Cognitive Reappraisal and Expressive Suppression Evoke Distinct Neural Connections during Interpersonal Emotion Regulation

The study investigates how interpersonal emotion regulation (IER)—deliberate attempts to change another person’s emotions—recruits neural systems and differs by strategy. Prior work has emphasized intrapersonal regulation (cognitive reappraisal [CR] vs expressive suppression [ES]); CR reframes meaning, while ES inhibits expression. CR is generally more adaptive intrapersonally, but its relative effectiveness and neural underpinnings in IER remain unclear. Using hyperscanning fNIRS, the authors examine inter- and intra-brain connectivity during three IER stages (emotion sharing, regulation, feedback), grounded in models positing involvement of social cognition/empathy, cognitive control, and affect systems. They hypothesize: (1) CR would more effectively downregulate targets’ negative emotions than ES; (2) CR would enhance intrapersonal functional connectivity (FC) in regions linked to mirror neuron, mentalizing, and cognitive control systems (e.g., STG, AG, DLPFC); and (3) CR would enhance interpersonal brain synchronization (IBS) in mentalizing/alignment regions (e.g., IFG, TPJ).

The paper reviews intrapersonal regulation literature showing CR’s benefits over ES for affect and well-being and ES’s limited effects and physiological costs. It highlights gaps in IER neural mechanisms, noting prior single-brain studies and human-computer paradigms. Hyperscanning research shows IBS supports social interaction and has implicated DLPFC, TPJ, IFG, and STG in dialogue, empathy, and alignment. Models of IER (e.g., Reeck et al., 2016) posit roles for social cognition/empathy, cognitive control, and emotion generation systems, and delineate stages: sharing, regulation, and feedback. Emotional empathy is emphasized in sharing; cognitive empathy in regulation. The review motivates examining strategy-specific neural coupling (FC and IBS) across stages.

Design and participants: 35 female friend dyads (N=70; mean age 19.60±1.48), right-handed, heterosexual, healthy, with normal/corrected vision, were recruited; 1 dyad was excluded (missing data), yielding 34 dyads (CR: n=18, 20.09±1.21; ES: n=16, 19.11±1.75; age matched, t(66)=0.99, p=0.31). A 2 (strategy: CR vs ES; between-dyad) × 3 (stage: sharing vs regulation vs feedback; within-dyad) factorial design was used. Power analyses (a priori f=0.25, power=0.80; post hoc G*Power) indicated sufficient power. Ethics approval and informed consent obtained. Measures and materials: Relationship Closeness Inventory assessed acquaintance length and closeness (Cronbach’s α=0.87). Regulators completed 5-day pre-experiment practice with definitions and examples for assigned strategy (CR or ES). Six ~2-min negative emotional video clips (natural disasters, assaults, wars) were pre-rated by an independent sample (n=30) for valence and arousal; three clips inducing moderate negative affect selected (no differences across clips; valence F(2,58)=1.20, p=0.31; arousal F(2,58)=1.08, p=0.42). Emotion Rating Scale (Gross, 1998): 6 positive and 8 negative emotion items (1–9), plus arousal (1–9) and IER effectiveness (1–9) per IER session. Procedure: Within each dyad, one participant was Target (emotion induction via 120 s clip) and the other Regulator (implements assigned strategy). Each experiment included 3 resting sessions and 3 IER sessions. Each IER session: (1) Target views clip (120 s) and both rate emotions/Regulator completes a strategy check; (2) Stage I—Sharing: Target describes feelings (3 prompts) while Regulator listens; (3) Stage II—Regulation: Regulator applies CR or ES with on-screen tips, using own words; (4) Stage III—Feedback: Target reports current feelings and perceived regulation effectiveness; both then rate emotions/arousal and IER effectiveness post-session. Negative emotion induction was successful (targets: low positive 1.64±0.78; moderate negative 5.00±1.00). fNIRS acquisition: Dual-brain hyperscanning with Hitachi ETG-7100, wavelengths 695/830 nm, 10 Hz sampling. HbO and HbR computed via modified Beer–Lambert; analyses focused on HbO. Two 3×5 optode arrays per participant (8 emitters, 7 detectors; 22 channels; 3 cm spacing) covering bilateral frontal, temporal, parietal regions; positioned via 10–20 system (T3/T4); MNI estimates via virtual registration provided. Preprocessing and analysis: PCA to remove global components; correlation-based signal improvement (CBSI) for motion artifact removal. FC and IBS computed as cross-correlations between HbO time series across channels within-person (FC) and between paired participants (IBS), for each stage and session; matrices averaged across 3 IER sessions and baseline-corrected by subtracting resting-session matrices; Fisher r-to-z transformed. Two-way mixed ANOVAs tested STRATEGY × STAGE effects on FC/IBS with FDR correction; Bonferroni post hocs/simple effects as needed. Permutation tests: phase randomization for FC (1,000 iterations) and dyad reshuffling for IBS (1,000 iterations). Brain–behavior Pearson correlations performed with FDR correction.

