On the relevance of symbolizing heart rate variability by means of a percentile-based coarse graining approach

Dirk Cysarz, Friedrich Edelhauser, Michal Javorka, Nicola Montano, Alberto Porta

Research output: Contribution to journalArticle

Abstract

Objective: Symbolic transformations of the cardiac interbeat interval series yield a coarse-grained description of the dynamical information of the underlying system and complement standard measures of heart rate variability. The most commonly utilized coarse graining procedures are strongly influenced by the presence of a few extreme values wasting precious symbols to code very unlikely values. Approach: Here, we used a transformation procedure that ensured the appearance of each symbol with equal probability using a short alphabet, A 4 = {0, 1, 2, 3}, and a long alphabet, A 6 = {0, 1, 2, 3, 4, 5}. The procedure was applied to the cardiac interbeat interval series of 17 healthy subjects, obtained during graded head-up tilt tests at tilt table inclinations of 0°, 15°, 30°, 45°, 60°, 75°, 90°. The dynamics of the symbolic series was assessed by the rate of symbolic pattern categories. Symbolic patterns of length three were grouped according to the variations of the symbols in each pattern: no variation (0V%), one variation (1V%), two like variations (2LV%) and two unlike variations (2UV%) of the symbols. Main results: As for the alphabet A 4, the linear regression analysis on tilt angle showed that 0V% increased with increasing tilt angle whereas 1V%, 2LV% and 2UV% decreased. As for the alphabet A 6, the categories 0V%, and 1V% increased with increasing tilt angle whereas 2LV% and 2UV% decreased. Significance: The symbolic transformation ensuring a uniform distribution of the symbols is capable of reflecting changes in the cardiac autonomic nervous system during graded head-up tilt. This approach is more robust against outliers and data with skewed distributions compared to previously used symbolizations.

Original languageEnglish
Article number105010
JournalPhysiological Measurement
Volume39
Issue number10
DOIs
Publication statusPublished - Oct 24 2018

Fingerprint

Neurology
Linear regression
Regression analysis
Heart Rate
Tilt-Table Test
Head
Autonomic Nervous System
Information Systems
Linear Models
Healthy Volunteers
Regression Analysis

ASJC Scopus subject areas

  • Biophysics
  • Physiology
  • Biomedical Engineering
  • Physiology (medical)

Cite this

On the relevance of symbolizing heart rate variability by means of a percentile-based coarse graining approach. / Cysarz, Dirk; Edelhauser, Friedrich; Javorka, Michal; Montano, Nicola; Porta, Alberto.

In: Physiological Measurement, Vol. 39, No. 10, 105010, 24.10.2018.

Research output: Contribution to journalArticle

@article{68bcef5ddae4421080203f92625a25e3,
title = "On the relevance of symbolizing heart rate variability by means of a percentile-based coarse graining approach",
abstract = "Objective: Symbolic transformations of the cardiac interbeat interval series yield a coarse-grained description of the dynamical information of the underlying system and complement standard measures of heart rate variability. The most commonly utilized coarse graining procedures are strongly influenced by the presence of a few extreme values wasting precious symbols to code very unlikely values. Approach: Here, we used a transformation procedure that ensured the appearance of each symbol with equal probability using a short alphabet, A 4 = {0, 1, 2, 3}, and a long alphabet, A 6 = {0, 1, 2, 3, 4, 5}. The procedure was applied to the cardiac interbeat interval series of 17 healthy subjects, obtained during graded head-up tilt tests at tilt table inclinations of 0°, 15°, 30°, 45°, 60°, 75°, 90°. The dynamics of the symbolic series was assessed by the rate of symbolic pattern categories. Symbolic patterns of length three were grouped according to the variations of the symbols in each pattern: no variation (0V{\%}), one variation (1V{\%}), two like variations (2LV{\%}) and two unlike variations (2UV{\%}) of the symbols. Main results: As for the alphabet A 4, the linear regression analysis on tilt angle showed that 0V{\%} increased with increasing tilt angle whereas 1V{\%}, 2LV{\%} and 2UV{\%} decreased. As for the alphabet A 6, the categories 0V{\%}, and 1V{\%} increased with increasing tilt angle whereas 2LV{\%} and 2UV{\%} decreased. Significance: The symbolic transformation ensuring a uniform distribution of the symbols is capable of reflecting changes in the cardiac autonomic nervous system during graded head-up tilt. This approach is more robust against outliers and data with skewed distributions compared to previously used symbolizations.",
author = "Dirk Cysarz and Friedrich Edelhauser and Michal Javorka and Nicola Montano and Alberto Porta",
year = "2018",
month = "10",
day = "24",
doi = "10.1088/1361-6579/aae302",
language = "English",
volume = "39",
journal = "Physiological Measurement",
issn = "0967-3334",
publisher = "Institute of Physics Publishing",
number = "10",

