Neural activation during emotional interference corresponds to emotion dysregulation in stressed teachers

Teaching is among the most stressful professions, with stressors including workload, interpersonal demands, role ambiguity, and emotional labor. Teacher stress and burnout relate to lower self-efficacy and poorer student outcomes. Emotion regulation (awareness/acceptance of emotions, goal-directed behavior under negative affect, and flexible strategy use) is a protective factor, while dysregulation predicts burnout. Theoretical models posit regulation targets across attention, appraisal, and response; attentional selection (disengaging from affective stimuli) and cognitive reappraisal both reduce negative emotion intensity. Burnout is also linked to deficits in executive functions and attention, making cognitive control crucial for effective emotion regulation. Emotional stimuli capture attention, creating interference measured with emotional Stroop paradigms that recruit prefrontal networks (ventrolateral, dorsomedial, dorsolateral PFC, ACC). This study asked: how do stressed teachers process emotional information neurally during an emotional interference task? It tested whether teacher-specific versus general emotional stimuli elicit distinct neural interference patterns, and whether individual differences in emotion dysregulation, emotion recognition, and cognitive control relate to these neural responses.

- Teacher stress/burnout contributes to attrition, reduced wellbeing and student outcomes; sources include workload, interpersonal demands, role conflict/ambiguity, and emotional demands. Resilience and emotion regulation buffer stress. - Emotion regulation frameworks (Gross; Gratz & Roemer) identify attentional selection and cognitive reappraisal as key strategies; deep acting strategies relate to reduced burnout, while surface acting relates to increased burnout. - Job Demands-Resources theory explains burnout via high demands and inadequate resources; burnout relates to cognitive/attentional deficits and maladaptive self-regulation. Empirical work shows deficits in executive functioning, attention, memory, increased cognitive load, and error/negative feedback processing in burnout. - Emotional Stroop tasks index automatic attentional capture by emotional stimuli; clinical populations show enhanced interference to disorder-specific stimuli. Neuroimaging meta-analyses show consistent engagement of prefrontal/ACC networks and, in some cases, amygdala and fusiform during emotional interference. - Gaps: Few studies focus on teachers; unclear how teacher-specific emotional content affects neural processing; need to relate individual differences in emotion dysregulation, recognition, and cognitive control to neural responses during interference.

Design: Event-related fMRI emotional counting Stroop task with multivariate Partial Least Squares (PLS) analyses to identify spatiotemporal neural patterns and brain–behavior relationships. Participants: 54 recruited Australian practicing teachers with self-reported workplace stress/burnout; 5 excluded for missing self-report, final N=49 (3 males), age 25–67 years (Mean 44.77, SD 10.77), mean 18.15 years teaching (SD 10.51). Right-handed; no neurological disease/trauma, no substance use, no regular mindfulness/vigorous exercise; MRI-compatible. Moderate-high perceived stress (PSS M=19.17, SD=4.85); DERS total M=76.90 (SD=18.23). Ethics approved; informed consent obtained. Tasks/Measures: - Self-report: Difficulties in Emotion Regulation Scale (DERS, 36 items; total score used), Perceived Stress Scale (PSS-10). - CANTAB: Attention Switching Task (AST) for cognitive control (side vs direction cue response mapping); Emotion Recognition Task (ERT) with 180 morphed facial emotion stimuli (six emotions), faces presented 200 ms; RT and accuracy recorded. Emotional counting Stroop (fMRI task): Participants pressed a button indicating the number of times (1–4) a word appeared, ignoring meaning. Conditions: General Negative (GenNeg), General Neutral (GenNeut), Teacher-Specific Negative (SpeNeg), Teacher-Specific Neutral (SpeNeut), plus Control (number words congruent with count). Word display 1000 ms; jittered ISI 1000–2000 ms with fixation. Four blocks; pseudorandomized within blocks; counterbalanced block order; instructions before each block. Behavioral analysis: Accuracy (%) and RT computed per condition relative to Control (condition RT – control RT) to adjust for age-related differences; RM-ANOVA with within-subject factor Condition (GenNeg, GenNeut, SpeNeg, SpeNeut); Pearson correlations with years teaching, stress, emotion measures. MRI acquisition: 3T Siemens Magnetom Trio Tim, 32-channel head coil. T1 MP2-RAGE: 176 sagittal slices, 1 mm isotropic, TR=4000 ms, TE=2.89 ms, FOV=256 mm. fMRI T2*-EPI: 45 slices, 2.5 mm thickness, voxel 2.5 mm^3, TR=3000 ms, TE=30 ms, FOV=190 mm, flip=90°. Four runs (~6:06 each). Preprocessing (SPM12): slice-timing correction, realignment, normalization to MNI (2 mm^3), smoothing 6 mm FWHM; motion <2 mm for all. fMRI analysis: Whole-brain event-related PLS (model-free, SVD decomposition) to identify latent variables (LVs) capturing covariance between brain activity and task design or behavioral covariates. Significance via 500 permutations; reliability via 100 bootstraps; peak voxels with bootstrap ratio (BSR) ≥3 and clusters ≥20 voxels reported. Spatiotemporal maps examined across lags (e.g., 3–6 s, 9–12 s, 12–15 s post-stimulus). Brain–behavior PLS related activation to DERS total, ERT RT, and AST RT. Initial task PLS included all conditions; a second PLS excluded SpeNeut to probe GenNeg vs SpeNeg effects.

