Recent machine learning advancements in sensor-based mobility analysis: Deep learning for Parkinson's disease assessment

Eskofier, Bjoern M.; Lee, Sunghoon I.; Daneault, Jean‐François; Golabchi, Fatemeh Noushin; Carvalho, Gabriela Ferreira; Vergara‐Diaz, Gloria; Sapienza, Stefano; Costante, Gianluca; Klucken, Jochen; Kautz, Thomas; Bonato, Paolo

doi:10.1109/embc.2016.7590787

Cited by 131 publications

(100 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the majority of studies (e.g. [16], [17], [19], [21]), wearable inertial measurement units (IMUs) are used with applications in remote home environments in mind, whereas marker-based motion tracking systems are often used in a controlled environment or when they provide a ground truth for other sensors [22], [23]. Existing studies largely focus on one or two PD-specific behavior classes including tremor, bradykinesia, dyskinesia, and gait disturbance, where the presence of abnormality in these classes is estimated by machine learning [22].…”

Section: A Related Workmentioning

confidence: 99%

See 1 more Smart Citation

A Multi-Layer Gaussian Process for Motor Symptom Estimation in People With Parkinson's Disease

Lang

Pfister

Fröhner

et al. 2019

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

The assessment of Parkinson's disease (PD) poses a significant challenge as it is influenced by various factors which lead to a complex and fluctuating symptom manifestation. Thus, a frequent and objective PD assessment is highly valuable for effective health management of people with Parkinson's disease (PwP). Here, we propose a method for monitoring PwP by stochastically modeling the relationships between their wrist movements during unscripted daily activities and corresponding annotations about clinical displays of movement abnormalities. We approach the estimation of PD motor signs by independently modeling and hierarchically stacking Gaussian process models for three classes of commonly observed movement abnormalities in PwP including tremor, (non-tremulous) bradykinesia, and (non-tremulous) dyskinesia. We use clinically adopted severity measures as annotations for training the models, thus allowing our multi-layer Gaussian process prediction models to estimate not only their presence but also their severities. The experimental validation of our approach demonstrates strong agreement of the model predictions with these PD annotations. Our results show the proposed method produces promising results in objective monitoring of movement abnormalities of PD in the presence of arbitrary and unknown voluntary motions, and makes an important step towards continuous monitoring of PD in the home environment.

show abstract

Section: A Related Workmentioning

confidence: 99%

“…For instance, Eskofier et al [17] compare several supervised machine learning pipelines and a deep learning algorithm to detect bradykinesia. Several specific motor tasks of 10 patients with idiopathic PD were recorded using IMU sensors.…”

Section: A Related Workmentioning

confidence: 99%

A Multi-Layer Gaussian Process for Motor Symptom Estimation in People With Parkinson's Disease

Lang

Pfister

Fröhner

et al. 2019

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

show abstract

“…Nowadays, DL is present in commercial systems for speech recognition and computer vision and it is expected to have an important role in smart healthcare in the future. It is expected that wearable healthcare systems with DL approaches will be successfully integrated in mobile systems (52)(53)(54)(55)(56)(57)(58)(59)(60)(61)(62)(63)(64). The direct consequence of the application of such systems, in both home and clinical practice, will be decreasing expenses for home and clinical healthcare, or supervising athletes.…”

Section: Limitations and Future Directionsmentioning

confidence: 99%

Deep learning approaches for human activity recognition using wearable technology

