical assessments can help determine whether the positioning of electrodes is acceptable (Chakrabarti et al., 2014). The programming of stimulation is also often suboptimally performed, and there are many examples of patients being suddenly liberated from poor movement when chronically applied settings are changed (Okun et al., 2005; Sommer et al., 2015). Programming also confers a high burden to patients and neurologists, especially in the era of directional electrodes, often requiring multiple sessions over several months to identify the most effective DBS settings (Cagnan et al., 2019). Moreover, once programmed, DBS is then applied invariantly without “adapting” to the real-time fluctuating needs of the patient (Little et al., 2013; Priori et al., 2013). A realistic solution to these issues is to use neuronal signals recorded from DBS electrodes to guide accurate electrode implantation (ideally with patients under general anesthesia), to automate programming, and to act as a feedback signal for continuous “adaptive” control. To achieve these aims, such a neuronal biomarker would likely need to localize to the STN (preferably the dorsal motor region where DBS is usually most effective; Herzog et al., 2004), reflect patient state and therapeutic effects with a reasonable time resolution, and, crucially, be reliably detectable in all patients and conditions, including under general anesthesia. Potential biomarkers that may fulfill these criteria for STN DBS in PD include measures from spontaneous neural activity, such as beta oscillations (Little and Brown, 2012; Priori et al., 2013). However, beta band (13–30 Hz) activity is typically of small magnitude (microvolts) and can be variable across patients (Giannicola et al., 2010), making it challenging to reliably record with high fidelity, particularly using implantable, miniaturized systems (Neumann et al., 2016). Evoked responses elicited by DBS pulses offer alternative biomarkers (Ashby et al., 2001; Baker et al., 2002; Walker et al., 2012; Gmel et al., 2015; Kent et al., 2015; Romeo et al., 2019), including the recently identified phenomenon of evoked resonant neural activity (ERNA) (Sinclair et al., 2018).