Behavioral: - Age correlated with years of teaching (r=0.82, p<0.001). Controlling for age, years teaching positively correlated with slower emotion recognition (ERT median RT; r=0.31, p<0.05). PSS positively correlated with DERS (significant; exact r not reported). Emotional Stroop RTs intercorrelated across conditions; no significant correlations with other variables. - RM-ANOVA on RT (relative to control): no main effect of condition, F(3,144)=0.820, p>0.05. Accuracy: significant main effect, F(3,144)=37.24, p<0.001, η²=0.44; SpeNeut accuracy lower than GenNeg (t(48)=9.37, p<0.001), GenNeut (t(48)=7.17, p<0.001), SpeNeg (t(48)=9.63, p<0.001). All other conditions showed ceiling accuracy (>97%). Task PLS (neural): - All conditions included: 1 significant LV (36.32% covariance, p=0.014) differentiating SpeNeut from SpeNeg at 3–6 s post-onset. SpeNeut associated with greater activation in right frontal pole, middle frontal gyrus, inferior temporal gyrus, temporal pole, supramarginal gyrus, lateral occipital cortex; left hippocampus and amygdala; bilateral middle temporal gyrus, precuneus, cingulate gyrus, thalamus, putamen. Anti-correlated with right caudate body engaged in SpeNeg. - Excluding SpeNeut: 1 significant LV (42.06% covariance, p=0.032) with two temporal peaks: • Early (3–6 s): common activation for GenNeg and SpeNeg in left OFC, middle/superior temporal gyri, precuneus, cuneus; bilateral occipital poles extending to lateral occipital, lingual, fusiform gyri; reduced activation in bilateral thalamus and right superior frontal, precentral, cingulate, angular gyri. • Late (12–15 s): SpeNeg uniquely engaged left hippocampus, differentiating it from Control, GenNeut, GenNeg which showed increased engagement of bilateral OFC/IFG (pars triangularis/opercularis), SFG, postcentral gyrus, fusiform (occipital), cingulate, cerebellum, right frontal pole/SFG, right inferior/middle temporal gyri, right insula/pallidum/thalamus, left precentral, left fusiform (temporal), left occipital pole/lingual/fusiform. Brain–behavior PLS: - DERS (emotion dysregulation): Significant LV (51.38% covariance, p<0.001), peak at 9–12 s. Greater DERS associated with increased activation in bilateral lingual gyrus/occipital fusiform, precuneus/cuneus/lateral occipital, angular gyrus, frontal pole, SFG/MFG, right IFG (opercularis/triangularis), right STG, left MTG, left supramarginal/central opercular/insular cortex, thalamus, cerebellum during GenNeut (r≈−0.61 in brain score plot indicating higher DERS linked to stronger engagement). Marginal reversal for SpeNeg (r=0.19) and Control (r=0.16). - ERT (emotion recognition speed): Significant LV (35.72% covariance, p=0.012), peak 12–15 s. Faster ERT (lower RT) correlated with greater activation during GenNeg (r=−0.57) and GenNeut (r=−0.44); slower ERT correlated with greater activation during SpeNeg (r=0.35) across a network including lateral occipital, precuneus/angular, supramarginal, superior parietal, middle/superior temporal, pre/postcentral, superior/middle frontal (including OFC), cingulate/paracingulate, thalamus, caudate, cerebellum. - AST (cognitive control): No significant LV.

