pith. machine review for the scientific record. sign in

arxiv: 2604.22907 · v1 · submitted 2026-04-24 · ⚛️ physics.med-ph · stat.ME

Recognition: unknown

Fingertip Micro-Motion as a Source of Respiratory Information During Sleep Using Triaxial Accelerometers

Jeanne Lin , Lily Liu , Hau-Tieng Wu
Authors on Pith no claims yet

Pith reviewed 2026-05-08 08:49 UTC · model grok-4.3

classification ⚛️ physics.med-ph stat.ME
keywords triaxial accelerometerfingertiprespiratory monitoringsleepinstantaneous respiratory raterespiratory effortmicro-motionpolysomnography
0
0 comments X

The pith

Fingertip triaxial accelerometers capture respiratory effort and rate during sleep.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper shows that accelerometer signals at the fingertip contain usable information about breathing movements while sleeping. The authors apply an antiderivative-based transformation to turn raw triaxial data into a respiratory surrogate signal. Validation against full polysomnography recordings finds that roughly one-fifth of the night yields high-quality segments on average, where the surrogate tracks chest and abdominal motion better than airflow. These segments support accurate estimates of instantaneous respiratory rate and a motion index that flags reliable data.

Core claim

Fingertip-mounted triaxial accelerometers encode meaningful respiratory information during sleep. An antiderivative-based nonlinear transformation produces a respiratory surrogate called TAA-resp that correlates more strongly with thoracic and abdominal motion than with airflow, indicating predominant capture of respiratory effort. High-quality segments of TAA-resp yield instantaneous respiratory rate estimates with root mean square error 0.027 ± 0.022 Hz, and a respiratory motion index derived from time-frequency analysis identifies reliable data with 0.74 sensitivity and 0.75 specificity in cross-validation.

What carries the argument

The antiderivative-based nonlinear transformation that converts fingertip triaxial accelerometer signals into the TAA-resp respiratory surrogate, isolating micro-motion tied to breathing.

If this is right

  • High-quality TAA-resp segments provide instantaneous respiratory rate estimates accurate to within 0.027 ± 0.022 Hz root mean square error.
  • TAA-resp correlates more strongly with thoracic and abdominal motion than with airflow, indicating it mainly tracks respiratory effort.
  • The respiratory motion index rises during REM, N2, and N3 sleep and falls during apnea or hypopnea events.
  • The respiratory motion index predicts signal quality labels at 0.74 sensitivity and 0.75 specificity in leave-one-subject-out cross-validation.
  • On average 22.2% ± 15.6% of full-night recordings contain high-quality respiratory information from the fingertip sensor.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This method could enable wearable devices for home sleep monitoring that avoid chest straps or nasal cannulas in channel-limited settings.
  • The approach might be tested for daytime respiratory tracking or in populations with frequent movement artifacts.
  • Consumer devices such as smart rings or watches could incorporate similar fingertip or wrist placements for continuous overnight breathing surveillance.
  • The transformation and index may apply to accelerometer data from other body sites during sleep.

Load-bearing premise

The antiderivative-based nonlinear transformation successfully isolates respiratory-induced micro-motion from other fingertip movements and artifacts during sleep without introducing significant distortion or bias.

What would settle it

A new dataset of simultaneous fingertip TAA and PSG recordings in which high-quality TAA-resp segments produce instantaneous respiratory rate estimates with root mean square error well above 0.027 Hz or fail to match thoracic effort waveforms.

Figures

Figures reproduced from arXiv: 2604.22907 by Hau-Tieng Wu, Jeanne Lin, Lily Liu.

