Abstract:Continuous, automated monitoring of Parkinsons Disease (PD) symptoms would provide clinicians with more information to understand their patients' disease progression and adjust treatment protocols, thereby improving PD care. Collecting precisely labeled data for Parkinson's symptoms, such as tremor, is difficult. Therefore, algorithms for monitoring should only require weakly-labeled training data. In this paper, we evaluate five standard weakly-supervised algorithms and propose a "stratified" version of three… Show more
“…The more recent work of [24], extended the previous approach by using a dataset of 5 PD patients all of which exhibited tremor. Data collection took place under laboratory conditions that resembled a home environment.…”
Parkinson's Disease (PD) is a slowly evolving neurological disease that affects about 1% of the population above 60 years old, causing symptoms that are subtle at first, but whose intensity increases as the disease progresses. Automated detection of these symptoms could offer clues as to the early onset of the disease, thus improving the expected clinical outcomes of the patients via appropriately targeted interventions. This potential has led many researchers to develop methods that use widely available sensors to measure and quantify the presence of PD symptoms such as tremor, rigidity and braykinesia. However, most of these approaches operate under controlled settings, such as in lab or at home, thus limiting their applicability under free-living conditions. In this work, we present a method for automatically identifying tremorous episodes related to PD, based on IMU signals captured via a smartphone device. We propose a Multiple-Instance Learning approach, wherein a subject is represented as an unordered bag of accelerometer signal segments and a single, expert-provided, tremor annotation. Our method combines deep feature learning with a learnable pooling stage that is able to identify key instances within the subject bag, while still being trainable end-to-end. We validate our algorithm on a newly introduced dataset of 45 subjects, containing accelerometer signals collected entirely in-the-wild. The good classification performance obtained in the conducted experiments suggests that the proposed method can efficiently navigate the noisy environment of in-the-wild recordings.
“…The more recent work of [24], extended the previous approach by using a dataset of 5 PD patients all of which exhibited tremor. Data collection took place under laboratory conditions that resembled a home environment.…”
Parkinson's Disease (PD) is a slowly evolving neurological disease that affects about 1% of the population above 60 years old, causing symptoms that are subtle at first, but whose intensity increases as the disease progresses. Automated detection of these symptoms could offer clues as to the early onset of the disease, thus improving the expected clinical outcomes of the patients via appropriately targeted interventions. This potential has led many researchers to develop methods that use widely available sensors to measure and quantify the presence of PD symptoms such as tremor, rigidity and braykinesia. However, most of these approaches operate under controlled settings, such as in lab or at home, thus limiting their applicability under free-living conditions. In this work, we present a method for automatically identifying tremorous episodes related to PD, based on IMU signals captured via a smartphone device. We propose a Multiple-Instance Learning approach, wherein a subject is represented as an unordered bag of accelerometer signal segments and a single, expert-provided, tremor annotation. Our method combines deep feature learning with a learnable pooling stage that is able to identify key instances within the subject bag, while still being trainable end-to-end. We validate our algorithm on a newly introduced dataset of 45 subjects, containing accelerometer signals collected entirely in-the-wild. The good classification performance obtained in the conducted experiments suggests that the proposed method can efficiently navigate the noisy environment of in-the-wild recordings.
“…At best, weak labels (presence, absence, or the approximate amount of a symptom within a time segment) can be provided by participants of the data collections. Weakly supervised algorithms, such as multiple-instance learning, explicitly account for weak labels, and tend to perform better than standard, fully supervised learning algorithms in these scenarios [ 4 ]. In summary, collecting data, labeling them, and learning from such data are all more challenging in wild environments.…”
Section: Introductionmentioning
confidence: 99%
“…Each feature set is followed with either a Random Forest (RF) or Multi-Layer Perceptron (MLP) classifier to distinguish the various performance contributions of features versus algorithms. We compare the performance of our best feature set and algorithm combination with the three most relevant systems described in the literature: [ 4 , 6 , 7 ]. We also compare our method with [ 4 ] (best among [ 6 , 7 ]) in the ability to reproduce patient self-reports, the current standard for in-home symptom monitoring.…”
Section: Introductionmentioning
confidence: 99%
“…We compare the performance of our best feature set and algorithm combination with the three most relevant systems described in the literature: [ 4 , 6 , 7 ]. We also compare our method with [ 4 ] (best among [ 6 , 7 ]) in the ability to reproduce patient self-reports, the current standard for in-home symptom monitoring. Finally, we propose a new metric for monitoring PD tremor (percentage of tremor time measured over longer periods), which is easier to detect than specific tremor events.…”
Continuous in-home monitoring of Parkinson’s Disease (PD) symptoms might allow improvements in assessment of disease progression and treatment effects. As a first step towards this goal, we evaluate the feasibility of a wrist-worn wearable accelerometer system to detect PD tremor in the wild (uncontrolled scenarios). We evaluate the performance of several feature sets and classification algorithms for robust PD tremor detection in laboratory and wild settings. We report results for both laboratory data with accurate labels and wild data with weak labels. The best performance was obtained using a combination of a pre-processing module to extract information from the tremor spectrum (based on non-negative factorization) and a deep neural network for learning relevant features and detecting tremor segments. We show how the proposed method is able to predict patient self-report measures, and we propose a new metric for monitoring PD tremor (i.e., percentage of tremor over long periods of time), which may be easier to estimate the start and end time points of each tremor event while still providing clinically useful information.
“…Radar based approaches used predefined features from the time-frequency representation (TFR) of the radar return signal to monitor human gait [2]. Sensor based approaches detect gait abnormality or tremors using contact sensors such as MetaMotionRan Inertial Measurement Unit (IMU) sensor from Mbientlab [3] and the Axivity AX3 accelerometer [4]. Camera based approaches recognizes gait abnormality or tremors using the Kinect camera [5], [6].…”
Cerebellar dysfunction (CD) is a neurological disorder that involves a number of abnormalities that affect the movement of various parts of the body such as gait abnormality or tremors in limbs such as hands or feet while reaching out for something. A user-friend tool that can objectively evaluate the aforementioned body movements in CD patients can aid the clinicians for an objective assessment in clinical settings. The objective of this work is to develop a method that quantifies the gait abnormality and tremors in hand using S-Band sensing technique. The S-Band sensing essentially leverages small wireless devices such as network interface card, omnidirectional antenna and router operating at 2.4 GHz to record the wireless channel data. Specifically, the aim is to use the variances of amplitude and phase information induced due to the human body movements. Each body movement leaves a unique imprint in the form of wireless channel information that is used to identify abnormalities in body motions. The proposed framework applied a linear transformation on raw phase data for calibrations since the data retrieved using interface card contain noise and is inapplicable for motion detection. The support vector machine used to classify the data achieved high classification accuracy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.