Prototype Implementation of a Wearable Device to Suppress Parkinson’s Tremor in Real Time
Tremor is defined as rhythmic oscillatory activity of bodyparts. An initial classification of pathological tremors is rest tremor (RT) and action tremor (AT). RT has been defined as a tremor affecting a part of the body that is not voluntarily activated; clinically it is assessed when the patient tries to relax and is given an adequate opportunity to relax the affected body part. AT has been defined as a tremor occurring in a body part while voluntarily maintaining a position against gravity (postural tremor), during any voluntary movement (kinetic tremor), or during muscle contraction against a rigid stationary object (isometric tremor). Although RT is one of the cardinal features of Parkinson’s Disease (PD), PD patients are also observed to have postural tremor, kinetic tremor or both. In spite of this classification helps in determining cause, the presentation of tremor syndromes varies, diagnosis is therefore based on a thorough evaluation of the patient’s medical history and physical examination, and in some cases, clinical examinations including single photon emission tomography (PET) and single photon emission computed tomography (SPECT) are required. Pathological tremors impact on several domains of quality of life, from physical to psychosocial, in a large proportion of patients. The drug therapy is mostly used to solve motor symptoms, Levodopa (commonly called LDopa) is the main treatment to treat Parkinson’s disease. Another approach is the use of implantable electrical stimulation; this surgical practice relies on high-frequency (100-200 Hz) electrical stimulation in specific areas of the brain. Both the approaches present some limits; as for the surgical approach, a notable limitation are the criteria for the selection of subjects to undergo surgery, and it is necessary to comply with strict international standardized criteria. As for drug therapy, high levodopa concentration is responsible for various side effects, especially in advanced disease, such as nasea, vomiting and othostatic hypotension. These limitations have prompted experts to investigate new diagnostic and therapeutic methods to suppress tremor efficiently. From a scientific literature and patent review emerged some clinical solutions that allows to detect the tremor, such as different sensors and electriomyography. However, it is necessary a continuous monitoring as long as the analysis of large amount of data at different times of the day. Hence the use of wearable technology may help to overcome these barriers since these devices can, not only monitor the motor symptoms, but also help to improve them. For this purpose, potential methods for suppress tremor that might be implementable on a wearable device were researched. Several 2 studies have shown the positive effects of external feedback (i.e., electrical stimulation) for suppressing tremor. In addition, possible detection algorithms are mostly based on spectral analysis, which allows to distinguish voluntary movements (low frequency) from pathological tremors that can be related to a limited spectrum. The input data to the algorithms are mostly data from inertial sensors or electromyographic signal. In this scenario, the aim of this work is to design from scratch a prototype wearable device with a closed-loop stimulation. The prototype is based on the use of a microcontroller and a 9-axis IMU sensor included in a commercially available Arduino board, while for the stimulation part it was necessary to design a custom-build circuit that would allow adjustment of the parameters of biphasic stimuli delivered to patients with surface electrodes. Both electronic design and firmware were developed. In more detail, the algorithm consists of two parts: the first that is dedicated to acquiring the data at a preset sampling rate, processing it, and detecting the tremor windows as a result of energy assessment around the frequency peak of the spectrum. The second part is the enabling of stimulation for a certain period that is regulated by a pid controller to ensure automatic feedback control. The amplitude of the output stimuli is defined during calibration, since each patient has his/her own different threshold of perception. The work was then focused on the design and fabrication of the case that would enclose the circuit so as to protect it and make it wearable on the wrist. During this phase, care was taken to ensure that the device was stable and adaptable to different wrist sizes.
The aim of this work is to design ad implement from scratch a prototype wearable device to suppress pathological tremors in real time, with closed-loop surface electrical stimulation.
Study of literature and patents
The device, in a prototype form, provides closed-loop electrical stimulation, in that it is applied only when pathological tremor is detected. The detection algorithms and stimulation parameters needed to suppress tremor were derived from a study of the literature. Stimulation parameters are adjusted to the specific patient, given the variability of his or her tremors. Specifically, the stimulation time is regulated by a pid controller, while the amplitude of the biphasic stimuli is determined during the calibration phase. The hardware part was designed to allow the adjustment of the stimuli. In conclusion, these results pave the way for the possible use of this medical device, leading to an improvement of patient quality of life, supporting them in daily routine movements that otherwise would be difficult for them, limiting their authonomy and independence.
Future developments include, first a miniaturization of the hardware since at present the components are being mounted on a matrix board. After which it is necessary to harden the control algorithm, this could be implemented through slide mode control that is a nonlinear control method. Another important future development is to add phase locking to closed-loop stimulation. Considering tremor as an oscillatory motion, it can be related to a sinusoidal signal, the starting point of stimulation must be coincident with the origin of the sinusoidal signal that is the one with 0-phase. Moreover, an in vivo test phase is neede since the device has been tested indirectly.