Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Behavioral and neural indices of affective coloring for neutral social stimuli

Behavioral and neural indices of affective coloring for neutral social stimuli Emotional processing often continues beyond the presentation of emotionally evocative stimuli, which can result in affective biasing or coloring of subsequently encountered events. Here, we describe neural correlates of affective coloring and examine how individual differences in affective style impact the magnitude of affective coloring. We conducted functional magnetic resonance imaging in 117 adults who passively viewed negative, neutral and positive pictures presented 2 s prior to neutral faces. Brain responses to neutral faces were modulated by the valence of preceding pictures, with greater activation for faces following negative (vs positive) pictures in the amygdala, dorsomedial and lateral prefrontal cortex, ventral visual cortices, posterior superior temporal sulcus, and angular gyrus. Three days after the magnetic resonance imaging scan, participants rated their memory and liking of previously encountered neutral faces. Individuals higher in trait positive affect and emotional reappraisal rated faces as more likable when preceded by emotionally arousing (negative or positive) pictures. In addition, greater amygdala responses to neutral faces preceded by positively valenced pictures were associated with greater memory for these faces 3 days later. Collectively, these results reveal individual differences in how emotions spill over onto the processing of unrelated social stimuli, resulting in persistent and affectively biased evaluations of such stimuli. Key words: affective coloring; amygdala; individual differences; faces; emotion in an intractable state of negative affect, which is subsequently Introduction reflected in overly critical evaluations of the statistics exams Emotional experiences are inherently ‘sticky’. Responses to she grades that evening. On the other hand, after having a affectively significant events (whether positive or negative) are manuscript accepted in a prestigious journal, the same student rarely temporally constrained to the time of exposure, but can may experience sustained positive affect that positively colors persist long after these events have passed (Frijda et al., 1991; her evaluations of the next stack of exams. Notably, under cer- Verduyn et al., 2015). Completely unrelated events, encountered tain circumstances such affective coloring has the potential to in the wake of these evocative emotional experiences, may be result in biased long-term evaluations of, or memory for, subse- processed through a particular affective filter, consequently quent unrelated events or interactions (Tambini et al., 2017; ‘coloring’ the perception or evaluation of unrelated events Lapate et al., 2017). While recent studies have begun to illumin- (Murphy and Zajonc, 1993; Anderson et al., 2012). For example, ate the neural correlates of affective coloring (Lapate et al., 2016, harsh feedback from a graduate student’s supervisor may result 2017; Tambini et al., 2017), it is unknown whether and how Received: 21 August 2017; Revised: 17 January 2018; Accepted: 9 February 2018 V C The Author(s) (2018). Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 311 inter-individual variability in neural mechanisms or trait-like All participants were right handed and were screened to ensure indices of affective style bear on affective coloring outcomes. MRI compatibility, and all provided informed consent prior to Individual differences in the temporal dynamics of emo- all study procedures. tional responding, or ‘affective chronometry’ (Davidson, 1998), may be critically important for the extent of such affective col- MRI data collection oring. It is now well established that individual differences in MRI data were collected on a 3 T scanner (MR750 GE Healthcare, affective chronometry are critical for both positive well-being Waukesha, WI) using an 8-channel head coil. We collected three (Davidson 2004; Heller et al. 2013; Schaefer et al. 2013) and risk of sets of echo planar images during the fMRI task (231 volumes, psychopathology (Siegle et al. 2002; Burke et al. 2005; Larson et al. TR ¼ 2000, TE ¼ 20, flip angle¼ 60 , field of view ¼ 220 mm, 9664 2006; Dichter and Tomarken 2008; Blackford et al. 2009; Lapate matrix, 3 mm slice thickness with 1 mm gap, 40 interleaved sa- et al. 2014; Schuyler et al. 2014; Heller et al. 2015) in addition to gittal slices, and ASSET parallel imaging with an acceleration negative physical health outcomes (Brosschot et al., 2006). factor of 2). A three-dimensional magnetization-prepared rapid Investigating affective coloring may afford insight into one gradient-echo sequence (Mugler and Brookeman, 1990) was pathway through which the temporal dynamics of emotion can used to acquire a T1-weighted anatomical image for functional impact psychological health: Individuals who quickly recover to data registration (TR ¼ 8.2, TE ¼ 3.2, flip angle ¼ 12 , field of a baseline state may exhibit minimal behavioral or biological in- view ¼ 256 mm, 256256 matrix, 160 axial slices, inversion time- dicators of spillover of emotion from one context to the next, ¼ 450 ms). The MRI session also included the acquisition of field whereas those who are slower to recover from negative events map images, a resting-state scan, diffusion-weighted data, and (or who are more capable of ‘savoring’ positive events) will be perfusion data. more susceptible to this spillover or coloring. For example, pre- vious studies utilizing daily diary methods have shown that while individuals with depression do not necessarily show Functional MRI task greater reactivity to stressful life events, the impact of interper- During fMRI scanning, participants passively viewed pictures sonal stressors ‘spills over’ into greater experiences of negative from the International Affective Picture System (IAPS; Lang affect on subsequent days (Gunthert et al., 2007), and that the et al., 2008), presented for 4 s (Figure 1). Each of three runs con- extent of this spillover predicts poorer response to cognitive be- tained 10 pictures each from negative, neutral and positive va- havioral therapy (Cohen et al., 2008). The laboratory investiga- lence categories, for a total of 90 picture presentations. Pictures tion of individual differences in affective coloring may thus were presented using a pseudorandom trial order with the re- provide novel mechanistic insight into affective pathology and quirement that no more than two pictures from the same va- its treatment in a more carefully controlled setting. lence category be presented in a row. Within each valence To that end, our lab has developed a functional magnetic category, stimuli were randomly selected for each participant. resonance imaging (fMRI) paradigm designed to measure the Each picture was followed by a 2 s inter-stimulus interval and a impact of well-characterized affective stimuli on immediate 0.5 s neutral face presentation. Participants were instructed to (neural) and long-term (behavioral) indices of affective coloring press one of two buttons to indicate the gender of the face, a (specifically, the coloring of neutral social stimuli; Figure 1). In low-level categorization task chosen to confirm task engage- this task, participants passively view negative, neutral and posi- ment without interfering with natural processing of the faces or tive pictures that are followed by a face displaying a neutral ex- preceding IAPS pictures. Faces were followed by an inter-trial pression, and are instructed to indicate the gender of the face (a interval between 3.5 and 27.5 s (mean duration ¼ 7.5 s). A 1 s low-level categorization task). To investigate affective coloring crosshair appeared before the start of each trial to orient partici- of neural responses, we compare brain responses to neutral pants’ attention. faces as a function of the preceding picture valence. To investi- This trial timing was chosen based on simulations conducted gate enduring behavioral indices of affective coloring, we pre- using optseq2 (https://surfer.nmr.mgh.harvard.edu/fswiki/opt sent the same faces to participants 3 days later and test the seq2), with the constraint that faces be presented within 4 s of extent to which memory or liking of these faces is modulated by picture presentation. This relatively brief ISI increased statistical the type of picture that initially preceded these faces. power to observe a spillover effect by maximizing the number of trials within the scan session. Specifically, we compared model efficiency for trial timings that included (i) average ISIs of 2 and Materials and methods 3 s, (ii) jittered ISIs of 0, 0.5 and 1 s and (iii) ITIs that varied (on Participants average) between 4.5 and 8 s. Based on these simulations, the fixed 2 s interval provided the greatest power to detect effects of Participants for this study were drawn from Midlife in the United States (MIDUS), a national longitudinal study of health interest during the face period. Because this fixed ISI may raise concerns regarding collinearity of picture and face regressors, we and well-being across the lifespan (http://www.midus.wisc. edu). For this study, we included participants from the MIDUS calculated the variance inflation factor (VIF) for each regressor. A ‘refresher’ sample who were enrolled beginning in 2012 in an ef- VIF cutoff of either 5 or 10 is typically used to indicate problem- atic levels of collinearity, and VIF values for the face regressors fort to expand and refresh the original MIDUS sample with younger age cohorts. The majority of the MIDUS refresher sam- fell close to 5 (mean for picture regressors¼ 4.22–4.72; mean for ple was recruited through random digit dialing of adults face regressors¼ 4.82–5.45). Although the VIF values for face throughout the U.S. The refresher sample also includes an over- regressors in some cases exceeded this more conservative sampling of African American