Multimodal assessment of acute stress dynamics using an aversive video paradigm (AVP)

Figures

• 

  Fig. 1. Experimental procedures. After arrival of the participants, the EEG cap was prepared. Then RS EEG was conducted (R.S.), with 2 min Eyes Open (EO) and 2 min Eyes Closed (EC) EEG. Clips 1–2, and Clips 3–4 denote the movie clips shown (aversive or neutral, according to the control or stress session). Blue bars show the time points of saliva (cortisol, cytokines) sampling and red bars show time points when STAI and PANAS were administered. The time points for saliva sampling and questionnaire conduction in relation to the start of the intervention (clip presentation) are shown below the bar in mins. HR (using bipolar electrodes positioned on chest) and EEG were recorded during the whole experiment starting from R.S.1 to R.S.3. The procedure was identical for control and stress sessions except for the kind of movie clips shown. The latter part of the experiment, including tDCS and working memory task performance, is shown in opaque color and will be reported in detail elsewhere as it is out of scope of the topic of this article.

• 

  Fig. 2. Behavioural and physiological data during stress exposure. a.) Positive affect scores during the experiment. b.) Negative affect scores during the experiment. c.) STAI-S scores during the experiment. For these subjective scores, the y axis denotes the scores in the questionnaire and the x axis shows the time course relative to the start of the movie clips. The last time point of subjective data refers to the time point after conduction of the N-back task, which is out of scope of this paper, and these changes might be also influenced by the task. d.) For cortisol levels, the y axis displays mean cortisol levels (nm/l), and the x axis shows the time course relative to the start of the movie clips. Error bars denote ± SEM. Asterisks (*) denote significant differences (critical p-value ≤0.05) between intervention conditions for each time point. The hash symbol (#) denotes significant differences (critical p-value ≤0.05) between time points during intervention and baseline (time point 1) in each condition. The respective color of (#) indicates the respective intervention condition. The last two time points of the cortisol measures were conducted after N-back task performance, which…

• 

  Fig. 3. Cytokine levels during the experiment. The Y axis denotes cytokine concentration levels (pg/ml) and the x axis denotes time relative to onset of the movie clips. Hash symbols (#) denote significant differences (critical p-value ≤0.05) between baseline (time point 1) and the other time points within each condition. The respective color of (#) denotes the respective condition r. Error bars show ± SEM. The last two time points were obtained are after N-back task performance, and the dynamics of these time points might be influenced also by the task.

• 

  Fig. 4. Cardiac data obtained during the experiment. a.) Heart rate during the experiment. The y axis denotes changes of heart rate from baseline, and the x axis denotes each minute of the experiment. Error bars denote ± SEM. Asterisks (*) denote significant differences (critical p-value ≤0.05) between conditions for each time point. The hash symbol (#) denotes significant differences (critical p-value ≤0.05) between time points and baseline (time point 1) in each condition. The respective color of (#) denotes the intervention condition. RS and clip time points are highlighted in the background with grey and orange colored watermarks (grey represents RS and orange clips clip presentation). b.) Heart rate variablility during the experiment. HRV was calculated based on RMSSD measures and plotted the same way as heart rate data.

• 

  Fig. 5. EEG power changes for each clip and all clips combined across the frequency bands of interest: Theta (4–8 Hz), Alpha (8–13 Hz), low Beta (13–15 Hz), high Beta (23–32 Hz), low Gamma (32–50 Hz), and high Gamma (50–80 Hz). Each topographical plot illustrates the mean difference (stress-control) between conditions. Red indicates a higher value, while blue indicates a lower value in the stress condition compared to the control intervention. To identify significant electrodes in each analysis, we employed Monte Carlo-based permutation statistics with cluster correction (see data analysis section). This permutation strategy involves permuting data to establish a data-driven null distribution. To address the issue of multiple comparisons, clusters (based on adjacency) were identified and compared with the null distribution using a critical p-value of < 0.01. Each EEG topographical plot features a cartoon head viewed from above, with the nasion at the top and the inion at the bottom. The color bar represents the range of difference values from 2 to −2. For a reference to a zoomed-in cartoon head with highlighted electrode locations and labels, please refer to supplementary figure 2.

• 

  Fig. 6. EEG connectivity differences between stress and control conditions are shown for each clip separately, as well as for all clips combined for each frequency band of interest (as in Fig. 5). Each topographical plot displays the significant connections (which survived after statistical thresholding) marked in two colors. Blue indicates a decrease in connectivity strength, while red indicates an increase in connectivity strength in the stress condition compared to the control condition. To identify the significant connections between the two conditions, the entire 62 × 62 connectivity matrix was analyzed using Monte Carlo-based permutation statistics with maximum statistics correction (see data analysis section), which identifies significant connections based on a distribution created by permuted data. The differences were marked as significant for a critical p-value of ≤0.05. The topographical plot illustrates significant connections using red or blue lines between electrode pairs. The cartoon head is depicted from above, with the nasion positioned at the top and the inion at the bottom. For a reference to a zoomed-in cartoon head with highlighted electrode locations and labels,…

• 

  Fig. 7. a.) EEG power changes during the RS for both EO and EC states. The topographical plot over the cartoon head (plotted similar to Fig. 5) displays mean power differences before and after intervention for EO and EC states, for stress and control conditions. Significant electrodes after corrections for multiple comparisons are marked in bold (as in Fig. 5). The color bar denotes the difference value range from 3 to −3. Red indicates higher power values in the post-intervention RS compared with pre-intervention, and blue indicates reduced power. b.) EEG connectivity differences during RS before and after intervention for both EO and EC states for control and stress conditions. Significant connections surviving multiple comparisons (with statistics as shown in Fig. 6) are marked in blue and red. Blue indicates a decrease of connectivity strength in the post-intervention RS (after clip presentation), and red denotes an increase of connectivity strength in the post-, as compared to the pre-intervention RS. The topographical plot shows significant connections on the cartoon head similar to Fig. 6. For a reference to a zoomed-in cartoon head with highlighted electrode locations and labels,…

• 

  Fig. 8. EEG preprocessing pipeline used for all datasets (movie clips and RS). We used EEGLAB (Delorme & Makeig, 2004) and custom MATLAB scripts for the analysis. Please refer to supplementary Table 3 for the average number of epochs and channels after rejection for each clip and RS (EO and EC) recordings.