<abstract xmlns="http://www.w3.org/1999/xhtml"><p>Sleep can be characterised as a dynamic process that has a finite set of sleep stages during the night. The standard Rechtschaffen and Kales sleep model produces discrete representation of sleep and does not take into account its dynamic structure. In contrast, the continuous sleep representation provided by the probabilistic sleep model accounts for the dynamics of the sleep process. However, analysis of the sleep probabilistic curves is problematic when time misalignment is present. In this study, we highlight the necessity of curve synchronisation before further analysis. Original and in time aligned sleep probabilistic curves were transformed into a finite dimensional vector space, and their ability to predict subjects’ age or daily measures is evaluated. We conclude that curve alignment significantly improves the prediction of the daily measures, especially in the case of the <italic>S2</italic>-related sleep states or slow wave sleep.</p></abstract>

Sleep can be characterised as a dynamic process that has a finite set of sleep stages during the night. The standard Rechtschaffen and Kales sleep model produces discrete representation of sleep and does not take into account its dynamic structure. In contrast, the continuous sleep representation provided by the probabilistic sleep model accounts for the dynamics of the sleep process. However, analysis of the sleep probabilistic curves is problematic when time misalignment is present. In this study, we highlight the necessity of curve synchronisation before further analysis. Original and in time aligned sleep probabilistic curves were transformed into a finite dimensional vector space, and their ability to predict subjects’ age or daily measures is evaluated. We conclude that curve alignment significantly improves the prediction of the daily measures, especially in the case of the S2-related sleep states or slow wave sleep.

Time Alignment as a Necessary Step in the Analysis of Sleep Probabilistic Curves

Institute of Measurement Science, Slovak Academy of Sciences

Faculty of Electrical Engineering and Information Technology

Measurement Science Review

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

{"article-title":"Time Alignment as a Necessary Step in the Analysis of Sleep Probabilistic Curves"}

Sleep can be characterised as a dynamic process that has a finite set of sleep stages during the night. The standard Rechtschaffen and Kales sleep model...

Time Alignment as a Necessary Step in the Analysis of Sleep Probabilistic Curves

Published Online: Mar 07, 2018

Page range: 1 - 6

Received: Sep 09, 2017

Accepted: Feb 12, 2018

DOI: https://doi.org/10.1515/msr-2018-0001

Keywords
Curve alignment, probabilistic sleep model, functional principal component analysis

© 2018 Zuzana Rošt’áková et al., published by De Gruyter Open

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Time Alignment as a Necessary Step in the Analysis of Sleep Probabilistic Curves

Published Online: Mar 07, 2018

Page range: 1 - 6

Received: Sep 09, 2017

Accepted: Feb 12, 2018

DOI: https://doi.org/10.1515/msr-2018-0001

KeywordsCurve alignment, probabilistic sleep model, functional principal component analysis

© 2018 Zuzana Rošt’áková et al., published by De Gruyter Open

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Keywords
Curve alignment, probabilistic sleep model, functional principal component analysis