Behavioral:

• Relationship closeness: No differences between CR and ES groups across closeness indices (all p>0.05).
• IER effectiveness and affect change (targets): Significant decreases after IER in negative emotion (−1.94±0.85; t(33)=−13.30, p<0.001, d=3.21) and arousal (−1.53±1.59; t(33)=5.60, p<0.001, d=1.36), and increase in positive emotion (+0.89±0.85; t(33)=6.11, p<0.001, d=1.48). CR rated more effective than ES: CR 6.54±0.79 vs ES 4.88±1.81; t(20.05)=3.40, p=0.003, d=1.19. No other target group differences in overall emotion/arousal change.
• Regulators: Post-IER increases in arousal (1.04±1.26; t(33)=4.82, p<0.001, d=1.17) and negative emotion (0.29±0.55; t(33)=3.08, p<0.001, d=0.75) and decrease in positive emotion (−0.57±0.82; t(33)=−4.04, p<0.001, d=0.98). CR vs ES: greater arousal increase in CR (1.44±1.14) than ES (0.58±1.26), t(32)=2.09, p=0.044, d=0.72; specific emotions—regulators showed higher amusement change in CR (0.26±1.36) vs ES (−0.90±0.86), t(32)=2.92, p=0.006, d=1.02, and greater fear decrease in CR (−0.22±0.38) vs ES (0.17±0.50), t(32)=−2.57, p=0.015, d=0.88.
• Cross-partner correlations: CR group—targets’ IER effectiveness correlated with regulators’ IER effectiveness (r=0.54, p=0.02) and regulators’ positive emotion change (r=0.51, p=0.03). ES group—targets’ positive emotion change correlated with regulators’ arousal change (r=0.71, p=0.002); targets’ negative emotion change inversely correlated with regulators’ IER effectiveness (r=−0.52, p=0.04). Neuroimaging—Main effects of STRATEGY:
• Targets’ FC: Higher in CR than ES across multiple connections spanning cognitive control and social cognition/mirror systems, including lSTG–lIFG (T8–T23), lSTG–lDLPFC (T30–T32), rSTG–lSTG (T3–T25), rFG–rV3 (T4–T18), among others (FDR-corrected; see Table 4 in text). ES>CR for rSTG–lMotor (T3–T41), rV3–lPAC (T9–T24), SMG–lAG (T12–T44).
• Regulators’ FC: CR>ES for rIFG–lMotor (R1–R28); ES>CR for rIFG–lFEF (R1–R36).
• IBS: CR>ES for lIFGReg–rSTGTar (R23–T3); ES>CR for rV3Reg–rSMGTar (R22–T21). Neuroimaging—STRATEGY × STAGE interactions (selected):
• Targets’ FC: Significant interactions for rDLPFC–rSTG (T5–T8), rV3–lMotor (T22–T28), rPSC–rSMG (T33–T34), rPSC–lSMG (T42–T43). At sharing stage: CR>ES for rDLPFC–rSTG (p=0.002, d=1.16) and rPSC–rSMG (p<0.001, d=1.35); ES>CR for rV3–lMotor (p<0.001, d=1.59) and rPSC–lSMG (p=0.008, d=0.98). Stage-wise within-strategy differences reported (e.g., sharing>feedback for several CR FCs).
• Regulators’ FC: lSMG–lMotor (R29–R37) interaction; at sharing, ES>CR (p=0.002, d=1.17). Marginal interaction for rFG–lAG (R4–R44): at feedback, CR>ES (p=0.002, d=1.12).
• IBS: Interactions at rFGReg–lDLPFCTar (R4–T27), rSMGReg–lDLPFCTar (R7–T32), rAGReg–lFEFTar (R17–T36), lV3Reg–rSMGTar (R31–T21), lPSCReg–lSMGTar (R42–T29). Notably, sharing: CR>ES for rSMGReg–lDLPFCTar (p=0.013, d=0.90); regulation: ES>CR for rFGReg–lDLPFCTar (p=0.048, d=0.73) and rAGReg–lFEFTar (p=0.04, d=0.75); feedback: ES>CR for lPSCReg–lSMGTar (p=0.002, d=1.17). Stage-wise trends differed across strategies (e.g., ES showed increasing IBS from sharing to regulation/feedback in several pairs; CR often decreased by feedback). Validation and brain–behavior links:
• Permutation tests placed observed F statistics for significant effects in top ~1% of null distributions for both FC and IBS.
• Brain–behavior correlations (FDR-corrected): CR—regulators’ positive emotion change correlated with FC rIFG–lFEF (R1–R36) at feedback (r=0.60, Pcorr=0.049); regulators’ positive emotion change also correlated with targets’ FC SMG–lAG (T12–T44) at feedback (r=0.81, Pcorr=0.006). ES—targets’ negative emotion change correlated with targets’ FC lSTG–lDLPFC (T30–T32) at regulation (r=0.79, Pcorr=0.007). IBS: CR—targets’ IER effectiveness correlated with IBS rAGReg–lFEFTar at regulation (r=0.72, Pcorr=0.023). ES—regulators’ positive emotion change correlated with IBS lPSCReg–lSMGTar at feedback (r=0.72, Pcorr=0.042).

