CoChat

Development of digital biomarkers for resting tremor and bradykinesia using a wrist-worn wearable device

Open access

The DOI matches its Crossref record; 57 of the 58 references it cites were found in a registry or an index.

Abstract

Objective assessment of Parkinson’s disease symptoms during daily life can help improve disease management and accelerate the development of new therapies. However, many current approaches require the use of multiple devices, or performance of prescribed motor activities, which makes them ill-suited for free-living conditions. Furthermore, there is a lack of open methods that have demonstrated both criterion and discriminative validity for continuous objective assessment of motor symptoms in this population. Hence, there is a need for systems that can reduce patient burden by using a minimal sensor setup while continuously capturing clinically meaningful measures of motor symptom severity under free-living conditions. We propose a method that sequentially processes epochs of raw sensor data from a single wrist-worn accelerometer by using heuristic and machine learning models in a hierarchical framework to provide continuous monitoring of tremor and bradykinesia. Results show that sensor derived continuous measures of resting tremor and bradykinesia achieve good to strong agreement with clinical assessment of symptom severity and are able to discriminate between treatment-related changes in motor states.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41746-019-0217-7

Cites

57 of 57 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 Flow chart illustrating the hierarchical approach for detection and assessment of resting tremor and hand bradykinesia using accelerometer data from a wrist-worn device. This approach utilizes context detection (hand movement and gait) to identify periods of interest from raw sensor data and subsequently performs detection and assessment of motor symptoms (tremor and bradykinesia). Tremor is assessed when the hand is at rest and bradykinesia is assessed during periods of hand movement that are not due to gait.
Fig. 1 Flow chart illustrating the hierarchical approach for detection and assessment of resting tremor and hand bradykinesia using accelerometer data from a wrist-worn device. This approach utilizes context detection (hand movement and gait) to identify periods of interest from raw sensor data and subsequently performs detection and assessment of motor symptoms (tremor and bradykinesia). Tremor is assessed when the hand is at rest and bradykinesia is assessed during periods of hand movement that are not due to gait.CC BY · PubMed Central
Fig. 2 Agreement between the heuristic algorithm and ML classifier for resting tremor detection. Resting tremor constancy was estimated as a fraction of the total visit duration when resting tremor was detected from sensor data. Each point corresponds to one subject visit, solid line represents line of best fit and shaded region represents the confidence interval.
Fig. 2 Agreement between the heuristic algorithm and ML classifier for resting tremor detection. Resting tremor constancy was estimated as a fraction of the total visit duration when resting tremor was detected from sensor data. Each point corresponds to one subject visit, solid line represents line of best fit and shaded region represents the confidence interval.CC BY · PubMed Central
Fig. 3 Sensor-based assessments of resting tremor and clinical ratings. Agreement between a sensor-derived measures of resting tremor constancy and MDS-UPDRS score of rest tremor constancy, and b sensor-derived measures of resting tremor amplitude and MDS-UPDRS score of rest tremor amplitude. Box plot center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers.
Fig. 3 Sensor-based assessments of resting tremor and clinical ratings. Agreement between a sensor-derived measures of resting tremor constancy and MDS-UPDRS score of rest tremor constancy, and b sensor-derived measures of resting tremor amplitude and MDS-UPDRS score of rest tremor amplitude. Box plot center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers.CC BY · PubMed Central
Fig. 4 Sensor-based assessment of hand bradykinesia and clinical rating. Agreement between predicted hand bradykinesia score and live rater’s hand bradykinesia score. Each point corresponds to one subject visit, solid line represents line of best fit and shaded region represents the confidence interval. Hand bradykinesia score (range: 0–12) is a composite score derived by taking the sum of finger tapping (MDS-UPDRS 3.4), hand movement (MDS-UPDRS 3.5) and hand pronation-supination (MDS-UPDRS 3.6) scores.
Fig. 4 Sensor-based assessment of hand bradykinesia and clinical rating. Agreement between predicted hand bradykinesia score and live rater’s hand bradykinesia score. Each point corresponds to one subject visit, solid line represents line of best fit and shaded region represents the confidence interval. Hand bradykinesia score (range: 0–12) is a composite score derived by taking the sum of finger tapping (MDS-UPDRS 3.4), hand movement (MDS-UPDRS 3.5) and hand pronation-supination (MDS-UPDRS 3.6) scores.CC BY · PubMed Central
Fig. 5 Motor state associated changes in sensor-based assessment of resting tremor and hand bradykinesia. Distribution of difference between ON and OFF states of sensor-derived measures of a tremor constancy, b tremor amplitude, and c predicted hand bradykinesia score grouped by the randomized order of motor state in visit 1. Box plot center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers; **p < 0.01, ***p < 0.001.
Fig. 5 Motor state associated changes in sensor-based assessment of resting tremor and hand bradykinesia. Distribution of difference between ON and OFF states of sensor-derived measures of a tremor constancy, b tremor amplitude, and c predicted hand bradykinesia score grouped by the randomized order of motor state in visit 1. Box plot center line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers; **p < 0.01, ***p < 0.001.CC BY · PubMed Central
Fig. 6 Willingness of subjects to wear rigid (Opal) or flexible patch-like (BioStamp) devices continuously at home. A marked preference for a flexible patch-like device was observed for subjects who were very likely to wear devices continuously at home.
Fig. 6 Willingness of subjects to wear rigid (Opal) or flexible patch-like (BioStamp) devices continuously at home. A marked preference for a flexible patch-like device was observed for subjects who were very likely to wear devices continuously at home.CC BY · PubMed Central
Fig. 7 An example showing continuous assessment of resting tremor and bradykinesia using accelerometer data from a device worn on the most affected wrist. Data was collected on one PD patient across two in-lab visits. The first visit was in the OFF state and second visit in the ON state. a Continuous tremor detection and tremor amplitude measurement in the OFF (left panel) and ON (right panel) state. Inset shows aggregate measures of tremor constancy and tremor amplitude. Tremor constancy was rated 3 in the OFF state and 1 in the ON state. Tremor amplitude was rated 2 in the OFF and 0 in the ON. b Continuous hand movement detection and hand amplitude measurement in the OFF (left panel) and ON (right panel) state. Inset shows aggregate measures of hand movement amplitude and length of no hand movement bouts. Hand bradykinesia was rated 9 in the OFF state and 2 in the ON state.
Fig. 7 An example showing continuous assessment of resting tremor and bradykinesia using accelerometer data from a device worn on the most affected wrist. Data was collected on one PD patient across two in-lab visits. The first visit was in the OFF state and second visit in the ON state. a Continuous tremor detection and tremor amplitude measurement in the OFF (left panel) and ON (right panel) state. Inset shows aggregate measures of tremor constancy and tremor amplitude. Tremor constancy was rated 3 in the OFF state and 1 in the ON state. Tremor amplitude was rated 2 in the OFF and 0 in the ON. b Continuous hand movement detection and hand amplitude measurement in the OFF (left panel) and ON (right panel) state. Inset shows aggregate measures of hand movement amplitude and length of no hand movement bouts. Hand bradykinesia was rated 9 in the OFF state and 2 in the ON state.CC BY · PubMed Central
Fig. 8 Sensor setup and experimental protocol. a An illustration of the sensor setup used for data collection. Sensor data from only one device (blue) attached to the wrist of the most affected side (PD) or dominant side (HC) was used for the analysis presented in this paper. b Both the PD study and the HC study consisted of two in-clinic visits. Subjects performed a series of activities that were repeated in each visit. Activities were categorized into controlled speech activities, scripted MDS-UPDRS tasks, and activities of daily living. For HC’s both visits were identical whereas PD patients were randomly assigned to be in their self-reported ON or OFF motor state for a given visit.
Fig. 8 Sensor setup and experimental protocol. a An illustration of the sensor setup used for data collection. Sensor data from only one device (blue) attached to the wrist of the most affected side (PD) or dominant side (HC) was used for the analysis presented in this paper. b Both the PD study and the HC study consisted of two in-clinic visits. Subjects performed a series of activities that were repeated in each visit. Activities were categorized into controlled speech activities, scripted MDS-UPDRS tasks, and activities of daily living. For HC’s both visits were identical whereas PD patients were randomly assigned to be in their self-reported ON or OFF motor state for a given visit.CC BY · PubMed Central
Fig. 9 Method for detecting hand movement from raw accelerometer data. Vector magnitude derived from raw triaxial accelerometer data is filtered to remove high frequency components. An empirically derived threshold is then applied to coefficient of variation of the filtered signal to detect hand movement. HM, hand movement; NHM, no hand movement.
Fig. 9 Method for detecting hand movement from raw accelerometer data. Vector magnitude derived from raw triaxial accelerometer data is filtered to remove high frequency components. An empirically derived threshold is then applied to coefficient of variation of the filtered signal to detect hand movement. HM, hand movement; NHM, no hand movement.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 57 references