participants from Milwaukee, WI threshold, collinearity generally does not affect the Type I error rate, but rather can lead to more variable estimates and thus who were recruited by door-to-door solicitation. fMRI task data were collected on a subset of 122 MIDUS refresher participants negatively impact statistical power (Mumford et al., 2015). In par- living in the Midwest who were able to travel to our laboratory, ticular, more variable estimates at the level of individual subjects including 35 participants from the Milwaukee subsample. will average out across subjects, and should thus stabilize with a Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 312 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 Fig. 1. (A) Schematic of fMRI trial structure. On each trial, participants passively viewed a 4 s picture from the IAPS (Lang et al., 2008). Following a 2 s delay, a neutral face was presented and participants pressed a button to indicate the gender of the face. (B) Three days following the scan, participants used an online or paper rating scale to provide memory then likability ratings for 45 neutral faces viewed during the fMRI scan and 45 novel neutral faces. larger sample size. In contrast, collinearity in higher-level ana- participants and two runs from one participant; for the other lyses (which was not the case in this study) may result in invalid 106 participants, we included data from all three runs in group- inferences regarding the magnitude of effect sizes (for discussion, level analyses. The first-level general linear model included six regressors see Mumford et al., 2015). Average normative arousal ratings were equivalent for nega- of interest, consisting of separate picture and face regressors for tive (5.466 0.66) and positive (5.476 0.53) IAPS pictures, which each picture valence (Negative/Neutral/Positive). Regressors of were both greater than those for neutral pictures (3.166 0.42; no interest included mean-centered response times to face ts> 16.1, Ps< 0.001). Pictures within each valence category were presentations, six motion parameters with their first- and matched on luminosity and visual complexity, and for the num- second-order derivatives, and a confound regressor for each ber of pictures rated to be social in content. Faces were drawn volume with> .9 mm framewise displacement (Siegel et al., from the XM2VTSDB multi-modal face database (Messer et al., 2014). Additional processing steps included correction of time 1999), the NimStim database (Tottenham et al., 2009), and the series autocorrelation, resampling of functional data to 2 mm Montreal Set of Facial Displays of Emotion (Beaupre et al., 2005). isotropic voxels, and registration to Montreal Neurological In light of our diverse subject population, these faces included Institute template space. equal proportions of male and females with a broad range of As a complementary approach, we modeled amygdala time ages and ethnicities. Faces were cropped just above the hair and series data using a set of seven finite impulse response (FIR) below the chin, converted to black and white, and edited (e.g. to basis functions. This allowed for model-free estimation of the remove distinctive facial hair and eyeglasses). A total of 45 neu- hemodynamic response from the time of picture onset through tral faces were presented, and each face was presented follow- the first 14 s of each trial (corresponding to the minimum trial ing two randomly selected pictures from the same valence length). We then extracted empirically derived activation esti- category. mates for each trial type separately. For group-level analyses we averaged these estimates across individuals at each volume, calculated valence and arousal contrast estimates at each time FMRI data processing and analysis point, and conducted paired-sample t tests for these contrasts. FMRI data processing was carried out using FEAT (FMRI Expert Analysis Tool) Version 6.00, part of FSL (FMRIB’s Software Group analysis of fMRI data Library, www.fmrib.ox.ac.uk/fsl). Preprocessing steps included removal of the first four volumes, motion correction using Primary group analyses for the fMRI data were conducted using MCFLIRT, removal of non-brain regions using BET, spatial two orthogonal contrasts for both the picture and face epochs: a smoothing using a Gaussian kernel with 5 mm full-width at ‘Valence’ contrast (Negative vs Positive) and an ‘Arousal’ con- half-maximum, grand-mean intensity normalization, and high- trast [0.5*(Negative þ Positive) Neutral]. For significant effects pass temporal filtering. We excluded five participants following involving the Arousal contrast, follow-up analyses examined whether effects were driven more by the Negative or Positive first-level data processing (1 for excessive movement [>1mm mean framewise displacement across the entire run], 2 for fail- condition. For primary analyses focused on the amygdala, we ing to provide behavioral responses, and 2 for distorted func- extracted mean parameter estimates for each