Showing 1–6 of 6 results for author: Baumert, M

L-SeqSleepNet: Whole-cycle Long Sequence Modelling for Automatic Sleep Staging

Authors: Huy Phan, Kristian P. Lorenzen, Elisabeth Heremans, Oliver Y. Chén, Minh C. Tran, Philipp Koch, Alfred Mertins, Mathias Baumert, Kaare Mikkelsen, Maarten De Vos

Abstract: Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when develo** sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of… ▽ More Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when develo** sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of-the-art deep learning models are inefficient for that purpose. We thus introduce a method for efficient long sequence modelling and propose a new deep learning model, L-SeqSleepNet, which takes into account whole-cycle sleep information for sleep staging. Evaluating L-SeqSleepNet on four distinct databases of various sizes, we demonstrate state-of-the-art performance obtained by the model over three different EEG setups, including scalp EEG in conventional Polysomnography (PSG), in-ear EEG, and around-the-ear EEG (cEEGrid), even with a single EEG channel input. Our analyses also show that L-SeqSleepNet is able to alleviate the predominance of N2 sleep (the major class in terms of classification) to bring down errors in other sleep stages. Moreover the network becomes much more robust, meaning that for all subjects where the baseline method had exceptionally poor performance, their performance are improved significantly. Finally, the computation time only grows at a sub-linear rate when the sequence length increases. △ Less

Submitted 4 August, 2023; v1 submitted 9 January, 2023; originally announced January 2023.

Comments: This article has been published in IEEE Journal of Biomedical and Health Informatics (JBHI). Source code is available at http://github.com/pquochuy/l-seqsleepnet

Pediatric Automatic Sleep Staging: A comparative study of state-of-the-art deep learning methods

Authors: Huy Phan, Alfred Mertins, Mathias Baumert

Abstract: Background: Despite the tremendous progress recently made towards automatic sleep staging in adults, it is currently unknown if the most advanced algorithms generalize to the pediatric population, which displays distinctive characteristics in overnight polysomnography (PSG). Methods: To answer the question, in this work, we conduct a large-scale comparative study on the state-of-the-art deep learn… ▽ More Background: Despite the tremendous progress recently made towards automatic sleep staging in adults, it is currently unknown if the most advanced algorithms generalize to the pediatric population, which displays distinctive characteristics in overnight polysomnography (PSG). Methods: To answer the question, in this work, we conduct a large-scale comparative study on the state-of-the-art deep learning methods for pediatric automatic sleep staging. Six different deep neural networks with diverging features are adopted to evaluate a sample of more than 1,200 children across a wide spectrum of obstructive sleep apnea (OSA) severity. Results: Our experimental results show that the individual performance of automated pediatric sleep stagers when evaluated on new subjects is equivalent to the expert-level one reported on adults. Combining the six stagers into ensemble models further boosts the staging accuracy, reaching an overall accuracy of 88.8%, a Cohen's kappa of 0.852, and a macro F1-score of 85.8%. At the same time, the ensemble models lead to reduced predictive uncertainty. The results also show that the studied algorithms and their ensembles are robust to concept drift when the training and test data were recorded seven months apart and after clinical intervention. Conclusion: However, we show that the improvements in the staging performance are not necessarily clinically significant although the ensemble models lead to more favorable clinical measures than the six standalone models. Significance: Detailed analyses further demonstrate "almost perfect" agreement between the automatic stagers to one another and their similar patterns on the staging errors, suggesting little room for improvement. △ Less

Submitted 10 May, 2022; v1 submitted 23 August, 2021; originally announced August 2021.

Comments: This article has been published in IEEE Transactions on Biomedical Engineering

Progress in Brain Computer Interfaces: Challenges and Trends

Authors: Simanto Saha, Khondaker A. Mamun, Khawza Ahmed, Raqibul Mostafa, Ganesh R. Naik, Ahsan Khandoker, Sam Darvishi, Mathias Baumert

Abstract: Brain computer interfaces (BCI) provide a direct communication link between the brain and a computer or other external devices. They offer an extended degree of freedom either by strengthening or by substituting human peripheral working capacity and have potential applications in various fields such as rehabilitation, affective computing, robotics, gaming and artificial intelligence. Significant r… ▽ More Brain computer interfaces (BCI) provide a direct communication link between the brain and a computer or other external devices. They offer an extended degree of freedom either by strengthening or by substituting human peripheral working capacity and have potential applications in various fields such as rehabilitation, affective computing, robotics, gaming and artificial intelligence. Significant research efforts on a global scale have delivered common platforms for technology standardization and help tackle highly complex and nonlinear brain dynamics and related feature extraction and classification challenges. Psycho-neurophysiological phenomena and their impact on brain signals impose another challenge for BCI researchers to transform the technology from laboratory experiments to plug-and-play daily life. This review summarizes progress in BCI field and highlights critical challenges. △ Less

Submitted 10 January, 2019; originally announced January 2019.

