Fig 1.
Trial cadence for emotion regulation task depicting the instructed conditions (“REAPPRAISE” and “SUPPRESS” cues) and uninstructed condition (“LOOK” cue). Participants were given a cue at the beginning of a block of twenty trials. Participants first saw an original stimulus for 5000 ms. On the subsequent answer screen, participants either saw the identical image again, as depicted in the middle (REAPPRAISE) row, or a slightly altered image, as depicted in the top (LOOK) and bottom (SUPPRESS) rows. With a button press, participants indicated whether the image was identical (“correct”) or had been altered (“incorrect”). Stimuli were not repeated between conditions and were randomly presented within blocks. The order of emotion regulation conditions was randomized. Images from the International Affective Picture System were used in the experiment. However, the above image is not from this database. The photo used is credited to https://litfl.com/clinical-cases/ and was edited and reused here with permission under the Creative Commons Attribution-ShareAlike 3.0 Unported license (https://creativecommons.org/licenses/by/3.0/legalcode).
Fig 2.
Sample patterns of psychophysiological activity and multivariate distance.
Pattern of psychophysiological activity for each regulation condition for a single sample participant (left) and correlation distance among regulation conditions based on sample participant’s unique profile (right). CS = corrugator supercilii, GSR = galvanic skin response (skin conductance) and LL = levator labii. The value corresponding to the psychophysiological signal for each regulation condition (left) is the average activity for that signal across all twenty trials in that regulation condition and participant. Psychophysiological data are within-subject z-scored. Dissimilarity (right) is measured with correlation distance (ranging from 0 to 2, with 0 representing perfect correlation, 1 representing no correlation and 2 representing perfect anti-correlation). Smaller y-axis values represent greater similarity.
Fig 3.
Psychophysiological activity by regulation condition and trait anxiety.
Mean physiological activity for each measure in each emotion regulation condition, grouped by trait anxiety. Although trait anxiety (State-Trait Anxiety Inventory; STAI) is a continuous measure, for visualization purposes it is divided here into three groups corresponding to below one SD below the mean (Low), within one SD of the mean (Middle) and above one SD above the mean (High). All psychophysiological data are within-subject z-scored across all conditions in the study.
Fig 4.
Psychophysiological patterns associated with reappraisal versus suppression.
Logit function from a mixed-effects logistic regression predicting if participants were in the reappraise or suppress condition. Top: significant effect of skin conductance (left), significant effect of corrugator (middle) and non-significant effect of levator activity (right) on the predicted probability of being in the reappraisal condition as opposed to the suppress condition. Bottom: non-significant interaction between skin conductance and trait anxiety (left), significant interaction between corrugator and trait anxiety (middle) and significant interaction between levator and trait anxiety (right) on the predicted probability of being in the reappraisal condition as opposed to the suppress condition. All psychophysiological data are within-subject z-scored.
Fig 5.
Individual differences in multivariate psychophysiological dissimilarity.
Participants’ self-reported regulation tendency did not differentiate how they naturally regulated, but their trait anxiety did. Effects from Models 1, 2 and 3 on multivariate psychophysiological dissimilarity between instructed and uninstructed conditions. Left (Model 1): significant interaction between trait anxiety (STAI) and instructed condition. Middle (Model 2): non-significant interaction between ERQ-Reappraisal and instructed condition. Right (Model 3): non-significant interaction between ERQ-Suppression and instructed condition. Dissimilarity is measured with correlation distance (ranging from 0 to 2, with 0 representing perfect correlation, 1 representing no correlation, and 2 representing perfect anti-correlation). Smaller y-axis values represent greater similarity. ERQ scores are from the Emotion Regulation Questionnaire [36] and trait anxiety scores are from the State Trait Anxiety Inventory-trait Subscale [52].