Automatic staging of infant sleep using novel mathematical analysis of breathing data

Terrill, P. I., Wilson, S. J., Suresh, S., Cooper, D. and Dakin, C. (2010). Automatic staging of infant sleep using novel mathematical analysis of breathing data. In: Sleep Down Under 2010: Biodiversity of Sleep. Australian Sleep Association and Australian Sleep Technologists Association 22nd Annual Scientific Meeting. Sleep DownUnder 2010: Biodiversity of Sleep. Australian Sleep Association and Australian Sleep Technologists Association 22nd Annual Scientific Meeting (ASM 2010), Christchurch, N.Z., (A21-A21). 21-23 October 2010. doi:10.1111/j.1479-8425.2010.00457_3.x


Author Terrill, P. I.
Wilson, S. J.
Suresh, S.
Cooper, D.
Dakin, C.
Title of paper Automatic staging of infant sleep using novel mathematical analysis of breathing data
Conference name Sleep DownUnder 2010: Biodiversity of Sleep. Australian Sleep Association and Australian Sleep Technologists Association 22nd Annual Scientific Meeting (ASM 2010)
Conference location Christchurch, N.Z.
Conference dates 21-23 October 2010
Proceedings title Sleep Down Under 2010: Biodiversity of Sleep. Australian Sleep Association and Australian Sleep Technologists Association 22nd Annual Scientific Meeting   Check publisher's open access policy
Journal name Sleep and Biological Rhythms   Check publisher's open access policy
Place of Publication Richmond, VIC, Australia
Publisher Wiley-Blackwell Publishing Asia
Publication Year 2010
Sub-type Published abstract
DOI 10.1111/j.1479-8425.2010.00457_3.x
ISSN 1446-9235
1479-8425
Volume 8
Issue Supp.1
Start page A21
End page A21
Total pages 1
Collection year 2011
Language eng
Formatted Abstract/Summary
Introduction: In previous work we have described the use of the nonlinear analysis tool, recurrence quantification analysis (RQA) of infant inter-breath interval (IBI) patterns, to discriminate between the infant sleep states, REM (R) and not REM (N). This discrimination tool has then been used to automatically classify 30 s epochs of infant sleep as R or N. For practical application however, it is necessary to also be able to stage wake from sleep. In this work, we propose the use of linear discriminant analysis (LDA) with features derived from statistical quantifications and RQA to classify Wake (W), R and N.

Results: When age specific classifiers were trained, the mean agreement rate at 3 months was 88% (Range: 80–96%) and the mean agreement rate at 6 months was 86% (Range: 64–94%). When data was pooled, and an age generalised classifier was trained, the mean agreement rate was 86% (Range: 72–96%).

Discussion: The performance of this automated sleep state classification tool corresponds well with inter-human-scorer agreement rates, and outperforms respiratory or cardio-respiratory only automated sleep staging systems documented in literature. The requirement for only respiratory data means that this tool has potential application in remote or minimal channel studies, to augment standard sleep lab practice, or in epidemiological sleep research. Further research will investigate performance in older age groups, and in disease states.
© 2010 The Authors
Sleep and Biological Rhythms © 2010 Japanese Society of Sleep Research
Q-Index Code EX
Q-Index Status Confirmed Code
Institutional Status UQ
Additional Notes Published October 2010. Authors prepress title: "Application of linear discriminant analysis (LDA) to features from recurrence quantification analysis (RQA) and statistical analysis of respiratory data to automatically stage infant sleep". Presented during Paediatrics Concurrent Oral Presentations as Presentation no. OP10.

 
Versions
Version Filter Type
Citation counts: Google Scholar Search Google Scholar
Created: Fri, 11 Mar 2011, 12:25:32 EST by Dr David Mason on behalf of School of Information Technol and Elec Engineering