Asymptotic analysis of synchrosqueezing transform—toward statistical inference with nonlinear-type time-frequency analysis

Sourisseau, Matt; Wu, Hau‐Tieng; Zhou, Zhou

doi:10.1214/22-aos2203

Cited by 6 publications

(2 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…is a sharpened spectrum of the oscillatory signal at time t. We refer interested readers to a nonmathematical tutorial (Wu et al 2021), and several recent developments (Chen et al 2014, Sourisseau et al 2022 for the idea and discussions of SST. Similarly, the deSTFT can be sharpened by…”

Section: Resultsmentioning

confidence: 99%

Application of de-shape synchrosqueezing to estimate gait cadence from a single-sensor accelerometer placed in different body locations

Harezlak

2023

Physiol. Meas.

Self Cite

View full text Add to dashboard Cite

Objective: Commercial and research-grade wearable devices have become increasinglypopular over the past decade. Information extracted from devices using accelerometers is fre- quently summarized as “number of steps” (commercial devices) or “activity counts” (research- grade devices). Raw accelerometry data that can be easily extracted from accelerometers used in research, for instance ActiGraph GT3X+, are frequently discarded.Approach: Our primary goal is proposing an innovative use of the de-shape synchrosqueez- ing transform to analyze the raw accelerometry data recorded from a single sensor installed in different body locations, particularly the wrist, to extract gait cadence when a subject is walking. The proposed methodology is tested on data collected in a semi-controlled experiment with 32 participants walking on a one-kilometer predefined course. Walking was executed on a flat surface as well as on the stairs (up and down).Main Results: The cadences of walking on a flat surface, ascending stairs, and descending stairs, determined from the wrist sensor, are 1.98±0.15 Hz, 1.99±0.26 Hz, and 2.03±0.26 Hz respectively. The cadences are 1.98±0.14 Hz, 1.97±0.25 Hz, and 2.02±0.23 Hz, respectively if determined from the hip sensor, 1.98±0.14 Hz, 1.93±0.22 Hz and 2.06±0.24 Hz, respectively if determined from the left ankle sensor, and 1.98±0.14 Hz, 1.97±0.22 Hz, and 2.04±0.24 Hz, respectively if determined from the right ankle sensor. The difference is statistically significant indicating that the cadence is fastest while descending stairs and slowest when ascending stairs. Also, the standard deviation when the sensor is on the wrist is larger. These findings are in line with our expectations.Conclusion: We show that our proposed algorithm can extract the cadence with high accuracy, even when the sensor is placed on the wrist.

show abstract

Section: Resultsmentioning

confidence: 99%

Application of de-shape synchrosqueezing to estimate gait cadence from a single-sensor accelerometer placed in different body locations

Harezlak

2023

Physiol. Meas.

Self Cite

View full text Add to dashboard Cite

show abstract

“…For instance, in the case of f(t) = A(t) cos(2πϕ(t)), the TFR determined by SST concentrates around ϕ ′ l (t) while encoding the AM function A(t) as the intensity. This representation is notably less influenced by the choice of window and robust the noise (Daubechies et al 2011, Chen et al 2014, Sourisseau et al 2022. Although not necessary for our current discussion, it is worth noting that SST allows users to achieve more signal processing missions, e.g.…”

Section: Data Availability Statementmentioning

confidence: 99%

Unveil sleep spindles with concentration of frequency and time (ConceFT)

Shimizu,

2024

Physiol. Meas.

View full text Add to dashboard Cite

Objective: Sleep spindles contain crucial brain dynamics information. We intro- duce the novel non-linear time-frequency analysis tool “Concentration of Fre- quency and Time” (ConceFT) to create an interpretable automated algorithm for sleep spindle annotation in EEG data and to measure spindle instantaneous frequencies (IFs). Approach: ConceFT effectively reduces stochastic EEG in- fluence, enhancing spindle visibility in the time-frequency representation. Our automated spindle detection algorithm, ConceFT-Spindle (ConceFT-S), is com- pared to A7 (non-deep learning) and SUMO (deep learning) using Dream and MASS benchmark databases. We also quantify spindle IF dynamics. Main Re- sults: ConceFT-S achieves F1 scores of 0.765 in Dream and 0.791 in MASS, which surpass A7 and SUMO. We reveal that spindle IF is generally nonlin- ear. Significance: ConceFT offers an accurate, interpretable EEG-based sleep spindle detection algorithm and enables spindle IF quantification.

show abstract