cluster:datasets:pediatric_mouth_breathing

VSN 18-206

The study cohort consists of Icelandic children between the age of 11 and 14. The children are part of the EuroPrevall/iFAAM (Integrated Approaches to Food Allergen and Allergy Risk Management) birth cohort study, which recruited 12 000 newborns in nine European Countries from 2005 to 2009. As part of this higher-level study, the parents answered a questionnaire with questions regarding the children’s sleeping behaviour, which served as a basis for the selection of participants for this sleep study. The study was approved by the Data Protection Agency of Iceland and includes a written consent by each child’s legal guardian.

The study was conducted at the Landspitali University Hospital in cooperation with the Reykjavik University Sleep Institute. The measurement was performed with a Nox Medical A1 Polysomnography device. This device needs to be set-up for the measurement in the hospital, but allows the participants to sleep at home, making it more convenient for the participant, especially children, in addition to reducing the risk of biased sleep due to an unknown sleeping environment.

The study was simultaneously conducted with a second Self Applied Somnography (SAS) setup, a simplified version of the PSG, which can be applied at home without the help of a sleep technologist. This setup includes an oronasal cannula, which captures the mouth breathing. Each PSG recording is approximately 8 hours long. The measurement of this study includes the following sensors: electroencephalogram (EEG), electrooculogram (EOG), chin and leg electromyogram (EMG), electrocardiogram (ECG), RIP belts for thorax and abdomen, a finger probe pulse oximeter, a microphone, an oronasal cannula, electrodermal activity (EDA) and accelerometry measuring the movement and body position. In total, this results in 84 different signals.

Excerpt of the signals recorded in the dataset:

Signal Sampling Frequency
Thorax RIP 200
Abdomenal RIP 200
Nasal Flow 200

In total, there are 111 recordings, from which 20 have been manually labelled by a sleep technologist who abides to the annotation rules version 2.5 by the American Academy of Sleep Medicine. The sleep recordings are displayed in a scoring system, where different signals can be selected, enlarged, distorted and overlapped. Based on the oral flow and its interaction with the thorax, abdomen, body movement, audio and nasal flow, a sleep technologist manually labels periods with mouth breathing. Through these annotations, a new binary variable, which holds the information whether a sequence contains mouth breathing, can be created, which will be the target variable of the classification. Hence, the classification aims to differentiate between sequences which contain and do not contain mouth breathing.

For each patient, there is one PSG recording, one SAS recording and for the scored patients, a .csv document with the annotations from the sleep technologist. In order to use this data for machine learning, it needs to be loaded, synced and pre- processed.

If you use this dataset, please include a citation.

Plaintext

Holm, B. (2022). Slapp Open Access Dataset. Journal of Very Small Datasets, 1(3), 257–276.
https://doi.org/42.4242/012345678901234567

BibTex

@ARTICLE{slappopenaccess,
  title    = "Slapp Open Access Dataset",
  author   = "Benedikt Holm",
  journal  = "Journal of Very Small Datasets.",
  volume   =  1,
  number   =  1,
  pages    = "257--276",
  month    =  jan,
  year     =  2022,
  language = "en"
}
  • cluster/datasets/pediatric_mouth_breathing.txt
  • Last modified: 2022/03/01 14:21
  • by lukab