Asthmatic subjects stratification using autonomic nervous system information

Javier Milagro, Lorena Soto-Retes, Jordi Giner, Carolina Varon, Pablo Laguna, Raquel Bailon, Vicente Plaza, Eduardo Gil

Research output: Contribution to journalArticleScientificpeer-review


Objective: the aim of this study is to evaluate whether noninvasive autonomic activity assessment could represent a potential tool for the stratification of asthmatic subjects based on symptoms control, using only 10-min electrocardiographic and respiratory signals. Methods: several heart rate variability (HRV) derived indexes, which are regarded as surrogates of autonomic activity, were evaluated in a group of asthmatic patients classified based on their symptomatology control. The effect of respiration on HRV was mitigated by means of orthogonal subspace projection. The most relevant features were used for training different classifiers. Results: similar classification performance was obtained when using HRV or clinical features, with just a 10% decrease in accuracy when using the HRV features (80% vs. 70%). This classification performance is equivalent to that achieved in new patients using the current asthma control tests. Conclusion: results suggest that the noninvasive assessment of autonomic activity could represent an added value for the monitoring of asthmatic subjects outside the clinic, using less cumbersome equipment, and therefore being suitable for an objective asthma self-monitoring. Significance: This study provides evidence on the usefulness of noninvasive autonomic activity assessment for asthma control stratification, supporting it as a potential complement to the current clinical practice.

Original languageEnglish
Article number102802
Number of pages8
JournalBiomedical Signal Processing and Control
Publication statusPublished - 2021


  • Asthma
  • Asthma control
  • Autonomic nervous system
  • Heart rate variability
  • Machine learning

Cite this