Doctral school MIPTIS – Orleans University

Development and theoretical characterization of new complexity measures. Application to sEMG signals

Supervisors: Olivier Buttelli and Meryem JABLOUN, associate professors at Orleans University- PRISME laboratory, 12 rue de Blois, 45067 Orleans, France.

meryem.jabloun@univ-orleans.fr

olivier.buttelli@univ-orleans.fr

Thesis Proposal (September 2024): The concept of signal complexity is often associated with the amount of order or disorder in a system. Evaluating this complexity is a crucial task in the study of dynamic systems across various fields such as medical diagnostics, mechanical system fault analysis, astrophysics, and many others.

Although entropy measures, potential estimators of complexity, have undergone significant developments, the theoretical characterization of most of these measures has often been lacking. The work conducted within Mr. Davalos's thesis, followed by one year of post-doctoral research, focused on the statistical characterization of multi-scale permutation entropies (published in 5 international journals). These efforts contributed to improve parameter selection for these entropies, ensuring a better interpretation of results within the context of studying muscle fatigue using surface electromyographic (sEMG) signals.

The proposed thesis continues in the same vein as this previous work: Developing and statistically characterizing new complexity measurement tools that will allow for a better quantification of the complexity of nonstationary signals and system dynamics. Several strategies can be considered, such as modifying and/or combining entropy measures from different philosophies (signal value distribution, pattern distribution, frequency distribution, etc.)... We aim at identifying the best entropies offering the most effective classification of sEMG signals, for example, into pathological or normal classes, using sEMG data collected from a population of individuals with Parkinson's disease.
This classification will serve as evidence of the effectiveness of the medication and therapy utilized for patients with such a disease.

The subject of this thesis aligns with the research focus of the signal team at the PRISME laboratory, particularly concerning the analysis and processing of bioelectrical signals (sEMG). Previous master internships and completed theses have been supervised on this topic. Furthermore, this thesis work will be closely linked to the research group on complexity measures, a group we are establishing within the GDR IASIS. Additionally, a collaboration within the ATHENA project with Professor Alès Holobar from the University of Maribor is planned for the supervision of this thesis.

Resources available: The PHD student will have access to datasets annotated by experts. Signals were collected from patients with Parkinson's disease through previous project, the international ANR/Taiwan EcoTech project. The thesis is funded by a French ministerial scholarship.

Deadline for application: 15th of mai, 2024.

Publications related to the subject:

Meryem Jabloun, "Local Legendre polynomial fitting-based preprocessing for improving the interpretation of permutation entropy in stationary time series", 31st European Signal Processing Conference 2023 in Helsinki, Finland.

Meryem Jabloun, Philippe Ravier, and Olivier Buttelli, Improving the interpretation of linear filtering preprocessing-based multiscale permutation entropy", the 2023 IEEE Statistical Signal Processing Workshop (SSP), 190-194.

Meryem Jabloun, Philippe Ravier and Olivier Buttelli. « On the genuine relevance of the data-driven signal decomposition-based multiscale permutation entropy. Entropy, 24(10):1343, September 2022.

Philippe Ravier, Antonio Davalos, Meryem Jabloun, and Olivier Buttelli. The Refined Composite Downsampling Permutation Entropy Is a Relevant Tool in the Muscle Fatigue Study Using sEMG Signals. Entropy, 23(12):1655, December 2021

Antonio Davalos, Meryem Jabloun, Philippe Ravier, and Olivier Buttelli. The Impact of Linear Filter Preprocessing in the Interpretation of Permutation Entropy. Entropy, 23(7) :787, July 2021.

Antonio Davalos, Meryem Jabloun, Philippe Ravier, and Olivier Buttelli. Improvement of Statistical Performance of Ordinal Multiscale Entropy Techniques Using Refined Composite Downsampling Permutation Entropy. Entropy, 23(1):30, January 2021b.

Antonio Davalos, Meryem Jabloun, Philippe Ravier, and Olivier Buttelli. On the Statistical Properties of Multiscale Permutation Entropy: Characterization of the Estimator's Variance. Entropy, 21(5):450, May 2019a.

Abdelbassit Boualem, Meryem Jabloun, Philippe Ravier, Olivier Buttelli, “Legendre polynomial modeling of time-varying delay applied to surface EMG signals—Derivation of the appropriate time-dependent CRBs”, Signal Processing, Volume 114, 2015, Pages 34-44.