Figure 1
Figure 1. Figure 1: Illustration of s(t)g(t), where s(t) = cos(2πt) + 0.5 cos(2π2t + 1) and different smooth windows g’s (left column), and the corresponding Fourier magnitude (right column). From top to bottom, g is compactly supported on [−L, L] with L = 5, 2.5, 1.5, 0.5. In each case, g(t) = 1 on [−L + 0.2, L − 0.2] and tapers monotonically to 0 near the boundaries. Definition 2.3. Assume g is smooth, symmetric at 0, and c… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the proposed algorithm illustrated as a flowchart, with corresponding signal visualizations shown on the right. (a) Raw TAA-resp signal; (b) bandpass-filtered TAA-resp (0.1-1 Hz); (c) time-frequency representation (TFR) of (b) ob￾tained via the synchrosqueezing transform; (d) recovered phase of (b); and (e) unwrapped version of (b). The x-axis is time with unit second. See Figures 9 and 10 for … view at source ↗
Figure 3
Figure 3. Figure 3: The overall label rule and criteria. a more reliable and widely accepted framework would likely require either large￾scale studies involving multiple expert annotators or, more in line with established clinical practice, a process combining expert consensus, systematic literature re￾view, and iterative revision, as exemplified by the established apnea event scoring guidelines of the American Academy of Sle… view at source ↗
Figure 4
Figure 4. Figure 4: From left to right column: illustration of typical TAA￾resp signals labeled as good (RMI=0.88), moderate (RMI=0.68), and poor (RMI=0.31). The top row shows the raw acceleration signals after removing the mean (unit: m/s2 ), where ax is shifted above by 2×SD(ax) and az is shifted down by 2×SD(ax) to enhance the visualization, the second row shows the integrated acceleration signals after removing the mean (… view at source ↗
Figure 5
Figure 5. Figure 5: Illustration of the effect of unwrapping and the corre￾sponding spectrum. The top panel presents a TAA-resp segment labeled as high quality (RMI = 0.83), the middle panel presents a segment labeled as moderate quality (RMI = 0.64), whereas the bottom panel presents a segment labeled as poor quality (RMI = 0.44). The simultaneously recorded THOs are superimposed as gray curves. ratio 0 0.1 0.2 0.3 0.4 0.5 0… view at source ↗
Figure 6
Figure 6. Figure 6: Ratio of TAA-resp segments labeled as high-quality (moderate or good) over 39 subjects view at source ↗
Figure 7
Figure 7. Figure 7: Distributions of correlations between TAA-resp and THO, ABD, and airflow over all segments (top) and those labeled as high-quality (moderate or good). The mean and median of each group are marked as red and black vertical lines. Note that the unit of TAA-resp is arbitrary. We define the magnitude of a signal over a 1-minute segment as the 99% quantile of its absolute value. The relationships between TAA-re… view at source ↗
Figure 8
Figure 8. Figure 8: Two-dimensional histograms of log(1+magnitude of TAA-resp) versus log(1+magnitude of THO) (left), log(1+magnitude of ABD) (middle), and log(1+magnitude of airflow) (right), shown for all segments (top) and high-quality segments (bottom). 0.02 Hz. See view at source ↗
Figure 9
Figure 9. Figure 9: TAA-resp and THO are shown in the top left panel, and the associated TFRs are shown in the middle and bottom row. The THO is scaled to have the same magnitude of TAA-resp to enhance the visualization. The estimated IRRs are shown in the top right panel and superimposed in each TFR as red curves. The RMSE is 0.02. the odds ratio 2.64 (95% CI=[1.92, 3.63]) and p < 10−10. This result suggests that high-qualit… view at source ↗
Figure 10
Figure 10. Figure 10: TAA-resp and THO are shown in the top left panel, and the associated TFRs are shown in the middle and bottom row. The THO is scaled to have the same magnitude of TAA-resp to enhance the visualization. The estimated IRRs are shown in the top right panel and superimposed in each TFR. The RMSE is 0.015. Bad Moderate Good R M I 0 0.2 0.4 0.6 0.8 1 view at source ↗
Figure 11
Figure 11. Figure 11: Distribution of RMI over three groups. The mean and median of each group are marked as red and black vertical lines. 5.8. RMI distribution over different sleep stages view at source ↗
Figure 12
Figure 12. Figure 12: Violin plots of RMI over different sleep stages. The mean and median of each group are marked as red and black ver￾tical lines. 5.9. RMI distribution over different apnea events. Next, we study the re￾lationship between labels, RMI, and apnea events. Among the 11,502 1-minute segments, 9,312 had neither hypopnea nor apnea, 655 had hypopnea, and 1,535 had apnea. In segments without hypopnea or apnea, 7,168… view at source ↗
Figure 13
Figure 13. Figure 13: RMI distributions for segments with and without sleep apnea. The mean and median of each group are marked as red and black vertical lines. specificity of 0.74±0.27 and 0.75±0.22, with overall accuracy 0.78±0.20 and F1 0.55±0.24 if we replace NaN quantities associated with either row or column of the confusion matrix containing only zeros by 0. If we remove 3 cases with any row or column of the confusion m… view at source ↗
read the original abstract

