Model-based optimized phase-deviation deep brain stimulation for Parkinson ’s disease

Abstract

High-frequency deep brain stimulation (HF-DBS) of the subthalamic nucleus (STN), globus pallidus interna (GPi) and globus pallidus externa (GPe) are often considered as effective methods for the treatment of Parkinson’s disease (PD). However, the stimulation of a single nucleus by HF-DBS can cause specific physical damage, produce side effects and usually consume more electrical energy. Therefore, we use a biophysically-based model of basal ganglia-thalamic circuits to explore more effective stimulation patterns to reduce adverse effects and save energy. In this paper, we computationally investigate the combined DBS of two nuclei with the phase deviation between two stimulation waveforms (CDBS). Three different stimulation combination strategies are proposed, i.e., STN and GPe CDBS (SED), STN and GPi CDBS (SID), as well as GPi and GPe CDBS (GGD). Resultantly, it is found that anti-phase CDBS is more effective in improving parkinsonian dynamical properties, including desynchronization of neurons and the recovery of the thalamus relay ability. Detailed simulation investigation shows that anti-phase SED and GGD are superior to SID. Besides, the energy consumption can be largely reduced by SED and GGD (72.5% and 65.5%), compared to HF-DBS. These results provide new insights into the optimal stimulation parameter and target choice of PD, which may be helpful for the clinical practice.

Introduction

Parkinson’s disease (PD) is a degenerative disease associated with loss of dopaminergic cells in the substantia nigra pars compacta (SNc) (Obeso et al., 2000). These cells regulate the activities of the basal ganglia (BG) mainly through innervation of the striatum (Stein & Bar-Gad, 2013). Previous experimental studies have shown that pathological synchronous bursting activities occur in the BG of PD patients and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated primates (Galvan and Wichmann, 2008, Wichmann and Soares, 2006). To understand mechanisms underlying these pathological activities, knowledge of changes in neural nonlinear dynamic behaviors is required. The development of mathematical models provides useful tools for the study of abnormal synchronous oscillations of PD (Rubin, 2017). Using these ideas and methods of nonlinear dynamics, the pathogenic mechanism of these oscillations can be understood (Rubchinsky, Park, & Worth, 2012). Simulation results show that synchronized activity disrupts the thalamic reliability neurons (TC) to transmit sensory motor signals which may lead to akinesia, rigidity and other symptoms (Pirini et al., 2009, Rubin and Terman, 2004, van Albada and Robinson, 2009). These results can provide theoretical evidence for clinical applications.

So far, deep brain stimulation (DBS) has been considered as the most effective way to treat PD. It has been reported that the degree of Parkinson’s pathological symptoms decreases almost linearly with the increase of DBS frequency which is usually 130–185 Hz (Santaniello, Gale, & Sarma, 2018). The effectiveness of DBS relies on the parameters of the electrical pulses, including frequency, pulse width, and amplitude. In the treatment of most patients, the average stimulation frequencies are determined as 152 Hz for subthalamic nucleus-DBS (STN-DBS) and 162 Hz for globus pallidus internus-DBS (GPi-DBS) (Okun et al., 2012). The parameter configuration of DBS is different from patients to patients (Picillo, Lozano, Kou, Puppi Munhoz, & Fasano, 2016). Plenty of studies have shown that well-designed irregular, low-frequency stimulation patterns may have clinical advantages compared with the high-frequency, conventional DBS (Akbar et al., 2016, Brocker et al., 2017, Brocker et al., 2013, Santaniello et al., 2018), and they have highlighted the potential superiorities of low frequency DBS and encouraged further development of DBS optimization issues. Wongsarnpigoon and Grill proposed to replace the rectangular waveforms with the Gaussian energy efficient DBS waveforms (Wongsarnpigoon & Grill, 2010). Daneshzand et al. introduced a Gaussian Delay Gaussian (GDG) waveforms to disrupt pathological synchronization behavior (Daneshzand, Faezipour, & Barkana, 2017). The stimulation signal is usually performed on a fixed target area of the BG. STN and GPi are usually standard anatomical targets for DBS. Due to the characteristics of HF DBS and long-term processes (Rizzone, 2001), the stimulation intensity is often disruptive and excessive. In clinical applications, stimulation parameters are critical, and stimulation needs to be effective enough to achieve energy savings and avoid side effects. For this purpose, some multi-site and multi-nuclear stimulation strategies have developed. Guo and Rubin developed a multi-site stimulation of STN to diminish TC errors (Guo & Rubin, 2011). Popovych and Tass introduced a coordinated reset stimulation which was stimulated by multiple stimulation sites to achieve a spatially coordinated phase reset in the subpopulations (Popovych & Tass, 2014). Then, a new triplet-structure coordinated reset multi-nuclear stimulation paradigm was proposed to the treatment of PD by Fan and Wang (2015). Meanwhile, a novel pulsatile multisite linear delayed feedback stimulation method is developed (Popovych & Tass, 2018), which can achieve a better desynchronization effect. This feedback stimulus which is more intelligent than the normal DBS can be adjusted based on the brain’s electrical signals. So the energy required to perform this stimulus can be maintained at a low level.

BG is a set of interconnected subcortical nuclei of the brain which is related to PD. The loss of dopaminergic input changes the dynamics of the BG, which affects the thalamus and cortex and in turn affects the cerebellum (Bostan & Strick, 2010). The best targets for implementing DBS depend on electrophysiology and macro stimulation, and the most common targets for DBS in PD are GPi and STN. Although surgical experience with globus pallidus externa (GPe) is so far not that positive, some modeling and physiological results show that targeting GPe for DBS to eliminate Parkinson’s motor symptoms is much better than GPi-DBS (Daneshzand, Ibrahim, et al., 2017, Pirini et al., 2009, Vitek et al., 2004). Preceding multi-nuclear stimulation strategies do not consider the optimal combination of the stimulus targets. Innovations in stimulating targets have rarely been considered in previous studies, and it is necessary to find the best combination of stimulating targets to decrease stimulation frequency and reduce the side effects of stimuli. This paper aims to propose new stimulation methods to decrease the symptoms of PD and reduce energy consumption.

It is of great significance to study the dynamic behavior of PD, so various mathematical models have been proposed to reproduce the physiological and pathological activities of these nuclei (Pirini et al., 2009, Rubin and Terman, 2004, Schiff, 2010, So et al., 2012, Terman et al., 2002, van Albada and Robinson, 2009). Based on the basal ganglia-thalamic (BGTC) circuit, Rubin and Terman established a model composed of several different types of neuronal populations (Rubin & Terman, 2004), which is improved by reconstructing the network topology (So et al., 2012). Here we use this model of BGTC network to study the effect of combined DBS of two nuclei (CDBS) on PD. DBS techniques have been employed to disrupt the synchronization behavior (Liang and Wang, 2019, Montaseri et al., 2015, Qin et al., 2013). Therefore, the numerical study of CDBS is mainly based on its ability to de-synchronize, restore the relay capability of the TC, and reduce the energy of the beta band (13–30 Hz). We first propose three stimulation strategies based on different combination of stimulating targets: STN and GPe CDBS (SED), GPe and GPi CDBS (GGD), STN and GPi CDBS (SID). The phase difference between two DBSs would be an important factor of neuron behaviors in the BG, which will be considered in CDBS. Further, we explore the optimal stimulation parameters for the CDBS with phase difference.