}

TY - JOUR

T1 - On the relevance of symbolizing heart rate variability by means of a percentile-based coarse graining approach

AU - Cysarz, Dirk

AU - Edelhauser, Friedrich

AU - Javorka, Michal

AU - Montano, Nicola

AU - Porta, Alberto

PY - 2018/10/24

Y1 - 2018/10/24

N2 - Objective: Symbolic transformations of the cardiac interbeat interval series yield a coarse-grained description of the dynamical information of the underlying system and complement standard measures of heart rate variability. The most commonly utilized coarse graining procedures are strongly influenced by the presence of a few extreme values wasting precious symbols to code very unlikely values. Approach: Here, we used a transformation procedure that ensured the appearance of each symbol with equal probability using a short alphabet, A 4 = {0, 1, 2, 3}, and a long alphabet, A 6 = {0, 1, 2, 3, 4, 5}. The procedure was applied to the cardiac interbeat interval series of 17 healthy subjects, obtained during graded head-up tilt tests at tilt table inclinations of 0°, 15°, 30°, 45°, 60°, 75°, 90°. The dynamics of the symbolic series was assessed by the rate of symbolic pattern categories. Symbolic patterns of length three were grouped according to the variations of the symbols in each pattern: no variation (0V%), one variation (1V%), two like variations (2LV%) and two unlike variations (2UV%) of the symbols. Main results: As for the alphabet A 4, the linear regression analysis on tilt angle showed that 0V% increased with increasing tilt angle whereas 1V%, 2LV% and 2UV% decreased. As for the alphabet A 6, the categories 0V%, and 1V% increased with increasing tilt angle whereas 2LV% and 2UV% decreased. Significance: The symbolic transformation ensuring a uniform distribution of the symbols is capable of reflecting changes in the cardiac autonomic nervous system during graded head-up tilt. This approach is more robust against outliers and data with skewed distributions compared to previously used symbolizations.

AB - Objective: Symbolic transformations of the cardiac interbeat interval series yield a coarse-grained description of the dynamical information of the underlying system and complement standard measures of heart rate variability. The most commonly utilized coarse graining procedures are strongly influenced by the presence of a few extreme values wasting precious symbols to code very unlikely values. Approach: Here, we used a transformation procedure that ensured the appearance of each symbol with equal probability using a short alphabet, A 4 = {0, 1, 2, 3}, and a long alphabet, A 6 = {0, 1, 2, 3, 4, 5}. The procedure was applied to the cardiac interbeat interval series of 17 healthy subjects, obtained during graded head-up tilt tests at tilt table inclinations of 0°, 15°, 30°, 45°, 60°, 75°, 90°. The dynamics of the symbolic series was assessed by the rate of symbolic pattern categories. Symbolic patterns of length three were grouped according to the variations of the symbols in each pattern: no variation (0V%), one variation (1V%), two like variations (2LV%) and two unlike variations (2UV%) of the symbols. Main results: As for the alphabet A 4, the linear regression analysis on tilt angle showed that 0V% increased with increasing tilt angle whereas 1V%, 2LV% and 2UV% decreased. As for the alphabet A 6, the categories 0V%, and 1V% increased with increasing tilt angle whereas 2LV% and 2UV% decreased. Significance: The symbolic transformation ensuring a uniform distribution of the symbols is capable of reflecting changes in the cardiac autonomic nervous system during graded head-up tilt. This approach is more robust against outliers and data with skewed distributions compared to previously used symbolizations.

UR - http://www.scopus.com/inward/record.url?scp=85055908208&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055908208&partnerID=8YFLogxK

U2 - 10.1088/1361-6579/aae302

DO - 10.1088/1361-6579/aae302

M3 - Article

C2 - 30235164

AN - SCOPUS:85055908208

VL - 39

JO - Physiological Measurement

JF - Physiological Measurement

SN - 0967-3334

IS - 10

M1 - 105010

ER -