Objective: Triaxial accelerometers (TAAs) are widely used in homecare medicine. This study investigates whether TAA signals recorded at the fingertip encode respiratory information, particularly instantaneous respiratory rate (IRR) and respiratory effort, during sleep. Method: We propose an antiderivative-based nonlinear transformation to convert TAA signals into a respiratory surrogate, termed TAA-resp. To quantify the embedded respiratory-induced motion, a modern time-frequency analysis tool is applied to derive an index, referred to as the respiratory motion index (RMI). The proposed TAA-resp and RMI are validated on a dataset comprising 39 full-night recordings with simultaneous polysomnography (PSG) and a fingertip TAA measurements. Criteria for labeling TAA-resp signal quality as good, moderate, or poor are established, and expert annotations are obtained. Result: On average, TAA-resp over 22.2% $\pm$ 15.6% of full-night recordings encodes high-quality respiratory information, reaching up to 58.9% in some cases. TAA-resp shows stronger correlation with thoracic and abdominal motion than with airflow, indicating predominant capture of respiratory effort. High-quality TAA-resp offers an accurate IRR estimate with root mean square error $0.027 \pm 0.022$ Hz. RMI is higher for high-quality segments and lower for poor-quality segments, and its distribution aligns with physiology, with higher values during REM, N2, and N3 sleep and in the absence of apnea or hypopnea events. In leave-one-subject-out cross-validation, RMI predicts quality labels with 0.74 sensitivity and 0.75 specificity. Conclusion: Fingertip-mounted TAAs encode meaningful respiratory information. Leveraging this underutilized signal may enhance home-based sleep monitoring in channel-limited settings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 3 minor

Summary. The paper claims that fingertip-mounted triaxial accelerometers (TAAs) encode respiratory information during sleep. It proposes an antiderivative-based nonlinear transformation to derive a respiratory surrogate (TAA-resp) and a respiratory motion index (RMI) from time-frequency analysis of the transformed signal. These are validated on 39 full-night simultaneous PSG and fingertip TAA recordings, with expert quality labeling of TAA-resp segments. Results indicate that high-quality TAA-resp (present in 22.2% of recordings on average) correlates more strongly with thoracic/abdominal effort than airflow, yields IRR estimates with RMSE 0.027 Hz, and that RMI predicts quality labels in leave-one-subject-out cross-validation while aligning with sleep-stage physiology.

Significance. If the transformation reliably extracts respiratory micro-motion, the work could enable low-channel home sleep monitoring by repurposing fingertip accelerometers already common in wearables. Strengths include the use of full-night PSG ground truth for validation, reporting of both effort and rate metrics, and cross-validation of the RMI quality predictor. The observation that TAA-resp tracks effort better than airflow is physiologically grounded and potentially useful for distinguishing central vs. obstructive events in future extensions.

major comments (3)
  1. [Method] Method section: The antiderivative-based nonlinear transformation used to obtain TAA-resp is described only at a high level. No explicit mathematical definition, derivation, choice of integration limits or filtering parameters, or sensitivity analysis is provided. This is load-bearing for the central claim that the transformation isolates respiratory-induced micro-motion without distortion from other fingertip movements or artifacts.
  2. [Results] Results section: All quantitative performance metrics (correlations with PSG channels, IRR RMSE of 0.027 ± 0.022 Hz) are reported exclusively on the expert-labeled 'high-quality' subset (average 22.2% of recording time). No breakdown of moderate/poor segments, failure-rate statistics, or characteristics of discarded data is given, raising the possibility that reported accuracy reflects selection of favorable segments rather than robust performance of the transformation.
  3. [Results] Results section: No controlled experiments, simulations, or ablation studies are described to test whether the transformation suppresses or distorts non-respiratory components (e.g., cardiac ballistocardiography, voluntary micro-movements, or sensor drift) that are known to be present in fingertip TAA signals. Such tests would be required to substantiate that TAA-resp primarily reflects respiratory effort.
minor comments (3)
  1. [Abstract] Abstract and Methods: The 'modern time-frequency analysis tool' used to compute RMI is not named (e.g., whether it is a specific synchrosqueezed transform or wavelet method), making it difficult to reproduce the index.
  2. [Results] Results: The claim that RMI distributions 'align with physiology' is supported only by qualitative description; quantitative tables or statistical tests comparing RMI across sleep stages and apnea/hypopnea events would improve clarity.
  3. [Figures] Figure legends: Legends for any time-series or spectrogram figures should explicitly state the time scale, frequency range, and which PSG channels are overlaid for visual comparison.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed review. The comments highlight important areas for improving clarity, completeness, and robustness. We address each major comment below and will revise the manuscript to incorporate the suggested enhancements.

read point-by-point responses

  1. Referee: [Method] Method section: The antiderivative-based nonlinear transformation used to obtain TAA-resp is described only at a high level. No explicit mathematical definition, derivation, choice of integration limits or filtering parameters, or sensitivity analysis is provided. This is load-bearing for the central claim that the transformation isolates respiratory-induced micro-motion without distortion from other fingertip movements or artifacts.

    Authors: We agree that the Method section requires greater detail for reproducibility and to fully support the central claim. In the revised manuscript, we will provide the explicit mathematical definition of the antiderivative-based nonlinear transformation, including the integration limits, filtering parameters, and a derivation explaining its isolation of respiratory micro-motion. We will also add a sensitivity analysis to assess robustness to parameter variations. revision: yes

  2. Referee: [Results] Results section: All quantitative performance metrics (correlations with PSG channels, IRR RMSE of 0.027 ± 0.022 Hz) are reported exclusively on the expert-labeled 'high-quality' subset (average 22.2% of recording time). No breakdown of moderate/poor segments, failure-rate statistics, or characteristics of discarded data is given, raising the possibility that reported accuracy reflects selection of favorable segments rather than robust performance of the transformation.

