Tremor evaluation using smartphone accelerometry in standardized settings

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Associated Data

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Abstract

Tremor can be highly incapacitating in everyday life and typically fluctuates depending on motor state, medication status as well as external factors. For tremor patients being treated with deep-brain stimulation (DBS), adapting the intensity and pattern of stimulation according the current needs therefore has the potential to generate better symptomatic relief. We here describe a procedure for how patients independently could perform self-tests in their home to generate sensor data for on-line adjustments of DBS parameters. Importantly, the inertia sensor technology needed exists in any standard smartphone, making the procedure widely accessible. Applying this procedure, we have characterized detailed features of tremor patterns displayed by both Parkinson’s disease and essential tremor patients and directly compared measured data against both clinical ratings (Fahn-Tolosa-Marin) and finger-attached inertia sensors. Our results suggest that smartphone accelerometry, when used in a standardized testing procedure, can provide tremor descriptors that are sufficiently detailed and reliable to be used for closed-loop control of DBS.

Introduction

Therapeutic neuromodulation, for example deep-brain stimulation (DBS), can effectively ameliorate tremor in neurological conditions such as essential tremor (ET) and Parkinson’s disease (PD). It is thought, however, that significant improvement in treatment efficacy could be achieved if the stimulation protocols were better adapted to the changing unique needs of the patient. For this reason, closed-loop neuromodulatory systems, where sensor data, informing on the current tremor state, is used as feedback control of the stimulation device has been attracting a growing interest. In principle, any electrical closed-loop neuromodulation system contains three main components that jointly determine system characteristics, and which each presents a separate engineering challenge. That is, (1) sensors that can reliably estimate severity and characteristics of symptoms being treated, (2) control algorithm converting sensor signals into appropriate adjustments of the neuromodulatory stimulation patterns, and (3) one or more stimulating devices passing current into the tissue to modulate neuronal activity. Designing an efficient closed-loop neuromodulation system to improve treatment of tremor therefore requires knowledge of both symptom characteristics and of the underlying pathophysiology. During the last three decades, certain advances have been made in this latter respect in ET and PD. When bilateral upper limb action tremor is present in the absence of other neurological signs, the condition is commonly classified as ET (Bhatia et al., 2018). Several studies point to a central origin of ET and it has been hypothesized that synchronized rhythms in brain networks may generate oscillation frequencies that are being transmitted to the muscles. In line with this notion, thalamotomy or DBS of the ventral intermediate nucleus (VIM) of the thalamus is known to result in effective symptomatic treatment (Vaillancourt et al., 2003; Pedrosa et al., 2014; Huss et al., 2015). Consequently, VIM stimulation has been the preferred surgical target for the treatment of ET and is recommended primarily in elderly with medication refractory ET (Dallapiazza et al., 2019). In a similar way, in PD, thalamic neurons have been found to discharge rhythmic bursts at 3–6 Hz that correlate with limb tremor (Lenz et al., 1988) and VIM stimulation is known to be an effective treatment for PD tremor as well (Cury et al., 2017). However, to achieve symptomatic relief covering the broader range of motor complications in PD, including rigidity and bradykinesia, the subthalamic nucleus (STN) is generally considered a preferred target. Interestingly, in PD patients with severe tremor at rest, the thalamic bursting rhythm can be the predominant oscillation frequency also in the STN (Alonso-Frech et al., 2006).

With respect to characterization of symptoms, a rapid development has recently taken place in the use of inertia sensor techniques. In particular, the development of a range of body-worn sensors incorporated in consumer products such as watches and the almost ubiquitous use of smartphones in everyday life has opened-up for many new ways to monitor movement disorders using commercially available products. However, although several studies have shown a good correspondence between accelerometer data and clinical scores (see e.g., Senova et al., 2015; Haubenberger et al., 2016; Longardner et al., 2019; Vescio et al., 2021) only a few technical solutions have yet been developed to a technological readiness level that is approaching clinical requirements (Luis-Martínez et al., 2020). Thus, in this study we have investigated to what extent the accelerometry technology widely available in smartphones could provide sufficiently detailed characterizations of tremor in ET and PD patients to be used in a closed-loop controlled DBS device; under the assumption that the metric created could be implemented as a simple control algorithm to modify stimulation features. In this context, we foresee that a calibration procedure using a standardized setting will be faster and more reliable than on-line monitoring of movements across widely differing behavioral states. Hence, the data analyzed have been recorded during standard neurological assessments of postural tremor that would be trivial for the patients to carry out independently at home on demand. Finally, to assess the potential limitations of smartphone accelerometry data, phone recordings were here directly compared to miniature inertia sensors attached to the index finger.

Materials and methods

Subjects

In total, 33 subjects were included in the study. Of these, 17 with ET (12 males, 5 females) with an average age of 68 years (28–89), 9 with PD tremor (5 males, 4 females) with an average age of 75 years (57—79) were recruited. In addition, seven subjects (1 male, 6 females) with an average age of 51 years (21–77) with diabetes mellitus type 2 without any neurological complications were also included in the study as control group (the study was performed during Covid-19 pandemics, which created difficulties to recruit research persons to the control group. We have therefore used patients with diabetes as control group because of practical and safety issues). The diagnosis of ET was confirmed using the neurological examination and the WHIGET diagnostic criteria (Louis and Pullman, 2001), whereas diagnosis of PD was confirmed using MDS clinical diagnostic criteria for Parkinson’s disease (Postuma et al., 2015). Subjects who were diagnosed with other forms of chronic motor system dysfunctions (e.g., previous stroke or tumor with persistent, significant motor impairments), hallucinations, alcoholism, drug addiction, dementia, or were on medications that can cause tremor or motor impairments, were excluded from the study.

This study was approved by the Swedish Ethical Review Authority with a diary number of 2021-00503 and all of the participants included signed written informed consent. The detailed descriptions of participants are listed in Table 1 .

TABLE 1

Clinical profile of subjects with essential tremor and Parkinson’s disease.