Alexander Neergaard Zahid

Postdoctoral researcher

I am a research scientist and biomedical engineer with special interests in signal processing and machine learning for computational sleep science. I design algorithms and data processing pipelines for modeling, analysis and visualisation of physiological signals such as EEG, EOG and EMG obtained from nocturnal PSG recordings. Through this work, I hope to further our collective understanding of sleep pathologies, and how they impact human health.

Currently, I work as a postdoctoral researcher at the Section for Cognitive Systems at the Department of Applied Mathematics and Computer Science, Technical University of Denmark, through a LF Postdoc grant from the Lundbeck Foundation.

I also provide consulting services to academic and industry research groups, which include data processing and machine learning pipelines, algorithm development, and data visualization.

In 2020, I graduated with a PhD in Biomedical Engineering from the Technical University of Denmark under the supervision of the late Associate Professor Helge B. D. Sørensen, PhD; Professor Poul Jennum, MD, PhD, from the Danish Center for Sleep Medicine; and Professor Emmanuel Mignot, MD, PhD, from Stanford University.

Interests

Deep learning
Computational sleep science
Biomedical signal processing

Education

PhD in Biomedical Engineering
Technical University of Denmark
MScEng in Biomedical Engineering
Technical University of Denmark
BScEng in Biomedical Engineering
Technical University of Denmark

Featured Publications

CSLP-AE: A Contrastive Split-Latent Permutation Autoencoder Framework for Zero-Shot Electroencephalography Signal Conversion

Dec 9, 2023

MSED: A Multi-Modal Sleep Event Detection Model for Clinical Sleep Analysis

Mar 3, 2023

Automatic sleep stage classification with deep residual networks in a mixed-cohort setting

Manual annotation of polysomnography studies is subject to human bias with multiple studies showing variations in how sleep experts score sleep. Most research in automatic sleep stage classification models use small-scale data from a single origin, and it is unknown how these models generalize to new data. We developed an algorithm for automatic scoring of sleep stages using raw polysomnography data and obtain state-of-the-art classification performance on a large number of test subjects. Our algorithm was tested under different conditions to compare generalizability. We found that using data from many different sources improves classification performance, and that models trained on single-origin data generalize inconsistently to new data. Future researchers should take multiple datasets into account when developing sleep scoring models.

Jan 1, 2021

Neural Network Analysis of Sleep Stages Enables Efficient Diagnosis of Narcolepsy

Dec 6, 2018

Recent Publications

Javier García Ciudad, Morten Mørup, Birgitte Rahbek Kornum, Alexander Neergaard Zahid (2024). Evaluating the Influence of Temporal Context on Automatic Mouse Sleep Staging through the Application of Human Models. IEEE EMBC.

Cite arXiv

Anders Nørskov, Alexander Neergaard Zahid, Morten Mørup (2023). CSLP-AE: A Contrastive Split-Latent Permutation Autoencoder Framework for Zero-Shot Electroencephalography Signal Conversion. NeurIPS.

PDF Cite Code arXiv

Alexander Neergaard Zahid, Poul Jennum, Emmanuel Mignot, Helge B. D. Sørensen (2023). MSED: A Multi-Modal Sleep Event Detection Model for Clinical Sleep Analysis. IEEE TBME.

PDF Cite DOI arXiv

Alexander Neergaard Zahid, Poul Jørgen Jennum, Emmanuel Mignot, Helge Bjarup Dissing Sørensen (2021). Automatic sleep stage classification with deep residual networks in a mixed-cohort setting. Sleep.

PDF Cite Code DOI arXiv

Aditya Ambati, Yo-El Ju, Ling Lin, Alexander Neergaard Zahid, Henriette Koch, Julien Jacques Hedou, Eileen B. Leary, Vicente Peris Sempere, Emmanuel Mignot, Shahrad Taheri (2020). Proteomic biomarkers of sleep apnea. Sleep.

Cite DOI

See all