    Authors: The emphasis on high-quality segments reflects the study's aim to characterize performance where respiratory information is reliably present, as determined by expert labeling. We acknowledge the need for fuller context on the entire dataset. In the revision, we will include a breakdown of quality label distributions, statistics on moderate and poor segments, failure rates, and characteristics of lower-quality or discarded data to provide a more complete assessment and address potential selection concerns. revision: yes

  3. Referee: [Results] Results section: No controlled experiments, simulations, or ablation studies are described to test whether the transformation suppresses or distorts non-respiratory components (e.g., cardiac ballistocardiography, voluntary micro-movements, or sensor drift) that are known to be present in fingertip TAA signals. Such tests would be required to substantiate that TAA-resp primarily reflects respiratory effort.

    Authors: The current validation leverages full-night PSG recordings across varied sleep stages and events, with correlations and RMI physiological alignment providing supporting evidence. However, we agree that targeted controlled tests would strengthen substantiation of the transformation's specificity. In the revised manuscript, we will add simulations or ablation studies to evaluate the transformation's handling of non-respiratory components such as cardiac signals, voluntary movements, and drift. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation and validation remain independent of fitted inputs or self-referential definitions.

full rationale

The paper proposes an antiderivative-based nonlinear transformation to obtain TAA-resp from raw fingertip TAA signals and applies standard time-frequency analysis to compute the respiratory motion index (RMI). These steps are presented as methodological choices rather than derived from or fitted to the target metrics. Validation uses simultaneous PSG recordings (independent ground truth for respiratory effort and IRR) plus expert annotations of signal quality on the transformed output. Reported quantities—correlation strengths, IRR RMSE on high-quality segments only, and RMI's ability to predict expert labels in leave-one-subject-out cross-validation—are computed against these external references. No equation or procedure in the provided text reduces the central performance numbers to a statistical tautology (e.g., no parameter fitted on a subset is then relabeled as a prediction of a closely related quantity on the same subset). Quality labeling criteria are established separately and applied by experts; the conditional reporting on high-quality segments is therefore a deliberate scope limitation rather than a circular selection that forces the result. Self-citation, if present in the full text, is not load-bearing for the uniqueness or correctness of the transformation or RMI. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 2 invented entities

Only the abstract is available, so the precise mathematical assumptions inside the antiderivative transformation and any implicit signal-modeling choices cannot be audited. No explicit free parameters, axioms, or invented physical entities are stated in the abstract.

invented entities (2)
  • TAA-resp no independent evidence
    purpose: Respiratory surrogate signal obtained from fingertip TAA via antiderivative nonlinear transformation
    Introduced as the output of the proposed method; no independent evidence outside the paper is provided in the abstract.
  • RMI no independent evidence
    purpose: Respiratory motion index derived from time-frequency analysis of TAA-resp
    New index created to quantify respiratory content; no independent evidence outside the paper is provided in the abstract.

pith-pipeline@v0.9.0 · 5652 in / 1455 out tokens · 72823 ms · 2026-05-08T08:49:18.094571+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

57 extracted references · 1 canonical work pages

  1. [1]

    Reconsider phase reconstruction in signals with dynamic periodicity from the modern signal processing perspective.Founda- tions of Data Science, 4(3):355–393, 2022

    Aymen Alian, Yu-Lun Lo, Kirk Shelley, and Hau-Tieng Wu. Reconsider phase reconstruction in signals with dynamic periodicity from the modern signal processing perspective.Founda- tions of Data Science, 4(3):355–393, 2022

    2022

  2. [2]

    New insights into the timing and potential mechanisms of respiratory-induced cortical arousals in obstructive sleep apnea.Sleep, 41(11):zsy160, 2018

    Jason Amatoury, Amy S Jordan, Barbara Toson, Chinh Nguyen, Andrew Wellman, and Danny J Eckert. New insights into the timing and potential mechanisms of respiratory-induced cortical arousals in obstructive sleep apnea.Sleep, 41(11):zsy160, 2018. RESPIRATION FROM FINGERTIP 29

    2018

  3. [3]

    An IMU-based wearable system for respiratory rate estimation in static and dynamic conditions.Cardiovascular Engineering and Technol- ogy, 14(3):351–363, 2023

    Alessandra Angelucci and Andrea Aliverti. An IMU-based wearable system for respiratory rate estimation in static and dynamic conditions.Cardiovascular Engineering and Technol- ogy, 14(3):351–363, 2023

    2023

  4. [4]