of these contrasts tional data), resulting in a final sample size of 117 for fMRI from voxels with  50% probability of being assigned to the bi- analyses (64 female/53 male; Milwaukee subsample N¼ 35; lateral amygdala from the Harvard–Oxford probabilistic atlas mean age6 s.d. ¼ 48.36 11.8; range ¼ 26–76). In addition, due to (volume ¼ 4040 mm ; Desikan et al., 2006; Figure 2A), which was excessive motion and/or failure to provide behavioral responses chosen as an a priori ROI prior to data analysis. In addition to on at least 25% of trials, we excluded one run from 10 this a priori region of interest, we conducted whole-brain Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 313 Fig. 2. (A) Anatomically defined amygdala region of interest (ROI). (B) Mean percent signal change values across the amygdala ROI for neutral faces presented following negative, neutral and positive pictures. (C) Estimated signal from the amygdala ROI using a set of finite impulse response basis functions revealed complementary evi- dence for greater amygdala activation to faces following negative vs positive pictures. Notes: Error bars represent 95% confidence intervals. *P < 0.05. voxelwise analyses on the same two contrasts for the picture with higher values indicating greater liking of, or greater confi- and face periods. Cluster threshold correction was applied using dence of having previously seen, each face. These ratings were a Z-threshold of 3.09 and a cluster-corrected significance value completed exactly 3 days after the scan in 85 out of 89 subjects of P< 0.05. who completed ratings online (two participants completed rat- ings 4 days after the scan, and one each completed ratings 7 and 9 days after the scan). All 16 participants completing paper rat- Behavioral data collection and analysis ings reported completing these exactly 3 days after the scan. Participants were asked to log onto the internet 3 days after the Analysis of these behavioral data tested the extent to which MRI session, and complete ratings of the 45 faces they viewed each participant’s long-term memory or liking of neutral faces during the MRI task, which were interspersed with 45 novel was colored by the content of preceding IAPS pictures from the faces. Each of these 90 faces was presented for an unlimited scan session. We conducted paired t tests on these behavioral duration with one of two continuous rating scales that asked data using the same contrasts as with the fMRI data in order to participants to rate their memory and liking of these faces. The test whether the valence or arousal of pictures had a persistent memory scale had equidistant anchor points labeled ‘com- impact on behavioral evaluations of subsequently presented pletely certain new’, ‘somewhat certain new’, ‘unsure’, ‘some- faces. We predicted that (i) memory for neutral faces would be what certain old’ and ‘completely certain old’; the liking scale enhanced for faces following emotionally arousing pictures had equidistant anchor points labeled ‘really dislike’, ‘some- (Negative or Positive, relative to Neutral), and (ii) liking of neu- what dislike’, ‘unsure’, ‘somewhat like’ and ‘really like’. Memory tral faces would be enhanced for faces following Positive rela- ratings were always completed prior to liking ratings. tive to Negative pictures. Memory and liking ratings were obtained from 105 and 104 participants, respectively. Of these participants, 16 were not Individual differences analyses comfortable using the online rating scale or did not have access to a computer; these individuals instead completed ratings We conducted analyses to explore the impact of individual dif- using paper packets. Ratings collected electronically and on ferences in both affective style and amygdala activation on the paper were both converted to values ranging between 1 and 1, affective coloring of behavior, as assayed by 3-day liking and Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 314 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 memory ratings. We examined four self-report measures of af- Amygdala responses to neutral faces were not modulated by fective style: state Positive and Negative Affect from the Positive the arousal contrast for preceding pictures [t(116)¼0.33, and Negative Affect Schedule (Watson et al., 1988); the P> 0.7]. However, amygdala responses to neutral faces were Reappraisal subscale of the Emotion Regulation Questionnaire modulated by the valence of preceding pictures, with greater ac- tivation when the preceding picture was negative vs positive (Gross and John, 2003); and total score on the Psychological [t(116) ¼ 2.39, P ¼ 0.027, d ¼ 0.22; Figure 2B]. We also estimated Well-Being Scale (Ryff and Keyes, 1995). For each of these four indices we calculated Pearson correlation coefficients with sub- amygdala responses directly from the data using a set of FIR basis functions (Figure 2C). This model-free approach revealed sequent ratings of face liking and memory, both as a function of clearly distinct hemodynamic