Observation of Stable Jones-Roberts Solitons in Bose-Einstein Condensates

Authors: Nadine Meyer, Harry Proud, Marisa Perea-Ortiz, Charlotte O'Neale, M. Baumert, M. Holynski, Jochen Kronjäger, Giovanni Barontini, Kai Bongs

Abstract: We experimentally generate two-dimensional Jones-Roberts solitons in a three-dimensional atomic Bose-Einstein condensate by imprinting a triangular phase pattern. By monitoring their dynamics we observe that this kind of solitary waves are resistant to both dynamic (snaking) and thermodynamic instabilities, that usually are known to strongly limit the lifetime of dark plane solitons in dimensions… ▽ More We experimentally generate two-dimensional Jones-Roberts solitons in a three-dimensional atomic Bose-Einstein condensate by imprinting a triangular phase pattern. By monitoring their dynamics we observe that this kind of solitary waves are resistant to both dynamic (snaking) and thermodynamic instabilities, that usually are known to strongly limit the lifetime of dark plane solitons in dimensions higher than one. We additionally find signatures of a possible dipole-like interaction between them. Our results confirm that Jones-Roberts solitons are stable solutions of the non-linear Schrödinger equation in higher dimensions and promote these excitations for applications beyond matter wave physics, like energy and information transport in noisy and inhomogeneous environments. △ Less

Submitted 3 May, 2017; v1 submitted 27 September, 2016; originally announced September 2016.

Journal ref: Phys. Rev. Lett. 119, 150403 (2017)

Oscillations and interactions of dark and dark-bright solitons in Bose-Einstein condensates

Authors: C. Becker, S. Stellmer, P. Soltan-Panahi, S. Dörscher, M. Baumert, E. -M. Richter, J. Kronjäger, K. Bongs, K. Sengstock

Abstract: Solitons are among the most distinguishing fundamental excitations in a wide range of non-linear systems such as water in narrow channels, high speed optical communication, molecular biology and astrophysics. Stabilized by a balance between spreading and focusing, solitons are wavepackets, which share some exceptional generic features like form-stability and particle-like properties. Ultra-cold… ▽ More Solitons are among the most distinguishing fundamental excitations in a wide range of non-linear systems such as water in narrow channels, high speed optical communication, molecular biology and astrophysics. Stabilized by a balance between spreading and focusing, solitons are wavepackets, which share some exceptional generic features like form-stability and particle-like properties. Ultra-cold quantum gases represent very pure and well-controlled non-linear systems, therefore offering unique possibilities to study soliton dynamics. Here we report on the first observation of long-lived dark and dark-bright solitons with lifetimes of up to several seconds as well as their dynamics in highly stable optically trapped $^{87}$Rb Bose-Einstein condensates. In particular, our detailed studies of dark and dark-bright soliton oscillations reveal the particle-like nature of these collective excitations for the first time. In addition, we discuss the collision between these two types of solitary excitations in Bose-Einstein condensates. △ Less

Submitted 3 April, 2008; originally announced April 2008.

Comments: 9 pages, 4 figures

Scaling graphs of heart rate time series in athletes demonstrate the VLF, LF and HF regions

Authors: M Baumert, LM Brechtel, J Lock, A Voss, D Abbott

Abstract: Scaling analysis of heart rate time series has emerged as an useful tool for assessment of autonomic cardiac control. We investigate the heart rate time series of ten athletes (five males and five females), by applying detrended fluctuation analysis (DFA). High resolution ECGs are recorded under standardized resting conditions over 30 minutes and subsequently heart rate time series are extracted… ▽ More Scaling analysis of heart rate time series has emerged as an useful tool for assessment of autonomic cardiac control. We investigate the heart rate time series of ten athletes (five males and five females), by applying detrended fluctuation analysis (DFA). High resolution ECGs are recorded under standardized resting conditions over 30 minutes and subsequently heart rate time series are extracted and artefacts filtered. We find three distinct regions of scale-invariance, which correspond to the well-known VLF, LF, and HF bands in the power spectra of heart rate variability. The scaling exponents alpha are alphaHF: 1.15 [0.96-1.22], alphaLF: 0.68 [0.57-0.84], alphaVLF: 0.83[0.82-0.99]; p<10^-5). In conclusion, DFA scaling exponents of heart rate time series should be fitted to the VLF, LF, and HF ranges, respectively. △ Less

Submitted 21 July, 2006; originally announced July 2006.

Journal ref: 2006 Physiol. Meas. 27 N35-N39