Capturing inter-individual variability in stress dynamics with heart rate traces reveals activity in the bilateral hippocampus, amygdala, and insula

Figures

• 

  Fig. 1.. Illustration of the study purpose: refined capturing of the individual stress wave by incorporation of heart rate traces. Three example heart rates traces (brown, pink, green) throughout the whole stress experiment are illustrated together with the classical box-car regressors (grey) contrasting each StressPhase uniformely against the Pre and PostStress phases without individual weighting. The differences among the individual’s rate traces illustrate the high interindividual variability and stress dynamic. This figure has been modified withBioRender.com.

• 

  Fig. 2.. Schematic illustration of the task. The task was imbedded into the second day of the beCOME study and consisted of three different phases (PreStress, Stress, and PostStress) with interleaved blocks of active calculations (60 s) and rest (40 s). East task phase consisted of 5 blocks; 3 rest blocks verbal instructions or negative verbal feedback were presented. PPG, ECG, and skin conductance were measured simultaneously; saliva cortisol was collected at four different times (in minutes) throughout the task. Figure adapted fromKühnel et al. (2020)usingBioRender.com.

• 

  Fig. 3.. Regressors for the model with blockwise pulse rates as parametric modulator and for the control analysis. Dark grey boxes represent box-car regressors for active task blocks (either regardless of the phase or for task phases sepereted), and blue/orange boxes represent either blockwise or volumewise heart rates/high-frequency power as a parametric modulator of the box-car regressors. This figure has been modified withBioRender.com.

• 

  Fig. 4.. Heart rate traces across the task. Solid black line shows the mean heart rate changes across the task; red dots= active calculation blocks, black dots= rest blocks; grey lines show individual heart rates; and all data have been normalized to the mean heart rate during rest in the PreStress phase. Thus, values >1 represent heart rate increases relative to the rest in the PreStress phase and <1 vice versa. This figure has been modified withBioRender.com.

• 

  Fig. 5.. Correlation of the heart rate up- and downswings with changes in self-reported physical arousal and relaxation from the BSKE. Left: correlations with mean heart rate upswing from the PreStress to Stress phase with changes in physical arousal and relaxation; Right: same but with proportional heart rate downswing (normalized to the participant’s upswing); r = correlation coefficient, p = uncorrected p-value (after correcting for multiple comparisons with the number of effective tests calculated with the R-based poolr package the p-threshold for significance was 0.01). This figure has been modified withBioRender.com.

• 

  Fig. 6.. Main results of the model incorporating individual mean heart rates. Depicted are neural correlates with a parametric modulator across all 15 active blocks of individual mean heart rates. The map is thresholded at pvoxel.FWE<0.05 and cluster extent >25. Red = positive correlation with heart rate. Blue = negative correlation. This figure has been modified withBioRender.com.

• 

  Fig. 7.. Mean z-transformed regional time-courses across blocks per task phase. Depicted are Mean z—transformed first principal components for left hippocampus, left amygdale, and left insula for all three task phases (blue =PreStress, red =Stress, yellow =PostStress) averaged across all active and rest blocks of each respective task phase and 95% confidence intervals (shaded areas). Similar results were obtained also for the right hippocampus, right amygdala, right insula (not shown). This figure has been modified withBioRender.com.