PD Disease State Assessment in Naturalistic Environments Using Deep Learning

By Nils Y. Hammerla, J. Fisher, Péter András, L. Rochester, R. Walker, Thomas Plötz. 2015

193 citations recorded by the source corpus.

Management of Parkinson's Disease (PD) could be improved significantly if reliable, objective information about fluctuations in disease severity can be obtained in ecologically valid surroundings such as the private home. Although automatic assessment in PD has been studied extensively, so far no approach has been devised that is useful for clinical practice. Analysis approaches common for the field lack the capability of exploiting data from realistic environments, which represents a major barrier towards practical assessment systems. The very unreliable and infrequent labelling of ambiguous, low resolution movement data collected in such environments represents a very challenging analysis setting, where advances would have significant societal impact in our ageing population. In this work we propose an assessment system that abides practical usability constraints and applies deep learning to differentiate disease state in data collected in naturalistic settings. Based on a large data-set collected from 34 people with PD we illustrate that deep learning outperforms other approaches in generalisation performance, despite the unreliable labelling characteristic for this problem setting, and how such systems could improve current clinical practice.

Subfields: Neurology, Physical Therapy, Sports Therapy and Rehabilitation

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Neurology

1. Zachary Chase Lipton

Citations in the corpus

• Distilling Knowledge from Deep Networks with Applications to Healthcare Domain (2015, background): “Other recent works [18, 30] also leverage the power of deep learning approaches to model diseases and clinical time series data.”
• Learning to Diagnose with LSTM Recurrent Neural Networks (2015, background): “Multiple recent papers apply modern deep learning techniques (but not RNNs) to modeling psychological conditions (Dabek & Caban, 2015), head injuries (Rughani et al., 2010), and Parkinson’s disease (Hammerla et al., 2015).”
• Doctor AI: Predicting Clinical Events via Recurrent Neural Networks (2015, background): “Previous studies such as phenotype learning (Lasko et al., 2013; Che et al., 2015; Hammerla et al., 2015) or representation learning (Choi et al.”
• Benchmark of Deep Learning Models on Large Healthcare MIMIC Datasets (2017, background): “Other recent works [38, 39, 40] have leveraged the power of deep learning approaches to model diseases and clinical time series data.”
• An Early Resource Characterization of Deep Learning on Wearables, Smartphones and Internet-of-Things Devices (2015, background): “Enabling wide-spread device-side deep learning inference will require a range of brand-new techniques for optimized resource sensitive execution.”
• Deep Learning for Sensor-based Human Activity Recognition (2020, methodology): “1:17 Table 2. Deep Learning Works for Annotation Scarcity in Activity Recognition Training Fashion Learning Approach References Unsupervised Pretraining Alsheikh et al. 2016 [6]; Hammerla et al. 2015 [56]; Plötz et al. 2011 [111]; Chikhaoui and Gouineau [33]; Gu et al. 2018 [47]; Zeng et al. 2017 [1”
• Ensembles of Deep LSTM Learners for Activity Recognition using Wearables (2017, methodology): “Based on this, subsequent works explored the e↵ectiveness of pre-trained, fully-connected RBM networks, for example, for automated assessments of disease states in Parkinson’s patients [Hammerla et al. 2015], in combination with more traditional sequence models (HMMs) [Zhang et al.”
• Data augmentation of wearable sensor data for parkinson’s disease monitoring using convolutional neural networks (2017, methodology): “In [11], a restricted Boltzmann machine (RBM) was used as pretraining for classifying four PD motor states.”
• Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances (2021, background): “In [118], DBNs are employed in Parkinson’s disease diagnosis to explore if they can cope with the unreliable labelling that results from naturalistic recording environments.”
• Using recurrent neural network models for early detection of heart failure onset (2016, background): “Both of these studies 30,31 used multivariate time series data from patients, which focused on very different clinical conditions, with continuous time series data.”
• DeepEar: robust smartphone audio sensing in unconstrained acoustic environments using deep learning (2015, background): “Deep learning has the potential to broadly impact the fields of activity recognition and the modeling of user behavior and context; early explorations of this potential are already underway [47, 35].”
• A Semisupervised Recurrent Convolutional Attention Model for Human Activity Recognition (2020, methodology): “[14], [15] resort to deep generative models such as Restricted Boltzmann Machines (RBMs) and”
• Deep Learning for Sensor-based Activity Recognition: A Survey (2017, methodology): “For multi-modal signals, it is better to use CNN since (Bhattacharya and Lane, 2016) Body-worn, ambiemt RBM Gesture, ADL, transportation Self, D01 (Chen and Xue, 2015) Body-worn CNN ADL Self (Chen et al., 2016b) Body-worn CNN ADL D06 (Edel and K¨oppe, 2016) Body-worn RNN ADL D01, D04, Self (Fang and”
• Deep, Convolutional, and Recurrent Models for Human Activity Recognition Using Wearables (2016, background): “However, more elaborate behaviours which are, for example, of interest in medical applications, pose a significant challenge to this manually tuned approach [Hammerla et al., 2015].”