Negative Emotion Induction Manipulation Check
To assess whether the negative films elicited negative emotion, we contrasted watch-negative versus watch-neutral conditions on three indices of negative emotion: experience, expressive behavior and neural response.
The watch-negative versus watch-neutral contrast was significant (t=7.46, p<.000005, η2=.78), confirming the induction of negative emotion experience ().
Effect of reappraisal and suppression on negative emotion ratings and disgust facial expression intensity during the four film viewing conditions. Error bars = SEM.
Two raters coded participants' facial expressions during MR scanning for disgust expressions. Inter-rater reliability during the watch-negative condition was adequate (kappa=.76). The watch-negative versus watch-neutral contrast was significant (t(14)=2.96, p<.05, η2=.40), confirming that watching negative films induced emotion-expressive behavior ().
The watch-negative versus watch-neutral contrast resulted in enhanced responses in bilateral dorsal amygdala and anterior insula (), frontal cortex (medial, dorsomedial, dorsolateral, ventrolateral), temporal cortex (inferior, superior), occipital (lingual gyrus), parietal cortex (superior parietal lobule), and subcortical regions (caudate, thalamus, hypothalamus) (). There were no significant brain responses for watch-neutral versus watch-negative.
fMRI BOLD signal time-series in percent signal change relative to the watch-neutral condition across 15s (10 time points × 1.5s each) for reappraisal, suppression, and watch-negative conditions in bilateral amygdala and insula.
Watch-Negative versus Watch-Neutral Film BOLD Response
Emotion Regulation Analyses
We examined the effects of reappraisal and suppression on (a) emotion experience and behavior, (b) emotion-related brain regions (amygdala and insula regions of interest identified in the contrast of watch-negative versus watch-neutral films), and (c) regulation-related brain regions (identified in the reappraise versus watch-negative and suppress versus watch-negative film contrasts. Based on our model of emotion regulation, we expected differential temporal effects of (antecedent-focused) reappraisal and (response-focused) suppression on emotion-generative and emotion regulatory brain regions. To test for hypothesized differential ER effects on neural temporal dynamics, analyses focused on neural responses during the early (0−4.5s), middle (4.5−10.5s), and late (10.5−15s) periods of each 15s film. We first examined each ER strategy separately, and then directly compared reappraisal versus suppression.
Reappraisal (versus watch-negative) reduced negative emotion experience (t=4.70, p<.0005, η2=.58), approached significant reduction of disgust facial behavior (t=1.82, p=.09, η2=.20), and reduced emotion-related neural signal during the late (but not early and middle) components in right amygdala (reappraisal vs. watch-negative, t=2.20, p<.05, η2=.21), left insula (t=2.22, p<.05, η2=.24) and marginally in left amygdala (t=1.89, p<.08, η2=.18). These results show that reappraisal effectively down-regulated negative emotion experience with a concomitant reduction of emotion-related neural signal by the end of the 15s films.
During the early (but not middle and late) period, reappraisal produced enhanced responses in PFC-related cognitive control of emotion (medial, dorsolateral, ventrolateral PFC, and lateral OFC), linguistic processing (left inferior frontal gyrus, left posterior superior temporal gyrus), visual attention (precuneus, lingual and angular gyri), and feature detection (middle and superior temporal gyri) (). Three representative early reappraisal-related neural responses, including medial PFC, left inferior PFC and left OFC, are shown in . There were no areas of greater response for watch-negative versus reappraise-negative.
Reappraise versus Watch-Negative and Suppress versus Watch-Negative Film BOLD Response for Early (0−4.5s), Middle (4.5−10.5s), and Late (10.5−15s) Components
Figure 5 fMRI BOLD signal time-series in percent signal change relative to the watch-neutral condition for reappraisal, suppression, and watch-negative conditions in medial prefrontal cortex BA 10 (−11 67 18), left inferior prefrontal cortex BA 46 (−51 (more ...)
Correlation analyses demonstrated that increased early (0−4.5s) reappraisal-related medial PFC and left OFC responses were associated with significantly decreased late (10.5−15s) left amygdala and left insula responses (). These findings indicate that implementation of reappraisal strategies may initiate interactions of regulatory and regulated brain systems that evolve over time and influence subsequent neural, experiential and behavioral indices of emotion.
