Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 29 of the 29 references it cites were found in a registry or an index.

Increasing Robustness in the Detection of Freezing of Gait in Parkinson’s Disease

Open access

Abstract

This paper focuses on detecting freezing of gait in Parkinson’s patients using body-worn accelerometers. In this study, we analyzed the robustness of four feature sets, two of which are new features adapted from speech processing: mel frequency cepstral coefficients and quality assessment metrics. For classification based on these features, we compared random forest, multilayer perceptron, hidden Markov models, and deep neural networks. These algorithms were evaluated using a leave-one-subject-out (LOSO) cross validation to match the situation where a system is being constructed for patients for whom there is no training data. This evaluation was performed using the Daphnet dataset, which includes recordings from ten patients using three accelerometers situated on the ankle, knee, and lower back. We obtained a reduction from 17.3% to 12.5% of the equal error rate compared to the previous best results using this dataset and LOSO testing. For high levels of sensitivity (such as 0.95), the specificity increased from 0.63 to 0.75. The biggest improvement across all of the feature sets and algorithms tested in this study was obtained by integrating information from longer periods of time in a deep neural network with convolutional layers.

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 · 10.3390/electronics8020119

Cites

29 of 29 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Spectrogram for three accelerometer signals: forward axis of ankle, knee, and back sensors. The log energy at different frequencies is coded with blue for low energy and red for higher energy, ranging from 0 to 80 dB. During a FOG episode, the power increases in the frequency band of 3-8 Hz.
Figure 1 Spectrogram for three accelerometer signals: forward axis of ankle, knee, and back sensors. The log energy at different frequencies is coded with blue for low energy and red for higher energy, ranging from 0 to 80 dB. During a FOG episode, the power increases in the frequency band of 3-8 Hz.CC-BY · From the PDF
Figure 2 Deep learning structure including convolutional and fully connected layers.
Figure 2 Deep learning structure including convolutional and fully connected layers.CC-BY · From the PDF
Figure 3 Example of random assignment to train or test in a sequence of 50%-overlapped windows.
Figure 3 Example of random assignment to train or test in a sequence of 50%-overlapped windows.CC-BY · From the PDF
Image 4
CC-BY · From the PDF
Figure 5 Results for Mazilu et al.'s features [22] (curves) for each subject compared to [13] when using optimized thresholds and R10Fold evaluation.
Figure 5 Results for Mazilu et al.'s features [22] (curves) for each subject compared to [13] when using optimized thresholds and R10Fold evaluation.CC-BY · From the PDF
Figure 6 Specificity vs. sensitivity curves, AUC, and EER for the four sets of features using random forest with 4-s windows and 75% overlap, a leave-one-subject-out (LOSO) evaluation, and excluding subjects 4 and 10. The isolated point shows the results reported in Mazilu et al. [22] without subjects 4 and 10 (sensitivity = 0.55 and specificity = 0.95).
Figure 6 Specificity vs. sensitivity curves, AUC, and EER for the four sets of features using random forest with 4-s windows and 75% overlap, a leave-one-subject-out (LOSO) evaluation, and excluding subjects 4 and 10. The isolated point shows the results reported in Mazilu et al. [22] without subjects 4 and 10 (sensitivity = 0.55 and specificity = 0.95).CC-BY · From the PDF
Figure 7 Results per subject for MFCCs and Mazilu et al.'s features using a LOSO evaluation and excluding subjects 4 and 10.
Figure 7 Results per subject for MFCCs and Mazilu et al.'s features using a LOSO evaluation and excluding subjects 4 and 10.CC-BY · From the PDF
Figure 8 Specificity vs. sensitivity curves, AUC, and EER for the four classification algorithms with LOSO evaluation and excluding subjects 4 and 10. The two isolated points represent the performance using hidden Markov models with different probability transition between FOG and non-FOG models.
Figure 8 Specificity vs. sensitivity curves, AUC, and EER for the four classification algorithms with LOSO evaluation and excluding subjects 4 and 10. The two isolated points represent the performance using hidden Markov models with different probability transition between FOG and non-FOG models.CC-BY · From the PDF
Figure 9 Structure of the hidden Markov models.
Figure 9 Structure of the hidden Markov models.CC-BY · From the PDF
Figure 10 Specificity vs. sensitivity curves, AUC, and EER when including previous and posterior windows in the deep neural network, using a LOSO evaluation and excluding subjects 4 and 10.
Figure 10 Specificity vs. sensitivity curves, AUC, and EER when including previous and posterior windows in the deep neural network, using a LOSO evaluation and excluding subjects 4 and 10.CC-BY · From the PDF
Figure 11 Specificity vs. sensitivity curves, AUC, EER, and specificity (Spe) for a sensitivity of 0.95 when comparing the baseline and the best system in this paper with a LOSO evaluation and excluding subjects 4 and 10.
Figure 11 Specificity vs. sensitivity curves, AUC, EER, and specificity (Spe) for a sensitivity of 0.95 when comparing the baseline and the best system in this paper with a LOSO evaluation and excluding subjects 4 and 10.CC-BY · From the PDF

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 29 references

Cited by

See all 87

Metadata

Authors
San-Segundo, R., Navarro-Hellín, H., Torres-Sánchez, R., Hodgins, J., & De la Torre, F.
Affiliations
  • Information Processing and Telecommunications Center (ES)
  • Carnegie Mellon University (US)
  • Universidad Politécnica de Cartagena (ES)
  • Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA
Corresponding author
San-Segundo, Rubén
Venue
Electronics
Publisher
MDPI AG
Published
2019-01-22
Volume
8
Issue
2
Pages
119
ISSN
2079-9292 (ISSN-L)
OpenAlex
W2913141042
Cited by
86
References
29
Citations by year
2026: 3 · 2025: 12 · 2024: 17 · 2023: 11 · 2022: 14 · 2021: 12 · 2020: 14 · 2019: 3
FWCI
24.76
Type
journal article
Language
en
Open access
Locations
  • Electronics (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • National Science Foundation (1602337)
  • Ministerio de Educación, Cultura y Deporte and the Fulbright Commission in Spain (PRX16/00103)
Topics
  • Health Sciences › Health Professions › Physical Therapy, Sports Therapy and Rehabilitation › Balance, Gait, and Falls Prevention
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
  • Physical Sciences › Engineering › Biomedical Engineering › Gait Recognition and Analysis
Keywords
freezing of gait, Parkinson's disease, wearable accelerometers, gait analysis, feature robustness, Mel-frequency cepstral coefficients, leave-one-subject-out cross-validation, deep neural network, convolutional neural networks, hidden Markov models, random forest, multilayer perceptron

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: Increasing Robustness in the Detection of Freezing of Gait in Parkinson’s Disease. https://app.cochat.ai/papers/10.3390/electronics8020119
The paper
San-Segundo, R., Navarro-Hellín, H., Torres-Sánchez, R., Hodgins, J., & De la Torre, F. (2019). Increasing Robustness in the Detection of Freezing of Gait in Parkinson’s Disease. Electronics. https://doi.org/10.3390/electronics8020119

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.