    The upper airway in sleep: physiology of the pharynx

    Indu Ayappa and David M Rapoport. The upper airway in sleep: physiology of the pharynx. Sleep medicine reviews, 7(1):9–33, 2003

    2003

  5. [5]

    Respiratory rate and flow waveform estimation from tri-axial accelerometer data

    Andrew Bates, Martin J Ling, Janek Mann, and Damal K Arvind. Respiratory rate and flow waveform estimation from tri-axial accelerometer data. In2010 International Conference on Body Sensor Networks, pages 144–150. IEEE, 2010

    2010

  6. [6]

    Jennifer Beck, Stewart B Gottfried, Paolo Navalesi, Yoanna Skrobik, Norman Comtois, Mauro Rossini, and Christer Sinderby. Electrical activity of the diaphragm during pressure support ventilation in acute respiratory failure.American journal of respiratory and critical care medicine, 164(3):419–424, 2001

    2001

  7. [7]

    Measurement of res- piratory rate with inertial measurement units.Current Directions in Biomedical Engineering, 6(3):237–240, 2020

    Simon Beck, Bernhard Laufer, Sabine Krueger-Ziolek, and Knut Moeller. Measurement of res- piratory rate with inertial measurement units.Current Directions in Biomedical Engineering, 6(3):237–240, 2020

    2020

  8. [8]

    Breathing pattern in humans: diversity and individuality.Respiration phys- iology, 122(2-3):123–129, 2000

    Gila Benchetrit. Breathing pattern in humans: diversity and individuality.Respiration phys- iology, 122(2-3):123–129, 2000

    2000

  9. [9]

    Aasm scoring manual updates for 2017 (version 2.4), 2017

    Richard B Berry, Rita Brooks, Charlene Gamaldo, Susan M Harding, Robin M Lloyd, Stu- art F Quan, Matthew T Troester, and Bradley V Vaughn. Aasm scoring manual updates for 2017 (version 2.4), 2017

    2017

  10. [10]

    Respiratory quality index design and validation for ecg and ppg derived respiratory data.Report for transfer of status, Dept

    Drew Birrenkott. Respiratory quality index design and validation for ecg and ppg derived respiratory data.Report for transfer of status, Dept. Eng. Sci., Univ. Oxford, Oxford, UK, 2015

    2015

  11. [11]

    An inexpensive accelerometer-based sleep- apnea screening technique

    Christie L Bucklin, Manohar Das, and Sam L Luo. An inexpensive accelerometer-based sleep- apnea screening technique. InProceedings of the IEEE 2010 National Aerospace & Electronics Conference, pages 396–399. IEEE, 2010

    2010

  12. [12]

    Physiology of sleep.Diabetes spectrum: a publication of the American Diabetes Association, 29(1):5, 2016

    David W Carley and Sarah S Farabi. Physiology of sleep.Diabetes spectrum: a publication of the American Diabetes Association, 29(1):5, 2016

    2016

  13. [13]

    Assessment of breathing parameters using an inertial measurement unit (IMU)-based system.Sensors, 19(1):88, 2018

    Ambra Cesareo, Ylenia Previtali, Emilia Biffi, and Andrea Aliverti. Assessment of breathing parameters using an inertial measurement unit (IMU)-based system.Sensors, 19(1):88, 2018

    2018

  14. [14]

    Validation of a fingertip home sleep apnea testing system using deep learning ai and a temporal event localization analysis.Sleep, 48(5):zsae317, 2025

    Ke-Wei Chen, Chun-Hsien Tseng, Hsin-Chien Lee, Wen-Te Liu, Kun-Ta Chou, and Hau- Tieng Wu. Validation of a fingertip home sleep apnea testing system using deep learning ai and a temporal event localization analysis.Sleep, 48(5):zsae317, 2025

    2025

  15. [15]

    Signal quality assessment of peripheral venous pressure.Journal of clinical monitoring and computing, 38(1):101–112, 2024

    Neng-Tai Chiu, Beau Chuang, Suthawan Anakmeteeprugsa, Kirk H Shelley, Aymen Awad Alian, and Hau-Tieng Wu. Signal quality assessment of peripheral venous pressure.Journal of clinical monitoring and computing, 38(1):101–112, 2024

    2024

  16. [16]

    Clinical guidelines for the use of unattended portable monitors in the diagnosis of obstructive sleep apnea in adult patients

    Nancy A Collop, W McDowell Anderson, Brian Boehlecke, David Claman, Rochelle Goldberg, Daniel J Gottlieb, David Hudgel, Michael Sateia, Richard Schwab, and Portable Monitoring Task Force of the American Academy of Sleep Medicine. Clinical guidelines for the use of unattended portable monitors in the diagnosis of obstructive sleep apnea in adult patients. ...