responses in the amygdala to the preceding pictures’ Valence and Arousal. We corrected for mul- picture and face epochs, with peak responses three volumes tiple comparisons testing using a conservative Bonferroni- (6 s) after picture onset and two to three volumes (4–6 s) after corrected P value of 0.003125 [4 measures2 behavioral ratings face onset. Consistent with our primary analysis, we observed (liking/memory)2 contrasts (Valence/Arousal)]. greater amygdala activation for the negative vs positive condi- We also tested two specific hypotheses regarding relation- tion at volumes 5 [t(116) ¼ 2.76, P ¼ 0.0067, d ¼ 0.27] and 7 ships between individual differences in affectively-modulated [t(116) ¼ 3.68, P< 0.001, d ¼ 0.39], i.e. corresponding to 4 and 8 s, amygdala responses to neutral faces and long-lasting behav- respectively, after face onset. Results of the whole-brain, voxel- ioral evaluations of these faces. First, as an extension of find- wise FIR analysis for all volumes are available at https://neuro ings that the amygdala mediates arousal-enhanced memory for vault.org/collections/3154/. emotional stimuli (Canli et al., 2000; Kensinger and Corkin, Notably, when including age as a covariate, we no longer 2004), we hypothesized that amygdala responses to faces follow- observed a significant group effect of Valence [t(112) ¼ 1.24, ing arousing pictures would be associated with greater memory P ¼ 0.22]. Indeed, age showed a marginally significant correl- for those faces 3 days later. Second, we correlated amygdala re- ation with amygdala activation for the Valence contrast sponses to faces following Negative vs Positive pictures [r(115) ¼ 0.17, P ¼ 0.064], with greater (Negative> Positive) amyg- (Valence contrast) with liking ratings, with the prediction that dala activation in older adults. To visualize this age effect, we more negative affective coloring of faces (as manifest in stron- conducted a post hoc median split of the sample into adults ger amygdala responses) would be associated with relatively younger than 50 (N¼ 58) and older than 50 (N ¼ 59). Whereas lower levels of liking 3 days later. In addition to these a priori older adults had significant amygdala modulation by preceding hypotheses, tested by calculating bivariate Pearson correlation picture valence [t(58) ¼ 2.75, P ¼ 0.008, d ¼ 0.36], younger adults coefficients, we also conducted exploratory whole-brain, voxel- showed no statistical difference [t(57) ¼ 0.79, P ¼ 0.44, d ¼ 0.10; wise correlation analyses to identify other regions in which acti- Figure 3]. Valence modulation of the amygdala was also reduced vation was associated with subsequent behavioral evaluations to non-significance when controlling for gender [t(115) ¼ 1.62, of neutral faces. P ¼ 0.11], although there were no gender differences in amyg- For each of these individual differences analyses, partici- dala signal [t(115)¼0.02, P ¼ 0.98]. The amygdala effect re- pants with behavioral data> 3 s.d. from the mean were mained significant when controlling for Sample status excluded as univariate outliers. In addition, for all effects identi- [t(115) ¼ 2.20, P ¼ 0.03]. fied as significant, the primary analysis (t test or correlation) was repeated as a linear regression model, controlling separ- Whole-brain responses to faces modulated by the ately for effects of age, gender, or Sample (Milwaukee/main valence of preceding pictures sample). Statistical maps for all whole-brain, voxelwise ana- lyses are available at http://neurovault.org/collections/2803/. We conducted whole-brain-corrected voxelwise searches to identify whether regions beyond the amygdala showed differ- ential responses to neutral faces based on preceding pictures’ Results valence or arousal. For the valence contrast, we observed Amygdala responses to neutral faces are modulated by greater activation for neutral faces following negative vs posi- the valence of preceding pictures tive pictures in the dorsomedial prefrontal cortex (dmPFC), left inferior frontal gyrus, left cerebellum, and regions typically During picture viewing, amygdala activation was modulated by involved in face processing including lateral fusiform gyri bilat- emotionally arousing pictures [0.5(Negativeþ Positive) > Neutral]; erally, right posterior superior temporal sulcus (STS), and bilat- t(116)¼ 5.94, P< 0.001, d¼ 0.55), and also by the valence of these eral angular gyri (Figure 4A and https://neurovault.org/images/ pictures [Negative> Positive; t(116)¼ 3.33, P¼ 0.0012, d¼ 0.31]. 