Heart Rate Variability to Automatically Identify Hyperbaric States Considering Respiratory Component

Figures

• 

  Figure 1. Estimation of the up slope, down slope and R-wave angle, as extracted from [24]. The thick magenta lines represent the two straight lines from which the slope series were obtained. The R-wave angle series were obtained from the smallest angle formed by these two lines.

• 

  Figure 2. An example of the three EDR signals extracted from one ECG lead.

• 

  Figure 3. Boxplots of all the parameters of the ECG signal. The significance level α of the t-Student or Wilcoxon test was indicated with ★ for α=0.05 , ⋄ for α=0.01 , and † for α=0.001 . The arrows indicate the statistically significant differences between the compared groups, and this was achieved using ANOVA or Friedman statistics with the Bonferroni correction for multiple comparisons tests when the median value of one stage (arrow start) was significantly higher or lower than the other (arrow end). The colors of the start and end of the arrows indicate the stages analyzed.

• 

  Figure 4. Accuracy of the subject stage identification as the number of selected features rose in each of the four classifiers used. The 70% accuracy was highlighted in the images with a gray horizontal line. (a) Stages 3D and 5 (C.3D-5). (b) Stages 5 and 3A (C.5-3A). (c) Stages 5 and 1A (C.5-1A). (d) Stages 3A and 1A (C.3A-1A). (e) Stages 3D , 5, 3A , and 1A (C.3DA-5-1A).

• 

  Figure 5. Number of times each subject was misclassified. The caption of each panel indicates the number of cases considered, i.e., those whose accuracy exceeded 70% and those multiplied by the two classes considered in each classifier. The maximum possible number of misclassifications considered in each panel is highlighted in the graphs with a horizontal line in the color associated with each set of classes. (a) C.5-1A: 66 cases. C.5-3A: 70 cases. (b) C.3A-1A: 8 cases. C.3D-5: 6 cases.