Janković¹,

Savić²,

Novičić³

et al. 2018

Medicinski podmladak

View full text Add to dashboard Cite

The need for long-term monitoring of individuals in their natural environment has initiated the development of a various number of wearable healthcare sensors for a wide range of applications: medical monitoring in clinical or home environments, physical activity assessment of athletes and recreators, baby monitoring in maternity hospitals and homes etc. Neural networks (NN) are data-driven type of modelling. Neural networks learn from experience, without knowledge about the model of phenomenon, but knowing the desired "output" data for the training "input" data. The most promising concept of machine learning that involves NN is the deep learning (DL) approach. The focus of this review is on approaches of DL for physiological activity recognition or human movement analysis purposes, using wearable technologies. This review shows that deep learning techniques are useful tools for health condition prediction or overall monitoring of data, streamed by wearable systems. Despite the considerable progress and wide field of applications, there are still some limitations and room for improvement of DL approaches for wearable healthcare systems, which may lead to more robust and reliable technology for personalized healthcare. Keywords: deep learning, human activity monitoring, human machine interface, wearable sensors, smart sensors, multimodal interface Medicinski podmladak / Medical Youth Janković M. et al. Deep learning approaches for human activity recognition using wearable technology. MedPodml 2018, 69(3):14-24 15Potreba za dugotrajnim praćenjem osoba u njihovom svakodnevnom okruženju inicirala je razvoj velikog broja nosivih senzora (integrisanih u delove garderobe) s različitom primenom, kao što su medicinsko praćenje u kliničkim i kućnim uslovima, procena fizičke aktivnosti sportista i rekreativaca, praćenje beba u porodilištima i kućama i sl. Neuralne mreže (NM) predstavljaju tip modelovanja zasnovan na velikom broju podataka. Ove mreže uče na osnovu iskustva, bez poznavanja modela fenomena, ali znajući šta su željeni "izlazni" podaci za obučavajuće "ulazne" podatke. Koncept mašinskog učenja koji najviše obećava i uključuje NM jeste duboko učenje (DU). Fokus ovog preglednog rada je u pristupima DU u cilju prepoznavanja fizioloških aktivnosti i analize ljudskih pokreta primenom "odevne tehnologije". Ovaj rad pokazuje da su tehnike dubokog učenja korisne alatke za predikciju zdravstvenih stanja ili celokupno praćenje podataka koji se šalju sa "odevnih" senzora. Uprkos značajnom napretku i obećavajućoj oblasti primene, i dalje postoje ograničenja i prostor za unapređenje pristupa DU za "odevne" zdravstvene sisteme koji će dovesti do njihove pouzdane primene i omogućiti personalizovanu zdravstvenu zaštitu.Ključne reči: duboko učenje, praćenje ljudske aktivnosti, čovek-mašina interfejs, odevni senzori, pametni senzori, multimodalni interfejs

show abstract

“…The use of sensors in healthcare has been leading to significant achievements, and deep learning is being used to leverage the use of sensors and actuators for proper healthcare delivery. For instance, in assessing the level of Parkinson's disease, Eskofier et al [48] used convolutional neural networks (CNN) for the classification and detection of the key features in Parkinson's disease based on data generated from wearable sensors. The results of the research proved that deep learning techniques work well with sensors when compared to other methods.…”

Section: Deep Learning On Sensor Network Applicationsmentioning

confidence: 99%

Big Sensed Data Meets Deep Learning for Smarter Health Care in Smart Cities

Obinikpo

Kantarcı

2017

JSAN

View full text Add to dashboard Cite

Abstract:With the advent of the Internet of Things (IoT) concept and its integration with the smart city sensing, smart connected health systems have appeared as integral components of the smart city services. Hard sensing-based data acquisition through wearables or invasive probes, coupled with soft sensing-based acquisition such as crowd-sensing results in hidden patterns in the aggregated sensor data. Recent research aims to address this challenge through many hidden perceptron layers in the conventional artificial neural networks, namely by deep learning. In this article, we review deep learning techniques that can be applied to sensed data to improve prediction and decision making in smart health services. Furthermore, we present a comparison and taxonomy of these methodologies based on types of sensors and sensed data. We further provide thorough discussions on the open issues and research challenges in each category.

show abstract

Recent machine learning advancements in sensor-based mobility analysis: Deep learning for Parkinson's disease assessment

Cited by 131 publications

References 20 publications

A Multi-Layer Gaussian Process for Motor Symptom Estimation in People With Parkinson's Disease

A Multi-Layer Gaussian Process for Motor Symptom Estimation in People With Parkinson's Disease

Deep learning approaches for human activity recognition using wearable technology

Big Sensed Data Meets Deep Learning for Smarter Health Care in Smart Cities

Contact Info

Product

Resources

About