Despite moderate-high stress, teachers showed no behavioral emotional interference on the counting Stroop, suggesting effective control of responses to negative stimuli. Neuroimaging revealed early engagement (3–6 s) of networks associated with cognitive reappraisal and visual/semantic processing (OFC, fusiform, precuneus, temporal regions) during negative trials, consistent with strategic down-regulation rather than heightened amygdala-driven reactivity. A later temporal peak (12–15 s) showed hippocampal engagement uniquely for teacher-specific negative words, implying additional associative/relational processing when stimuli are occupation-relevant and negative. Individual differences clarified these effects: greater emotion dysregulation was linked to heightened activation to neutral words (GenNeut) in visual, parietal, frontal control, and insular regions, suggesting hyper-reactivity to ambiguous/neutral stimuli among those with regulation difficulties. Emotion recognition speed modulated activation: faster recognizers recruited the network more for general words, while slower recognizers relied more on visual/word-processing regions during teacher-specific negative words, potentially reflecting expertise-related processing trade-offs. Overall, findings indicate that stressed teachers may employ adaptive regulation (e.g., reappraisal) to mitigate interference from negative stimuli, with occupation-specific negative content engaging memory-related processes and emotion dysregulation manifesting as interference to neutral stimuli.

This study characterizes spatiotemporal neural correlates of emotional interference in stressed teachers. Key contributions include: (1) absence of behavioral interference despite moderate-high stress; (2) early engagement of reappraisal/cognitive control and visual-semantic regions during negative trials; (3) late hippocampal involvement specific to teacher-related negative stimuli; (4) association of emotion dysregulation with increased activation to neutral words; and (5) emotion recognition speed differentially relating to activation depending on stimulus specificity. These results suggest learned or strategic emotion regulation may buffer interference in this population. Future research should directly assess regulation strategies in-task, measure stimulus valence/arousal in the target population, include low-stress teacher and non-teacher controls, and use temporally sensitive modalities (MEG/EEG) to probe regulation dynamics. Exploring potential trade-offs between occupational expertise and face/emotion recognition, and the influence of aging, is also warranted.

- Sample: non-clinical stressed teachers without low-stress teacher or non-teacher control groups limits generalizability and causal inference. - Behavioral ceiling effects (accuracy >97% in most conditions) may mask subtle interference; specific neutral condition showed reduced accuracy. - Stimulus characteristics (teacher-specific neutral/negative) may have been similarly arousing; valence/arousal ratings were not collected. - Age range wide (25–67 years); potential confounding with years of teaching and emotion recognition; limited power for age-related subgroup analyses. - No direct assessment of emotion regulation strategy use during task; interpretations of reappraisal are inferential. - fMRI temporal resolution limits fine-grained temporal dynamics; modalities like MEG/EEG recommended. - Gender imbalance (predominantly female) may limit generalizability. - PLS is exploratory; findings need replication.