    2007

  17. [17]

    Marcelo A Colominas and Hau-Tieng Wu. An iterative warping and clustering algorithm to estimate multiple wave-shape functions from a nonstationary oscillatory signal.IEEE Transactions on Signal Processing, 71:701–712, 2023

    2023

  18. [18]

    Respiratory and circulatory control during sleep.Journal of Experimental Biology, 100(1):223–244, 1982

    JH Coote. Respiratory and circulatory control during sleep.Journal of Experimental Biology, 100(1):223–244, 1982

    1982

  19. [19]

    Respiratory monitoring: Current state of the art and future roads.IEEE Reviews in Biomedical Engineering, 15:103– 121, 2020

    Ian Costanzo, Devdip Sen, Lawrence Rhein, and Ulkuhan Guler. Respiratory monitoring: Current state of the art and future roads.IEEE Reviews in Biomedical Engineering, 15:103– 121, 2020

    2020

  20. [20]

    Synchrosqueezed wavelet transforms: an empirical mode decomposition-like tool.Appl

    Ingrid Daubechies, Jianfeng Lu, and Hau-Tieng Wu. Synchrosqueezed wavelet transforms: an empirical mode decomposition-like tool.Appl. Comput. Harmon. Anal., 30(2):243–261, 2011

    2011

  21. [21]

    A dif- ferential inertial wearable device for breathing parameter detection: hardware and firmware development, experimental characterization.Sensors, 22(24):9953, 2022

    Roberto De Fazio, Maria Rosaria Greco, Massimo De Vittorio, and Paolo Visconti. A dif- ferential inertial wearable device for breathing parameter detection: hardware and firmware development, experimental characterization.Sensors, 22(24):9953, 2022. 30 JEANNE LIN, LILY LIU, AND HAU-TIENG WU

    2022

  22. [22]

    Validation of respiratory signal derived from suprasternal notch accel- eration for sleep apnea detection

    Parastoo Kh Dehkordi, Marcin Marzencki, Kouhyar Tavakolian, Marta Kaminska, and Bozena Kaminska. Validation of respiratory signal derived from suprasternal notch accel- eration for sleep apnea detection. In2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pages 3824–3827. IEEE, 2011

    2011

  23. [23]

    Estimation of res- piration rate and sleeping position using a wearable accelerometer

    Emer P Doheny, Madeleine M Lowery, Audrey Russell, and Silke Ryan. Estimation of res- piration rate and sleeping position using a wearable accelerometer. In2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), pages 4668–4671. IEEE, 2020

    2020

  24. [24]

    Wireless respiratory monitoring and coughing detection using a wearable patch sensor network

    Tamer Elfaramawy, Cheikh Latyr Fall, Martin Morissette, Fran¸ cois Lellouche, and Benoit Gosselin. Wireless respiratory monitoring and coughing detection using a wearable patch sensor network. In2017 15th IEEE international new circuits and systems conference (NEW- CAS), pages 197–200. IEEE, 2017

    2017

  25. [25]

    Tidal volume variability and res- piration rate estimation using a wearable accelerometer sensor

    Atena Roshan Fekr, Katarzyna Radecka, and Zeljko Zilic. Tidal volume variability and res- piration rate estimation using a wearable accelerometer sensor. In2014 4th International Conference on Wireless Mobile Communication and Healthcare-Transforming Healthcare Through Innovations in Mobile and Wireless Technologies (MOBIHEALTH), pages 1–6. IEEE, 2014

    2014

  26. [26]

    Ronald A. Fisher. Tests of significance in harmonic analysis.Proc. R. Soc. Lond., 125(796):54– 59, 1929

    1929

  27. [27]

    Pedestrian tracking with shoe-mounted inertial sensors.IEEE Computer graphics and applications, 25(6):38–46, 2005

    Eric Foxlin. Pedestrian tracking with shoe-mounted inertial sensors.IEEE Computer graphics and applications, 25(6):38–46, 2005

    2005

  28. [28]

    Real-time detection of respiratory activity using an inertial measurement unit

    Henrik Gollee and Wei Chen. Real-time detection of respiratory activity using an inertial measurement unit. In2007 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pages 2230–2233. IEEE, 2007

    2007

  29. [29]

    Sleep apnea severity estimation from tracheal movements using a deep learning model.IEEE Access, 8:22641–22649, 2020

    Maziar Hafezi, Nasim Montazeri, Shumit Saha, Kaiyin Zhu, Bojan Gavrilovic, Azadeh Yadol- lahi, and Babak Taati. Sleep apnea severity estimation from tracheal movements using a deep learning model.IEEE Access, 8:22641–22649, 2020

    2020

  30. [30]

    Performance evaluation of a tri-axial accelerometry-based respiration monitoring for ambient assisted living