52492/). The arousal contrast revealed enhanced activation in Whole-brain activation during picture viewing was generally con- occipital cortex for faces following emotionally arousing (nega- sistent with previous studies investigating affective picture viewing tive or positive vs neutral) pictures (Figure 4B and https://neuro (e.g. Lindquist et al.,2015). For example, the arousal contrast yielded vault.org/images/52494/). All results held when controlling for activation in ventromedial and dorsomedial PFC, lateral orbitofron- covariates of age, gender, or Sample, and no regions were iden- tal cortex and inferior frontal gyrus, amygdala, midbrain, pre- tified that showed effects in the opposite directions. cuneus and occipital cortices; and deactivation in lateral frontal poles, precuneus, and lateral occipital cortex (https://neurovault. Individual differences in affective style and liking of org/images/52490/). Greater activation for negative vs positive pic- neutral faces 3 days later tures was seen in lateral and inferior frontal gyri, midbrain, occipi- tal cortices into the ventral visual stream and cerebellum; greater On an average of 3 days after the fMRI session, participants pro- activation for positive vs negative pictures was observed in the vided memory and then liking ratings of the 45 faces viewed dur- middle frontal gyrus and midline regions, including medial PFC, ing the fMRI session, intermixed with 45 face foils. Participants posterior cingulate, precuneus and cuneus (https://neurovault.org/ were able to successfully discriminate previously seen from novel images/52489/). foil faces [t(104)¼ 12.48, P< 0.001, d¼ 1.22], and they rated Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 315 Fig. 3. Mean percent signal change values across the amygdala ROI for neutral faces presented following negative, neutral and positive pictures, for younger and older adults (based on a post hoc median split of the data). Older but not younger adults showed significant modulation of amygdala activation to faces based on the preced- ing valence of pictures, driven by a (non-significant) attenuation of amygdala activation to faces following positive pictures in older vs younger adults [t(115) ¼ 1.43, P ¼ 0.16]. Notes: Error bars represent 95% confidence intervals. *P < 0.01. Fig. 4. (A) Whole-brain voxelwise analysis for the Valence contrast revealed greater activation to neutral faces following negative vs positive pictures in the dorsomedial pre- frontal cortex, left inferior frontal gyrus and lateral fusiform gyri bilaterally, in addition to the right posterior STS, bilateral angular gyri and left cerebellum. Activation esti- mates from a set of finite impulse response basis functions are presented for these functionally defined clusters for illustrative purposes only. (B) Whole-brain voxelwise analysis for the Arousal contrast revealed greater activation to neutral faces following negative and positive (relative to neutral) pictures only in the occipital cortex. Maps are presented at P< 0.05, using cluster threshold correction. Full statistical maps can be viewed at http://neurovault.org/collections/2803/. previously seen faces as more likable than novel faces [Zajonc, participants completing paper ratings [mean6 s.d.¼ 0.066 0.08; 1968; t(103)¼ 7.33, P< 0.001, d¼ 0.72].There was a significant ef- t(103)¼ 4.21, P< 0.001]. No such effect was observed for likability fect of rating modality on memory ratings: the 89 participants ratings [t(102)¼ 0.53, P¼ 0.60]. Contrary to hypotheses, there were who completed ratings online had greater memory for old vs no main effects of preceding pictures’ valence or arousal on long- foil faces (mean difference6 s.d.¼ 0.256 0.18) than the 16 term memory or liking ratings [jtsj< 0.9, Ps> 0.4]. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 316 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 Fig. 5. (A) Individuals higher in trait positive affect on the Positive and Negative Affect Schedule (PANAS) showed greater arousal-modulated liking of faces viewed dur- ing the MRI session, such that faces initially preceded by negative or positive (relative to neutral) faces were rated as more likable when viewed 3 days later. (B) A similar relationship was observed for the Reappraisal subscale of the Emotion Regulation Questionnaire (ERQ). We next asked whether individual differences in self-report evaluations (liking or memory) of those faces 3 days later measures of affective style were associated with long-lasting af- (rsj< 0.11, Ps> 0.3). Exploratory analyses did not reveal any fective coloring of neutral faces. Faces preceded by emotionally other brain regions in which affectively colored activation to arousing (vs neutral) pictures were rated 3 days later as more like- neutral faces were associated with face ratings 3 days later able by individuals higher in trait positive affect [r(100)¼ 0.25, (https://neurovault.org/images/52495/ and https://neurovault. P¼ 0.012, uncorrected; Figure 5A] and emotional reappraisal org/images/52496/). [r(99)¼ 0.34, P¼< 0.001, Bonferroni-corrected a< 0.05; Figure 5B]. Collectively, these results show that short-term (neural) re- The relationship between reappraisal and liking still survived sponse to neutral faces are affectively colored by preceding Bonferroni correction (a< 0.05) when controlling for age, gender, emotional pictures, and that long-term (behavioral) indices of Sample and rating modality (paper vs online). The relationship affective coloring vary as a function of individual differences