Association of early (0−4.5s) enhanced cognitive reappraisal-related BOLD responses in orbitofrontal cortex and medial prefrontal cortex with reduced late (10.5−15s) insula and amygdala responses.
Suppression (versus watch-negative) reduced negative emotion experience (t=3.30, p<.005, η2=.41) and facial behavior (t=2.49, p<.05, η2=.32). In limbic regions previously identified in the watch-negative versus watch-neutral contrast, responses in the late component were greater for suppression in right insula (t=2.42, p<.05, η2=.27) and marginally greater in right amygdala (t=1.86, p=.08, η2=.18), but similar in left amygdala (p>.57) and left insula (p>.90). These findings demonstrate no reduction in emotion-related neural activity in amygdala, and indeed suggest enhanced responding for suppression in right insula.
Suppression (versus watch-negative) produced greater responses only during the late (not early and middle) period in regions implicated in inhibitory control (right ventrolateral PFC, ), cognitive regulation (dorsomedial, dorsolateral PFC), visual-sensory multimodal association (posterior occipito-temporal lobes) and visual-spatial processing (precuneus and occipital areas) (). Greater responses for watch-negative (versus suppression) occurred only during the late period in visual processing areas (cuneus and lingual gyrus).
Figure 7 fMRI BOLD signal time-series in percent signal change relative to the watch-neutral condition for reappraisal, suppression, and watch-negative conditions in right inferior prefrontal cortex BA 46 (50 42 3) and right inferior prefrontal cortex BA 45 (52 (more ...)
Comparison of Reappraisal and Suppression
Direct contrast of ER strategies revealed that reappraisal produced greater down-regulation of negative emotion experience (t=3.29, p<.005, η2=.40), while suppression produced greater reduction in disgust facial behavior (t=2.94, p<.05, η2=.40). These findings highlight the differential impact of reappraisal and suppression on negative emotion experience and behavior.
Neurally, in the previously identified watch-negative emotion-related regions, a 2 (ER: reappraisal, suppression) × 2 (time: early, late) repeated-measures ANOVA of BOLD signal revealed an ER × time interaction in bilateral amygdala (left, F(2,15)=4.31, p=.054, η2=.21; right, F(2,15)=4.74, p<.05, η2=.23) and insula (left, F(2,15)=5.03, p<.05, η2=.24; right, F(2,15)=4.04, p=.06, η2=.20). Follow-up t-tests showed that reappraisal compared to suppression resulted in reduction at the late period in amygdala (left: t=2.72, p<.05, η2=.32; right: t=2.82, p<.05, η2=.33) and insula (left: t=2.04, p=.059, η2=.21; right: t=2.27, p<.05, η2=.24). These results show that reappraisal was more effective than suppression in down-regulation of BOLD responses in emotion-related limbic regions.
In the 3 reappraisal and 2 suppression related activation clusters displayed in & 9, a 2 (ER: reappraisal, suppression) × 2 (time: early, late) repeated-measures ANOVA of BOLD signal resulted in a significant interaction of ER × time in left IFG BA46, F(2,15)=8.38, P>.05, η2=.34, mPFC BA10, F(2,15)=7.89, P>.05, η2=.33, right IFG BA46, F(2,15)=5.38, P>.05, η2=.25, right IFG BA45, F(2,15)=13.31, P>.005, η2=.45, but only a trend towards significance in left OFC BA11, F(2,15)=3.69, P=.07, η2=.19.
Direct reappraisal versus suppression contrasts during the early, middle, and late film components produced neural activity very similar to the results reported above. Reappraisal generated significant responses only during the early period in PFC regions (medial, dorsomedial, left inferior frontal gyrus), insula, superior and middle temporal lobes, and cuneus. Suppression produced significant responses only during the late period in PFC regions (ventromedial, ventrolateral, dorsolateral, rostral and dorsal ACC), left fusiform gyrus, inferior parietal lobule, middle and superior occipital gyrus, and thalamus (). There were no differential BOLD responses during the middle period. These direct contrasts re-affirmed the pattern of results described above, namely, differentially greater PFC BOLD responses for reappraisal during the early period only and for suppression during the late period only.
Reappraise versus Suppress BOLD Response for Early (0−4.5s), Middle (4.5−10.5s) and Late (10.5−15s) Components