Abhijith Mundanad Narayanan

Ph.D.

About Me

I am a signal processing engineer with a Ph.D. in Electrical Engineering from KU Leuven, Belgium. My research has focused on signal processing and data analysis. My expertise is in developing and implementing algorithms to analyse multi-channel time series data, through the experience working with EEG data. Strong understanding of fundamentals of machine learning and experienced in implementation of linear through deep-learning models. I have had previous experience in audio signal processing and telecommunications engineering. Surely, you noticed the very long name, but you can call me Abhi.

I am most skilled in: Signal and data analysis in MATLAB and Python.

Data analysis tools: pandas,numpy,scipy

Visualization Tools: dash,plotly

I am also competent in: linux/bash,R, scikit-learn, tensorflow

Projects (Scientific)

Neureka Seizure Challenge 2020 1st place submission

https://github.com/mabhijithn/irregulars-neureka-codebase

Code developed as part of team Biomed Irregulars from STADIUS research group of KU Leuven. Winners of Neureka Seizure prediction challenge 2020

Scripts and notebooks developed by the team Biomed Irregulars from STADIUS research group of KU Leuven. The winning submission to Neureka 2020 epilepsy challenge(results). The challenge was to develop a model to detect epileptic seizures in the TUH Seizure Dataset.

Channel selection in least-squares (LS) problem

github.com/AlexanderBertrandLab/channel-select

A fast and easilty interpretable channel (feature) selection in LS-based problems.

MATLAB and python code implementing utility-based channel selection for least-squares(LS) problems. Developed to be applied in EEG-based auditory attention decoding problems, the tool can be used for any LS problem. In neural-decoding, the channel selection method can be used in neural decoding of speech for future neuro-steered hearing aids. It has been also used for feature-selection in classification.

A python toolbox for doctors, practioners and researchers implementing a web application to read, analyse and visualize EEG collected and distributed using various formats.

There are plenty of toolboxes available online for EEG analysis in python. However, most of these toolboxes involve a steep learning curve. Moreover, easy, interactive visualization is a challenge of many such toolboxes. This toolbox is start of a journey towards a simple, easily useable toolbox in python which will help researchers, doctors and practitioners use a web browser to read, analyse and visualize EEG recorded and distributed in various formats.

Projects (Recreational)

An Edge-AI project. Real-time sound classification model deployed on a Raspberry Pi. Github link https://github.com/mabhijithn/urbannoise-edgeai

Impact of goal scorers in football

https://goalscorers-21st-century.herokuapp.com/

Analysing the impact of goal scorers of the English Premier League in 21st century.

An interactive tool to visualize the different goal scorers based on the impact of their goals on their team’s results or performance in a match. The entire data used in the visualization was collected personally with the help web-scraping using Beautiful Soup and some nifty python based structuring of the data.

Achievements

The recipient of the IEEE SPMB Best Paper Award on their 10th anniversary.

For the paper titled Epileptic Seizure Detection in EEG via Fusion of Multi-View Attention-Gated U-net Deep Neural Networks by C. Chatzichristos, J.Dan, A.M.Narayanan, et. al. The paper described in detail the model used to win Neureka Seizure Challenge 2020.

First Place in Neureka Seizure Challenge 2020

https://neureka-challenge.com/results/

First place in the world-wide seizure detection challenge organized by Novela Neurotech and Temple University

The winning submission to Neureka 2020 epilepsy challenge(results). The challenge was to develop a model to detect epileptic seizures in the TUH Seizure Dataset.

Experience

Indigo Diabetes

Data Scientist

October 2021 - Present

https://indigomed.com/

Developers of the world's first implantable Continuous Metabolite Measurement (CMM) sensor.

Analysing data measured by world’s first CMM sensor for continuous monitoring and prediction of glucose, ketone and other analytes in blood.

KU Leuven

Doctoral Researcher

October 2017 - September 2021

https://www.kuleuven.be/kuleuven/

Part of biomedical data processing research group in the most innovative university of the world (Reuters).

Developing data processing, analysis and machine learning techniques on biomedical data. Specific focus on electroencephalography (EEG) data.

Qualcomm

Signal Processing Engineer

July 2014 - September 2016

www.qualcomm.com

Initially part of Ikanos communications, which merged with Qualcomm, working on signal processing aspects of residential gateways and COs.

Involved in software development of residential gateways through upgrading in echo canceller system. Contributed towards the development of firmware of Gigabit DSL chip of Qualcomm.

Education

KU Leuven, Belgium

PhD Electrical Engineering

December, 2016 - September, 2021

In a renowned biomedical signal processing research group in one of the top universities in the world.

Analysing biomedical data and developing data analysis algorithms from the perspective of simultaneous multiple sensors. The focus of research is building towards future miniature biomedical sensor networks, specifically neuro-sensor networks.

Indian Institute of Science, Bangalore India

Masters in Signal Processing

2012 - 2014

The No.1 scientific research institute in India

Worked along with the nation’s leading researchers in the area, as part of the graduate program in signal processing leading to a publication in the top international conference in signal processing (ICASSP, 2014).

Publications

Mundanad Narayanan A., Zink R. and Bertrand A., 2021 EEG miniaturization limits for stimulus decoding with EEG sensor networks, Journal of Neural Engineering, vol. 18, no. 5, Oct. 2021.

A. Villa, A.M. Narayanan, S. Van Huffel, A. Bertrand, C. Varon 2021. Utility metric for unsupervised feature selection. PeerJ Computer Science 7:e477 https://doi.org/10.7717/peerj-cs.477

A.M. Narayanan, P.Patrinos and A.Bertrand Optimal versus approximate channel selection methods for EEG decoding with application to topology-constrained neuro-sensor networks IEEE Transactions on Neural Systems and Rehabilitation Engineering (TNSRE) (2020), DOI: 10.1109/TNSRE.2020.3035499

C. Chatzichristos, J. Dan, A.M. Narayanan, N. Seeuws, K. Vandecasteele, M. De Vos, A. Bertrand and S. Van Huffel Epileptic Seizure Detection in EEG via Fusion of Multi-View Attention-Gated U-net Deep Neural Networks 2020 IEEE Signal Processing in Medicine and Biology Symposium (SPMB). IEEE, 2020.

A.M. Narayanan and A. Bertrand Analysis of miniaturization effects and channel selection strategies for EEG sensor networks with application to auditory attention detection IEEE Transactions on Biomedical Engineering, vol. 67, no.1, pp. 234-244, 2020.

A.M. Narayanan and A. Bertrand Group-utility metric for efficient sensor selection and removal in LCMV beamformers Proc. IEEE International Conference on Acoustics, Speech and Signal processing (ICASSP), Barcelona, Spain, May 2020.

A.M. Narayanan, A. Bertrand The effect of miniaturization and galvanic separation of EEG sensor devices in an auditory attention detection task Proc. International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Honolulu, Hawai (USA), July 2018.

A.M. Narayanan, P. K. Ghosh, and K. Rajgopal. Multi-pitch tracking using Gaussian mixture model with time varying parameters and grating compression transform 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2014.