Fig. 1: Experimental task and neural conflict-induced upregulation responses (CI > II).

a Example stimuli illustrating all four possible face/word combinations in the emotional-stroop task. Face stimuli used in our experiment were identical to the face stimuli used in Etkin et al. 2006. For illustrative purposes, we have replaced these images here with open access face stimuli (https://faces.mpdl.mpg.de/imeji/). Participants were instructed to react to the facial expression while ignoring the overlaid word and to answer as fast and accurately as possible. On each trial, the word color was randomly assigned in order to avoid adaptation effects. b Trial presentation schedule. A CI-trial is an incongruent trial preceded by a congruent trial. An II-trial is an incongruent trial preceded by an incongruent trial. Subtracting neural responses for II from CI trials reveals regions involved in the upregulation response (CI > II), while subtracting neural responses for CI from II trials reveals regions associated with implicit conflict adaptation (II > CI). See Supplemental Methods for details on stimulus presentation and counterbalancing of conditions. c Cortical and subcortical regions involved in generating an upregulation response to resolve conflict. Mid-saggital slice with activation clusters shows higher activity to incongruent trials preceded by a congruent trial (CI) as compared to incongruent trials preceded by an incongruent trial (II) (left superior temporal cortex (STC), posterior cingulate cortex (PCC), anterior visual cortex and a large subcortical cluster, FWE-cluster-correction at p = 0.05 with cluster-forming-threshold at p = 0.001, One-sample t test, one-sided, the pseudo-color-map illustrating the one-sample t statistic applies to all panels). Inset shows magnified lateral-view of subcortical cluster and an overlaid locus coeruleus mask in green (2SD-mask from Keren et al.⁸⁸). d Coronal view of standard brain and magnified view of bilateral LC upregulation response (hot colors) overlaid with LC mask (green). e Participants (N = 48) with high subsequent anxiety/depression symptom changes show significantly stronger LC-NE responsivity (CI > II) than participants with lower symptom changes (median split), two-sample t test, two-sided, anxiety: p = 0.019; T = 2.431, depression: p = 0.037; T = 2.154. Bar plots show the LC-responsivity strength extracted from the CI > II contrast in the physiological noise controlled, unsmoothed data as weighted-average of LC-1SD mask voxels (see Supplemental Fig. S8 for detailed statistics and comparison with LC-2SD mask voxels). Single dots show individual data. Errorbars represent ±SEM. *p < 0.05. Source data are provided as a Source Data file.