Back to search

Πανεπιστήμιο Μακεδονίας

Quantification of upper limb parkinsonian tremor using ubiquitous devices

Abstract

dc:description

Parkinson’s disease is a common neurodegenerative disease, mostly affecting people over 60 years old. Despite being a condition most often associated with the elderly, there are cases of newly diagnosed patients as young as 40 years old. The cardinal disease manifestations are generally evident only after the damage in the neurons has already progressed significantly, increasing the need for early biomarkers. The disease affects motor and executive function and cognitive control, and is chronic but not fatal. Depending on age of onset, people can live with Parkinson’s for as many as two or even more decades. However, as the disease progresses and symptoms persist and worsen, the quality of life for PD patients deteriorates gradually, to the point of patients being unable to execute daily activities without assistance. Unfortunately, there is no cure for PD. Thankfully, several drugs and treatment regimens have been developed, which can, depending on the situation, greatly improve the patients’ ability to move and function without being hindered by the symptoms of the disease, ultimately improving their overall quality of life. However, not all treatments are equally effective for all patients, while the most potent substances, such as levodopa, have side effects after a few years of use. For that reason, it is important that treating physicians keep medication prescriptions optimized, varying the medication and keeping dosages at the minimum effective level. To accomplish that, the current practice dictates multiple and long medication titration cycles; physicians, often use trial-and-error to optimize the medication regimen for each of their patients. During this process, however, it is very common for the patient to either have to tolerate untreated symptoms or to ingest more chemicals than required, due to the lack of objective symptom measures or availability of timely scheduled clinical examination visits. PD management suffers from: 1. the lack of objectivity in measuring PD symptoms, caused both by sparse and short clinical visits, and by the inherent inability of most patients to be precise in their self-assessment, particularly retrospectively; 2. the decreasing availability and high cost of expert medical service, and 3. the de facto need for tailored treatment, due to the nature of the disease and the manner in which it affects patients.These considerations create a high need for tools and methods for remote, continuous, objective, ubiquitous, unobtrusive and user-friendly identification, quantification, measuring and monitoring of the cardinal disease manifestations. This research was designed to investigate new methods that could be incorporated into clinical practice to help carers provide better, both in quality and in timing, thus reducing the severity of the symptoms and lessening the burden of PD on the affected population and healthcare institutions. The aim was to design tools based on ubiquitous technology and hardware that can identify, quantify and monitor PD motor symptoms, as they manifest in the patients’ upper limbs. More specifically, this thesis describes two distinct tools, one harnessing the inertial measurement units (IMU’s) embedded in modern smartphones, i.e., accelerometer and gyroscope, and another which exploits a line-drawing task performed on a widely available digital tablet, to initially identify, then quantify and finally monitor motor symptoms, such as tremor and rigidity in the hands of PD patients. For our smartphone-based tool, we built software that is accessible through a simple web page and collects signals from the smartphone’s IMU sensors. Then, through appropriate signal processing we extract useful features and train machine learning models to classify users with high accuracy. We explore the correlation between our measurements and PD-specific clinical scores assigned by an expert of the proposed solution with PD-specific scores acquired through clinical examination conducted by an expert physician. Our tablet-based tool involves software that can detect PD related manifestations when patients draw a simple horizontal line. This task greatly decreases the complexity and duration of the test required to quantify PD symptoms, without sacrificing accuracy. We explore the reliability of the approach by investigating how well it can detect disease characteristics, such as laterality, its correlation with a subject’s age, speed of execution and clinical assessment. Three pilot clinical trials were conducted in the context of this work, in order to validate the proposed tools and analysis methods, involving a total of 122 subjects, 57 PD patients and 65 healthy volunteers. These trials serve as proof of concept and explore the feasibility of incorporating the proposed quantification methods into clinical practice, either as clinical examination complementary tools or as remote management platforms. The tools proposed in this work could be used both during a traditional clinical visit examination or at home, without requiring the presence of a trained expert or even a caregiver, with the data collected being sent directly to the treating physician to be reviewed remotely and asynchronously. The proposed PD quantification methods, once incorporated as complementary clinical tools, could potentially 1. allow a smaller number of trained experts to provide services to more patients, in a continuous, unobtrusive and cost-effective manner, 2. make available objective measures of PD leading to actionable insights and increased assessment accuracy, and3. assist in tailoring the PD treatment regimens much faster than with traditional expert-opinion-based, trial-and-error iterations.

Degree

thesis:*
Grantor dc:publisher
Πανεπιστήμιο Μακεδονίας
Year dc:date
2019

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Kostikis, Nikolaos
  • Κωστίκης, Νικόλαος

Subjects

dc:subject × 27

Rights

dc:rights
Statement dc:rights
  • BY_NC_SA
Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/46709

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Kostikis, Nikolaos; Κωστίκης, Νικόλαος. Quantification of upper limb parkinsonian tremor using ubiquitous devices. Πανεπιστήμιο Μακεδονίας, 2019. http://hdl.handle.net/10442/hedi/46709