NZNejib Zemzemi

Nejib Zemzemi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Nejib Zemzemi has a total of 162 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Enhancing Asleep Deep Brain Stimulation Targeting for Essential Tremor Using Machine Learning: The OPTIVIM Phase 2 Study

    by Julien Engelhardt, Emile Simon, Nejib Zemzemi, Camille Dallies-Labourdette, Edouard Courtin, Remi Sitta, Antoine Benard, Olivier Branchard, Michel Goillandeau, Chloé Laurencin, et al.

    BACKGROUND AND OBJECTIVES: Deep brain stimulation (DBS) is an effective treatment of essential tremor, but the optimal target and how to reach it with the best accuracy remain controversial. The OptimDBS algorithm is based on Kernel Ridge Regression and was trained on a database of patients who underwent DBS with optimal postoperative outcomes. Using 3-dimensional T1-weighted MRI as the only input, it calculates the stereotactic coordinates of an effective DBS target for treating tremor. The aim of the study was to evaluate the efficacy and safety of OptimDBS-guided asleep DBS without peroperative clinical and electrophysiological controls in an independent cohort of patients. METHODS: The OPTIVIM study was a prospective, single-arm, multicenter Phase 2 trial. The primary outcome was tremor reduction and improvement in activities of daily living as assessed by the Fahn-Tolosa-Marin scale 3 months after surgery. Secondary outcomes included reduction in tremor amplitude on accelerometry, quality of life (modified Parkinson’s Disease Questionnaire-39), ataxia (Scale for the Assessment and Rating of Ataxia), adverse events, and anatomic location of active contacts and the volume of tissue activated, all measured 3 months after surgery. RESULTS: Twenty-two patients from 2 centers were enrolled. Fahn-Tolosa-Marin scores improved by 61.3% (95% CI: 53.7%-68.9%). Accelerometry (mean ± SD) showed a reduction in postural tremor amplitude of 81% ± 56% on the right side and 84% ± 35% on the left side. Modified Parkinson’s Disease Questionnaire-39 scores (median [Q1-Q3]) improved by 55% [24.3%-77.8%]. Scale for the Assessment and Rating of Ataxia scores remained stable (median [Q1-Q3], preoperative vs postoperative: 4 [3-6] vs 3 [2-6]). There were no serious adverse events. Active contacts were located in the posterior subthalamic area (59%) or the ventral-intermediate nucleus of the thalamus (20%), with 100% of volume of tissue activated overlapping the ventral-intermediate nucleus of the thalamus or posterior subthalamic area. CONCLUSION: The results of OptimDBS-guided asleep DBS seem to be comparable with those reported in the literature for awake and asleep DBS and ablative techniques. Long-term evaluations with larger cohorts are needed to confirm these results.

  • Open Access

    How does the optimal constant of the Carleman estimate depend on the size of the observation boundary?

    by Hamza Ammar, Moncef Mahjoub, Néjib Zemzemi

    In this paper, we would like to elucidate the question of the Carleman estimate constants dependency on the size of the observation boundary. This allows to quantify the effect of measurements uncertainty with respect to the size of the observation boundary in many inverse and control problems. We take an example of an inverse source problem for a parabolic equation and explicitly calculate Lipschitz stability constants that appears in the L 2 -estimate of the source function. We deliberately construct a class of space dependent weight functions that depend on the measurement boundary size. Then we identify the optimal weight function that allows to minimize the stability constant. Using our approach, we found that when the measurement boundary covers 80% or more of the domain boundary, we can explicitly provide the formula of the optimal constant. When the measurement boundary is less than 80%, we are not able to find the explicit formula of the optimal expression, but we are able to numerically approximate the optimal constant. This paper presents meticulous calculations that require a rigorous and careful reading to fully comprehend the technical intricacies of the analysis.

  • Open Access

    CEPS: a cardiac electrophysiology exploration tool for new mathematical models and methods

    by Michael Leguèbe, Yves Coudière, Loïc Calvez, Valentin Pannetier, Charlie Douanla-Lontsi, Anđela Davidović, Pierre-Elliott Bécue, Mehdi Juhoor, Nejib Zemzemi

    Cardiac activity is regulated by the propagation within the heart muscle of an electric action potential (AP), which follows a specific synchronized pattern in normal conditions. Defects in the tissue or metabolic changes can lead to pathological propagations and cardiac arrhythmias, which are a major cause of death around the world. Electrophysiology simulation tools at tissue scale are required to better understand the normal and pathological function of the heart, and could also be used as design tools for new medical devices and therapies.


  • Global Lipschitz stability of multiple space dependent ionic parameters for the MEA/hiPSC-CM assays

    by Hamza Ammar, Moncef Mahjoub, Nejib Zemzemi

    In this paper, we consider an inverse problem of determining two space dependent ionic parameters of a strongly coupled parabolic-elliptic reaction–diffusion system arising in cardiac electrophysiology modeling when simulating drugs action with multi-electrode array/human induced pluripotent stem cells-cardiomyocytes assays. We use the bidomain model coupled to an ordinary differential equation and we consider the classical phenomenological model in cardiac electrophysiology of FitzHugh–Nagumo to describe the ionic exchanges at the microscopic level. Our main result is the uniqueness and a Lipschitz stability estimate for two ionic parameters ( k , γ ) of the model using sub-boundary observations over an interval of time. The key ingredients are a global Carleman-type estimates with a suitable observations acting on a part of the boundary.