in with positive affect remained significant (not correcting for mul- trait affective style as well as emotionally modulated amygdala tiple comparisons) when controlling for age, gender, Sample and activation. rating modality (ts> 2.3, Ps< 0.03). Positive affect and reappraisal were significantly correlated with one another [r(113)¼ 0.23, Discussion P¼ 0.012], and only reappraisal accounted for unique variance in face liking in a simultaneous regression model [reappraisal: In a large and diverse sample of healthy adults, we have dem- t(98)¼ 3.01, P¼ 0.004; positive affect: t(99)¼ 1.73, P¼ 0.09]. onstrated an important role for individual differences in emo- Follow-up analyses showed that these effects were present tional responding on determining enduring behavioral effects for faces following both positive and negative (relative to neu- of affective coloring. We found that self-report measures of tral) pictures (positive affect: rs> 0.20, Ps< 0.05; reappraisal: positive affective style were associated with more positive ap- rs> 0.25, Ps< 0.02). No correlations were observed between lik- praisals of neutral faces that 3 days earlier had been presented ing ratings and trait negative affect or psychological well-being, in the wake of either positive or negative (relative to neutral) or between memory ratings and any of these four self-report pictures. We also found that greater amygdala responsivity for measures (jrsj< 0.14, Ps> 0.1). neutral faces appearing after emotional pictures was associated with increased memory for these faces 3 days later, an effect Individual differences in amygdala activation and that somewhat unexpectedly was present for positive but not memory for neutral faces 3 days later negative affective stimuli. Collectively, our results underscore the critical role of individual differences in determining how Consistent with predictions, individuals with increased amyg- emotions spill over onto unrelated social stimuli, thus coloring dala activation to faces presented following emotionally arous- our subsequent preferences and memory for these stimuli. ing pictures had greater memory for those same faces when Contrary to predictions, we observed no main effects of tested 3 days later [r(100) ¼ 0.21, P ¼ 0.039; Figure 6A]. Planned emotional pictures’ valence or arousal on behavioral evalu- follow-up analyses showed that this effect was driven by faces ations of the following neutral faces, assayed 3 days after the following positive vs neutral pictures [r(99) ¼ 0.21, P ¼ 0.036; fMRI scan, across the sample. Instead, our results indicate that Figure 6B], with no relationship for faces following negative vs certain individuals are more prone than others to the influence neutral pictures [r(100)¼0.03, P ¼ 0.78; Figure 6C]. The confi- of emotion stimuli on subsequent coloring. We found that indi- dence interval for the difference between these correlations vidual differences in questionnaire measures of positive affect- included zero (0.01, 0.48), indicating that the relationships be- ive style (emotional reappraisal and trait positive affect) were tween positive and negative conditions were not significantly robustly related to arousal-modulated liking of neutral faces different (Zou, 2007). Amygdala responses to faces following viewed 3 days after the fMRI scan. For individuals high in these positive pictures were associated with enhanced memory for dimensions, the arousal elicited by pictures may be appraised those faces when controlling for age, gender, Sample or rating modality (ts> 2.1, Ps< 0.04). In contrast, there was no relation- in a more positive manner due to differences in baseline emo- ship between differential amygdala responses to faces following tional traits (Schachter and Singer, 1962), and may thus negative vs positive pictures and subsequent behavioral have differential consequences for how other stimuli in the Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 317 Fig. 6. (A) Individuals with stronger amygdala responses to neutral faces following negative or positive (relative to neutral) pictures showed arousal-enhanced memory for neutral faces 3 days after the MRI session. This effect was specific to the positive valence condition (B), as amygdala responses for faces following positive (vs neu- tral) but not negative pictures (C) predicted greater memory for these same faces 3 days later. environment are processed—in this case, leading individuals to consider the possibility that altered affective coloring in psycho- form persistent positive appraisals of neutral faces encountered pathology may not reflect excessive coloring by negative events, in the wake of emotionally arousing pictures. but rather deficient coloring by positive events, or the absence Neurally, greater amygdala responses to neutral faces fol- of what speculatively may be a protective ‘positivity bias’ (cf. lowing more arousing pictures (in particular, positive pictures) Alloy and Abramson, 1979). were associated with increased memory for those faces when In