    Anmin Jin, Bin Yin, Geert Morren, Haris Duric, and Ronald M Aarts. Performance evaluation of a tri-axial accelerometry-based respiration monitoring for ambient assisted living. In2009 Annual international conference of the IEEE engineering in medicine and biology society, pages 5677–5680. IEEE, 2009

    2009

  31. [31]

    Respiratory system dynamics

    David A Kaminsky, Donald W Cockcroft, and Beth E Davis. Respiratory system dynamics. InSeminars in respiratory and critical care medicine, volume 44, pages 526–537. Thieme Medical Publishers, Inc., 2023

    2023

  32. [32]

    Effect of arousal on sympathetic overactivity in patients with obstructive sleep apnea.Sleep Medicine, 62:86–91, 2019

    Jung Bin Kim, Bo Sik Seo, and Ji Hyun Kim. Effect of arousal on sympathetic overactivity in patients with obstructive sleep apnea.Sleep Medicine, 62:86–91, 2019

    2019

  33. [33]

    Respiratory rate assessments using a dual-accelerometer device.Respiratory physiology & neurobiology, 191:60–66, 2014

    Sara Lapi, Federico Lavorini, Giovanni Borgioli, Marco Calzolai, Leonardo Masotti, Massimo Pistolesi, and Giovanni A Fontana. Respiratory rate assessments using a dual-accelerometer device.Respiratory physiology & neurobiology, 191:60–66, 2014

    2014

  34. [34]

    Assessment of transdiaphragmatic pressure in humans.Journal of applied physiology, 58(5):1469–1476, 1985

    D Laporta and A Grassino. Assessment of transdiaphragmatic pressure in humans.Journal of applied physiology, 58(5):1469–1476, 1985

    1985

  35. [35]

    Probabilistic analysis of scalogram ridges in signal processing.Quarterly of Applied Mathematics, 2026

    Gi-Ren Liu, Yuan-Chung Sheu, and Hau-Tieng Wu. Probabilistic analysis of scalogram ridges in signal processing.Quarterly of Applied Mathematics, 2026

    2026

  36. [36]

    Estimation of respiration rate from three-dimensional acceleration data based on body sensor network

    Guan-Zheng Liu, Yan-Wei Guo, Qing-Song Zhu, Bang-Yu Huang, and Lei Wang. Estimation of respiration rate from three-dimensional acceleration data based on body sensor network. Telemedicine and e-health, 17(9):705–711, 2011

    2011

  37. [37]

    Diaphragm electromyography using an oesophageal catheter: current concepts.Clinical science, 115(8):233–244, 2008

    Yuan Ming Luo, John Moxham, and Michael I Polkey. Diaphragm electromyography using an oesophageal catheter: current concepts.Clinical science, 115(8):233–244, 2008

    2008

  38. [38]

    Simultaneous activity and respiratory monitoring using an accelerometer

    Janek Mann, Roberto Rabinovich, Andrew Bates, S Giavedoni, W MacNee, and DK Arvind. Simultaneous activity and respiratory monitoring using an accelerometer. In2011 interna- tional conference on body sensor networks, pages 139–143. IEEE, 2011

    2011

  39. [39]

    Jacob McErlean, John Malik, Yu-Ting Lin, Ronen Talmon, and Hau-Tieng Wu. Unsupervised ensembling of multiple software sensors with phase synchronization: a robust approach for electrocardiogram-derived respiration.Physiological Measurement, 45(3):035008, 2024

    2024

  40. [40]

    A review of wearable sensors and systems with application in rehabilitation.Journal of neuroengineering and rehabilitation, 9(1):21, 2012

    Shyamal Patel, Hyung Park, Paolo Bonato, Leighton Chan, and Mary Rodgers. A review of wearable sensors and systems with application in rehabilitation.Journal of neuroengineering and rehabilitation, 9(1):21, 2012. RESPIRATION FROM FINGERTIP 31

    2012

  41. [41]

    Screening of obstructive sleep apnea syndrome by heart rate variability analysis.Circulation, 100(13):1411–1415, 1999

    Fr´ ed´ eric Roche, Jean-Michel Gaspoz, Isabelle Court-Fortune, Pascal Minini, Vincent Pi- chot, David Duverney, Fr´ ed´ eric Costes, Jean-Ren´ e Lacour, and Jean-Claude Barth´ el´ emy. Screening of obstructive sleep apnea syndrome by heart rate variability analysis.Circulation, 100(13):1411–1415, 1999

    1999

  42. [42]

    Nonrandom variability of respiration during sleep in healthy humans.Sleep, 28(4):411–417, 2005

    Sven Rostig, Jan W Kantelhardt, Thomas Penzel, Werner Cassel, J Hermann Peter, Claus Vogelmeier, Heinrich F Becker, and Andreas Jerrentrup. Nonrandom variability of respiration during sleep in healthy humans.Sleep, 28(4):411–417, 2005