Both CR and ES effectively reduced targets’ negative affect and arousal and increased positive affect during IER, with targets rating CR as more effective. Neurally, CR evoked broader intra- and inter-brain couplings across cognitive control and social cognition/mirror systems (e.g., PFC, TPJ/STG, sensory–motor), aligning with the cognitive reframing and mentalizing demands of CR. ES preferentially engaged connections involving visual and sensorimotor regions and increased IBS between regulators’ visual/face-processing and targets’ control/monitoring regions, consistent with suppression’s response-focused, monitoring and inhibitory control emphasis. Stage-specific effects indicate strategy influences emerge as early as the sharing stage: CR enhanced target FC (rDLPFC–rSTG; rPSC–rSMG) and IBS (rSMGReg–lDLPFCTar), consistent with empathy/mentalizing and executive preparation, whereas ES increased FC in sensorimotor networks in both partners. During regulation, ES increased IBS linking regulators’ fusiform/AG with targets’ control/attention systems, consistent with monitoring of expressions and inhibitory control. Brain–behavior relationships suggest functional couplings relate to perceived effectiveness and emotional change, supporting functional relevance of these networks. Together, findings support IER models implicating social cognition, cognitive control, and affective/mirror systems and reveal distinct neural coupling signatures for CR vs ES across IER stages.

This study demonstrates that interpersonal cognitive reappraisal and expressive suppression both downregulate others’ negative emotions, with CR perceived as more effective by targets. Using fNIRS hyperscanning, the study reveals distinct intra- and inter-brain connectivity patterns: CR engages broader couplings across prefrontal, temporal, and parietal regions and enhances cross-brain synchronization during sharing; ES preferentially increases cross-brain synchrony during regulation in visual/face-processing and control networks. These results refine IER models by mapping stage-specific neural dynamics and highlight the roles of cognitive control, mentalizing, and mirror systems in social regulation of emotion. Future work should test generalizability across genders and relationship types, incorporate objective physiological measures, examine individual differences (e.g., strategy preferences, regulation ability), and assess flexible, context-sensitive strategy deployment.

Generalizability and ecological validity are limited: strategies were assigned rather than chosen to match personal preferences or contexts; individual differences (e.g., habitual IER strategy, regulation ability) may moderate effects. Sample size, while powered and comparable to prior hyperscanning work, would benefit from replication in larger cohorts. Only female dyads were studied; findings may not extend to males or mixed-gender dyads. Emotional states were self-reported only; objective physiological indices (e.g., heart rate, skin conductance) were not collected. The study did not test flexible or context-adaptive strategy selection, which may influence IER efficacy.