The pre-ejection period is a highly stress dependent parameter of paramount importance for pulse-wave-velocity based applications

FIGURE 1. Pre-ejection period determination: Illustration of the pre-ejection period (PEP) as the time delay between an ECG’s Q-wave (upper panel) and the start of blood ejection (increased flow in the aorta, lower panel). The PEP cannot be detected by ECG based measurements of the pulse-wave-velocity.

FIGURE 2. Experiment timeline illustrated by example heart rate profile: Phases of the experiments are indicated by the uppermost arrows. Subsets of experimental phases are indicated by the second row of arrows. An example heart rate profile is displayed in blue. Gray dashed vertical lines indicate the initiation of blood pressure measurements. The measurement instance is denoted to the right of the dashed line. The participant did not reach load-phases E6 + E7 (see Figure 4). The upper panel (A) illustrates the first experimental phase (mental load), the lower panel (B) the second phase (physical load).

FIGURE 3. Individual PEP variability at rest of all subjects. Measurements were obtained in the 60 s before calibration: SD, standard deviation; PEP–PEP at first BP measurement (at rest) depicted for 60 s before the BP measurement. The SD of 4.5 ms symbolizes small intraindividual PEP variability at rest.

FIGURE 4. PEP changes under mental (TSST) and physical (ergometer) load: The upper panel (A) depicts the PEP reactivity during the TSST. “R” marks the measurement at rest while “Q1”–“Q8” demark TSST questions. The lower panel (B) depicts PEP reactivity and load-response dependency under physical stress. “R” marks a measurement at rest, “E1”–“E10” demark measurements under physical load and “B1”–“B6” demark measurements during the break-period. We depicted PEP measurements taken at the end of their according BP measurements. Declining n-numbers are due to dropouts (during TSST) or reaching the maximum heart rate at earlier load steps (ergometer). Whiskers show confidence intervals (p < 0.05) for relative PEP values. The absolute variance of PEP is more clearly depicted in Figure 8.

FIGURE 5. HR, heart rate; LVET, left-ventricular-ejection-time; Correlations of PEP with parameters connected to sympathetic tone. To enhance visibility, panels (A–C) show scatters of 5,000 randomly sampled datapoints while correlation coefficients were calculated for the whole dataset. Panel (B) features an additional polynomial fit (blue) to account for the non-linear data distribution. Panel (D) shows the correlation between PEP and SBP compared to their respective values at rest.

FIGURE 6. Regression between heart rate (HR) and PEP at rest (A), during TSST (B), and treading (C) the bike ergometer.

FIGURE 7. Differences of PEP at same heart rate and its predictive power for current load. Upper panel (A): Differences of PEP at same heart rates at rest and under mental and physical load. To enhance visibility, PEP values are averaged over a heart rate range of 12 bpm. Lower pane (B) l: k-nearest-neighbor predictor (k = 13) of load-state based on heart rate and PEP deviation from measurement at rest. Decision boundaries are marked by colored areas in the plot. Bisecting decision boundaries represent a coequal importance of PEP and HR for predictive power (e.g., solely heart rate based predictor would show vertical decision boundaries). Scattered elements represent patient specific mean values under rest (blue circle), mental (green triangle) and physical load (red square).

FIGURE 8. Interindividual variance of PEP: The upper panel shows a boxplot (A) of individual subject PEP values at rest, at the first question of TSST and the last step of bike ergometer treading. Broad quartiles and long whiskers symbolize a large interindividual variability. The lower panel shows interindividual differences in regression slopes (B) between heart rate and PEP. The black line indicates the dataset mean while the green and red lines represent the flattest and steepest slopes, respectively. * = p < 0.05.

FIGURE 9. Applying PEP to a published PWV-BP relation. Panel (A) depicts the relation, transformed to a PAT/BP relation for a “standard” human (1.80 m). Panel (B) represents the best polynomial fit for our data of PEP and SBP. Panel (C) depicts the resulting PTT-BP relation (PTT = PAT–PEP) when correcting PAT for PEP. The resulting line (blue) is the purest representation of a true PWV-BP relationship. Panel (D) highlights the (on average) constant proportion of PEP in PAT.

FIGURE 10. The upper panel (A) depicts confidence intervals for the PTT-BP relation when estimating PEP. The significant discrepancies are originated from the large PEP variability depicted in Figure 9B. The lower panel (B) highlights the SBP measurement uncertainty induced by estimating PEP with provided confidence intervals.