    2005

  43. [43]

    Validation and comparison of actigraph activity monitors.Journal of science and medicine in sport, 14(5):411–416, 2011

    Jeffer E Sasaki, Dinesh John, and Patty S Freedson. Validation and comparison of actigraph activity monitors.Journal of science and medicine in sport, 14(5):411–416, 2011

    2011

  44. [44]

    REM sleep behaviour dis- order: an update on a series of 96 patients and a review of the world literature.Journal of sleep research, 2(4):224–231, 1993

    Carlos H Schenck, Thomas D Hurwitz, and Mark W Mahowald. REM sleep behaviour dis- order: an update on a series of 96 patients and a review of the world literature.Journal of sleep research, 2(4):224–231, 1993

    1993

  45. [45]

    The use of joint time frequency analysis to quantify the effect of ventilation on the pulse oximeter waveform

    Kirk H Shelley, Aymen A Awad, Robert G Stout, and David G Silverman. The use of joint time frequency analysis to quantify the effect of ventilation on the pulse oximeter waveform. Journal of clinical monitoring and computing, 20(2):81–87, 2006

    2006

  46. [46]

    Entropy-based time-varying window width selection for nonlinear-type time-frequency analysis.Int

    Yae-Lin Sheu, Liang-Yan Hsu, Pi-Tai Chou, and Hau-Tieng Wu. Entropy-based time-varying window width selection for nonlinear-type time-frequency analysis.Int. J. Data. Sci. Anal., 3:231–245, 2017

    2017

  47. [47]

    Model- based assessment of photoplethysmogram signal quality in real-life environments

    Yan-Wei Su, Chia-Cheng Hao, Gi-Ren Liu, Yuan-Chung Sheu, and Hau-Tieng Wu. Model- based assessment of photoplethysmogram signal quality in real-life environments. In2024 32nd European Signal Processing Conference (EUSIPCO), pages 1726–1730. IEEE, 2024

    2024

  48. [48]

    Ridge detection for nonsta- tionary multicomponent signals with time-varying wave-shape functions and its applications

    Yan-Wei Su, Gi-Ren Liu, Yuan-Chung Sheu, and Hau-Tieng Wu. Ridge detection for nonsta- tionary multicomponent signals with time-varying wave-shape functions and its applications. IEEE Transactions on Signal Processing, 72:4843–4854, 2024

    2024

  49. [49]

    IET, 2004

    David Titterton and John L Weston.Strapdown inertial navigation technology, volume 17. IET, 2004

    2004

  50. [50]

    Sensing systems for respiration monitoring: A technical systematic review.Sensors, 20(18):5446, 2020

    Erik Vanegas, Raul Igual, and Inmaculada Plaza. Sensing systems for respiration monitoring: A technical systematic review.Sensors, 20(18):5446, 2020

    2020

  51. [51]

    John Wiley & Sons, 2009

    John G Webster.Medical instrumentation: application and design. John Wiley & Sons, 2009

    2009

  52. [52]

    Lippincott Williams & Wilkins, 2020

    John B West and Andrew M Luks.West’s respiratory physiology. Lippincott Williams & Wilkins, 2020

    2020

  53. [53]

    H.-T. Wu. Instantaneous frequency and wave shape functions (I).Appl. Comput. Harmon. Anal., 35:181–199, 2013

    2013

  54. [54]

    Hau-Tieng Wu and Jaroslaw Harezlak. Application of de-shape synchrosqueezing to estimate gait cadence from a single-sensor accelerometer placed in different body locations.Physiolog- ical Measurement, 44(5):055009, 2023

    2023

  55. [55]

    Frequency detection and change point estimation for time series of complex oscillation.Journal of the American Statistical Association, pages 1–12, 2024

    Hau-Tieng Wu and Zhou Zhou. Frequency detection and change point estimation for time series of complex oscillation.Journal of the American Statistical Association, pages 1–12, 2024

    2024

  56. [56]

    Uncertainty quantification of synchrosqueezing transform under complicated nonstationary noise.arXiv preprint arXiv:2506.00779, 2025

    Hau-Tieng Wu and Zhou Zhou. Uncertainty quantification of synchrosqueezing transform under complicated nonstationary noise.arXiv preprint arXiv:2506.00779, 2025

    work page arXiv 2025

  57. [57]

    Esophageal pressure monitoring: why, when and how?Current opinion in critical care, 24(3):216–222, 2018

    Takeshi Yoshida and Laurent Brochard. Esophageal pressure monitoring: why, when and how?Current opinion in critical care, 24(3):216–222, 2018. AppendixA.A quick review of synchrosqueezing transform We begin with the short-time Fourier transform (STFT), which forms the foun- dation of the synchrosqueezing transform (SST) [20]. The STFT is a widely used too...

    2018