dr. B. Hunyadi

Assistant Professor
Circuits and Systems (CAS), Department of Microelectronics


Promotor:

Expertise: Biomedical signal processing

Themes: Health and Wellbeing

Biography

Borbála (Bori) Hunyadi was born in Budapest, Hungary. She received a MSc degree in electrical and computer engineering from the Pazmany Peter Catholic University in 2009. In the same year she joined Stadius, Department of Electrical Engineering at KU Leuven, where she worked in close collaboration with the Laboratory for Epilepsy Research, and she obtained her PhD degree in 2014. She continued working in Stadius as a postdoctoral researcher on the ERC advanced grant Biotensors, and she served as the research lead on the Imec-ICON project SeizeIT. Between February and May 2016 she was a visiting researcher at the University of Oxford, and in May 2017 she visited the Christian Albrechts University of Kiel. In 2018 she was awarded one of the “Delft Technology Fellowships” for outstanding female academic researchers. In October 2018 she joined the Circuits and Systems group at TU Delft as an assistant professor.

Her research interests include biomedical signal processing and machine learning for biomedical pattern recognition. More specifically, she is interested in multimodal signal processing and fusion, blind source separation, tensor decompositions and wearable signal processing to better understand healthy and pathological physiology, in particular brain activity in epilepsy.

EE2S31 Signal processing

Digital signal processing; stochastic processes

EE4C03 Statistical digital signal processing

A second course on digital signal processing: random signals, covariances, linear prediction, Levinson and Schur algorithm, spectrum estimation, optimal filtering, Wiener and Kalman filters, LMS and RLS algorithm

Multimodal, multiresolution brain imaging

Developing a novel brain imaging paradigm combining functional ultrasound and EEG

Medical Delta Cardiac Arrhythmia Lab

Part of a larger program (with Erasmus MC) to unravel and target electropathology related to atrial arrhythmia

  1. Development of temporal lobe epilepsy during maintenance electroconvulsive therapy: A case of human kindling?
    C. Schotte; E. Cleeren; K. Goffin; B. Hunyadi; S. Buggenhout; K. Van Laere; W. Van Paesschen;
    Epilepsia Open,
    Volume 4, Issue 1, pp. 200-205, 2019. DOI: 10.1002/epi4.12294
    document

  2. Semi-automated EEG enhancement improves localization of ictal onset zone with EEG-correlated fMRI
    S. Van Eyndhoven; B. Hunyadi; P. Dupont; W. Van Paesschen; S. Van Huffel;
    Frontiers in Neurology,
    Volume 10, 2019. DOI: 10.3389/fneur.2019.00805
    document

  3. Nonconvulsive epileptic seizure monitoring with incremental learning
    Y.R. Rodriguez Aldana; E.J. Maranon Reyes; F. Sanabria Macias; V. Rodriguez Rodriguez; L. Morales Chacon; S. Van Huffel; B. Hunyadi;
    Computers in Biology and Medicine,
    Volume 114, pp. 103434, 2019. ISSN 0010-4825. DOI: 10.1016/j.compbiomed.2019.103434
    Keywords: ... Nonconvulsive epileptic seizures, Hilbert huang transform, Multiway data analysis, Incremental learning.

    Abstract: ... Nonconvulsive epileptic seizures (NCSz) and nonconvulsive status epilepticus (NCSE) are two neurological entities associated with increment in morbidity and mortality in critically ill patients. In a previous work, we introduced a method which accurately detected NCSz in EEG data (referred here as ‘Batch method’). However, this approach was less effective when the EEG features identified at the beginning of the recording changed over time. Such pattern drift is an issue that causes failures of automated seizure detection methods. This paper presents a support vector machine (SVM)-based incremental learning method for NCSz detection that for the first time addresses the seizure evolution in EEG records from patients with epileptic disorders and from ICU having NCSz. To implement the incremental learning SVM, three methodologies are tested. These approaches differ in the way they reduce the set of potentially available support vectors that are used to build the decision function of the classifier. To evaluate the suitability of the three incremental learning approaches proposed here for NCSz detection, first, a comparative study between the three methods is performed. Secondly, the incremental learning approach with the best performance is compared with the Batch method and three other batch methods from the literature. From this comparison, the incremental learning method based on maximum relevance minimum redundancy (MRMR_IL) obtained the best results. MRMR_IL method proved to be an effective tool for NCSz detection in a real-time setting, achieving sensitivity and accuracy values above 99%.

    document

BibTeX support

Last updated: 30 Apr 2019