Research works printed but not linked: 3

  1. Standaert, D. G. & Roberson, E. D. Treatment of Central Nervous System Degenerative Disorders. in Goodman & Gilman’s: The Pharmacological Basis of Therapeutics, 12e (eds Brunton, L. L., Chabner, B. A. & Knollmann, B. C.) (McGraw-Hill Education, 2015).
  2. Erb, K. et al. The BlueSky Project: monitoring motor and non-motor characteristics of people with Parkinson’s disease in the laboratory, a simulated apartment, and home and community settings. Mov. Disord. 33, 1990–1990 (2018).
  3. Guyon, I. & Elisseeff, A. An introduction to variable and feature selection. J. Mach. Learn. Res. 3, 1157–1182 (2003).

Cited by

See all 158

Metadata

Authors
Mahadevan, N., Demanuele, C., Zhang, H., Volfson, D., Ho, B., Erb, M. K., & Patel, S.
Affiliations
  • Pfizer (United States) (US)
  • Tufts Medical Center (US)
Corresponding author
Patel, Shyamal
Venue
npj Digital Medicine
Published
2020-01-15
Volume
3
Issue
1
Pages
5-5
ISSN
2398-6352 (ISSN-L)
OpenAlex
W2999749474
Cited by
138
References
58
Citations by year
2026: 13 · 2025: 25 · 2024: 24 · 2023: 19 · 2022: 25 · 2021: 24 · 2020: 8
FWCI
7.39
Type
journal article
Language
en
Open access
Licence
Locations
  • npj Digital Medicine (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Pfizer
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
  • Health Sciences › Medicine › Clinical Neurology › Parkinson's Disease Mechanisms and Treatments
  • Physical Sciences › Engineering › Biomedical Engineering › Non-Invasive Vital Sign Monitoring
Keywords
Parkinson's disease, resting tremor, bradykinesia, digital biomarkers, wrist-worn accelerometers, wearable sensors, free-living conditions

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Development of digital biomarkers for resting tremor and bradykinesia using a wrist-worn wearable device. https://app.cochat.ai/papers/10.1038/s41746-019-0217-7
The paper
Mahadevan, N., Demanuele, C., Zhang, H., Volfson, D., Ho, B., Erb, M. K., & Patel, S. (2020). Development of digital biomarkers for resting tremor and bradykinesia using a wrist-worn wearable device. npj Digital Medicine. https://doi.org/10.1038/s41746-019-0217-7

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.