  • Open Access

    MUSIC: Cardiac Imaging, Modelling and Visualisation Software for Diagnosis and Therapy

    by Mathilde Merle, Florent Collot, Julien Castelneau, Pauline Migerditichan, Mehdi Juhoor, Buntheng Ly, Valery Ozenne, Bruno Quesson, Nejib Zemzemi, Yves Coudière, et al.

    The tremendous advancement of cardiac imaging methods, the substantial progress in predictive modelling, along with the amount of new investigative multimodalities, challenge the current technologies in the cardiology field. Innovative, robust and multimodal tools need to be created in order to fuse imaging data (e.g., MR, CT) with mapped electrical activity and to integrate those into 3D biophysical models. In the past years, several cross-platform toolkits have been developed to provide image analysis tools to help build such software. The aim of this study is to introduce a novel multimodality software platform dedicated to cardiovascular diagnosis and therapy guidance: MUSIC. This platform was created to improve the image-guided cardiovascular interventional procedures and is a robust platform for AI/Deep Learning, image analysis and modelling in a newly created consortium with international hospitals. It also helps our researchers develop new techniques and have a better understanding of the cardiac tissue properties and physiological signals. Thus, this extraction of quantitative information from medical data leads to more repeatable and reliable medical diagnoses.


  • Open Access

    In-Silico Data Based Machine Learning Technique Predicts Premature Ventricular Contraction Origin Coordinates

    by Andony Arrieula, Hubert Cochet, Pierre Jais, Michel Haissaguerre, Nejib Zemzemi, Mark Potse

    Premature ventricular contraction (PVC) can induce ventricular tachycardia or ventricular fibrillation. Drug-resistant PVCs can be cured by catheter ablation, but the accurate localization that this requires can be difficult and time-consuming. An accurate pre-procedural estimate of the origin could make the procedure more efficient. We propose a machine-learning method for accurate pre-procedural origin estimation. It uses a database of paced 12-lead ECGs with known pacing locations and presents its results on an imaging-based model of the patient. The method was tested using 7 realistic heart-torso models with hundreds of PVCs everywhere in the ventricles. We found that increasing the number of patients in the training database increased the accuracy of the predictions. The optimal number of pacing sites per patient in the training dataset was about 25, resulting in a prediction error around 15 mm. We conclude that our method gives a good indication to clinicians to efficiently start a pace-mapping during a catheter ablation procedure. It can be complemented with an intra-procedural method that uses the patients own paced beats to refine the prediction.


  • Open Access

    Benchmark of deep learning algorithms for the automatic screening in electrocardiograms transmitted by implantable cardiac devices

    by Narimane Gassa, Benjamin Sacristan, Nejib Zemzemi, Maxime Laborde, Juan Garrido Oliver, Clara Matencio Perabla, Guillermo Jimenez-Perez, Oscar Camara, Sylvain Ploux, Marc Strik, et al.

    The objective of this work was to benchmark different deep learning architectures for noise detection against cardiac arrhythmia episodes recorded by pacemakers and implantable cardioverter-defibrillators (PM/ICDs) and transmitted for remote monitoring. Up to now, most signal processing from ICD data has been based on classical hand-crafted algorithms, not AI or DL-based ones. The database consist of PM/ICD data from 805 patients representing a total of 10471 recordings from three different channels: the right ventricular (RV), the right atria (RA), and the shock channel. Four deep learning approaches were trained and optimized to classify PM/ICDs' records as actual ventricular signal vs noise episodes. We evaluated the performance of the different models using the F2 score. Results show that the use of 2D representations of 1D signals led to better performances than the direct use of 1D signals, suggesting that the detection of noise takes advantage of a spectral decomposition of the signal, which remains to be confirmed in other contexts. This study proposes deep learning approaches for the analysis of remote monitoring recordings from PM/ICDs. The detection of noise allows efficient management of this large daily flow of data.


  • Open Access

    Cardiac Activation Maps Reconstruction: A Comparative Study Between Data-Driven and Physics-Based Methods

    by Amel Karoui, Mostafa Bendahmane, Nejib Zemzemi

    One of the essential diagnostic tools of cardiac arrhythmia is activation mapping. Noninvasive current mapping procedures include electrocardiographic imaging. It allows reconstructing heart surface potentials from measured body surface potentials. Then, activation maps are generated using the heart surface potentials. Recently, a study suggests to deploy artificial neural networks to estimate activation maps directly from body surface potential measurements. Here we carry out a comparative study between the data-driven approach DirectMap and noninvasive classic technique based on reconstructed heart surface potentials using both Finite element method combined with L1-norm regularization (FEM-L1) and the spatial adaptation of Time-delay neural networks (SATDNN-AT). In this work, we assess the performance of the three approaches using a synthetic single paced-rhythm dataset generated on the atria surface. The results show that data-driven approach DirectMap quantitatively outperforms the two other methods. In fact, we observe an absolute activation time error and a correlation coefficient, respectively, equal to 7.20ms, 93.2% using DirectMap, 14.60ms, 76.2% using FEM-L1 and 13.58ms, 79.6% using SATDNN-AT. In addition, results show that data-driven approaches (DirectMap and SATDNN-AT) are strongly robust against additive gaussian noise compared to FEM-L1.