addition to the long-term behavioral consequences of af- viewed 3 days later. This result provides a novel contribution to fective coloring, our experimental paradigm allows for the studies on the role of the amygdala in emotional memory, examination of short-term neural correlates of affective color- which have previously shown that amygdala activation at ing. We provide a novel theoretical contribution by demonstrat- encoding mediates the relationship between emotionally arous- ing that amygdala responses to neutral social stimuli were ing stimuli and increased memory encoding or retrieval for differentially modulated by the valence of preceding emotional those stimuli (Cahill et al., 1996; LaBar and Phelps, 1998; Canli pictures, with greater responses following negative vs positive et al., 2000; Kensinger and Corkin, 2004). Here, we demonstrate pictures. Results from our model-free approach corroborated that memory for neutral faces, presented shortly after positive those of primary analyses, and provided visual evidence that affective stimuli, is enhanced specifically in individuals who this effect resulted from differential increases in amygdala acti- show elevated amygdala responses to those neutral faces (see vation following neutral face onset (as opposed to a sustained also Tambini et al., 2017). It is notable that emotionally modu- response to emotional pictures themselves). This coloring of lated amygdala activation was linked to greater memory for amygdala responses was observed over a very short timescale, positive and not for negative stimuli. Similarly, an earlier report with neutral faces appearing only 2 s after picture offset. It found memory enhancement for neutral stimuli presented prior would be informative to probe the system at later time points to to positive stimuli, as opposed to a retrograde amnesic effect for track the persistence of this effect, and to vary the picture-face neutral stimuli presented before negative stimuli (Hurlemann ISI to probe individual differences in this persistence. This amygdala effect was no longer significant when control- et al., 2005). Indeed, emotionally enhanced memory of neutral stimuli is more frequently observed when these neutral stimuli ling for age, due to a (non-significant) increase in amygdala acti- precede rather than follow emotional stimuli (Knight and vation to faces following negative–positive pictures with Mather, 2009). increasing age. These findings may seem inconsistent with the Notably, we did not identify a main effect of memory en- literature on the positivity effect, which refers to biased informa- hancement or impairment for emotional relative to neutral tion processing in older adults for positive vs negative informa- stimuli; again, as with findings related to affective style, this tion. The positivity effect is consistently reflected in increased memory effect emerged purely as a function of our individual attention or memory for positive vs negative information (Reed differences analytical approach. If replicated and extended in et al., 2014); importantly, the relationship we observed between other samples, this finding may have important real-world con- amygdala responses to faces following positive pictures and sub- sequences: those individuals who show prolonged amygdala ac- sequent memory for these faces remained significant while con- tivation in the immediate aftermath of emotionally arousing trolling for age. In addition to biased information processing, events may be more apt to process unrelated (but temporally aging is also associated with relatively greater amygdala activa- proximal) neutral stimuli in the environment in a way that en- tion for positive vs negative stimuli (Mather, 2016). This literature hances encoding or later retrieval. Research into the temporal has focused primarily on amygdala activation during the process- dynamics of emotional responses may thus bear on our under- ing of affective stimuli, and does not speak to post-stimulus acti- standing of emotional memory and its disruption in conditions vation or the temporal dynamics of amygdala responding. such as posttraumatic stress disorder (Hayes et al., 2011) or de- Indeed, in a study using corrugator EMG to assess affective chron- pression (Burt et al., 1995; Mather et al., 2006). Depression in par- ometry, we previously identified slower recovery from negative ticular is increasingly linked to deficiencies in the experience of, IAPS stimuli in older women relative to younger women and men or ability to sustain, positive affect (Heller et al., 2009, 2013; (van Reekum et al., 2011). This underscores the possibility that the Kovacs et al., 2016). Based on our results, it is important to relationship between aging and differential responses to positive Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 318 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 vs negative information may be qualitatively different during the Acknowledgements processing of emotional stimuli vs after stimulus processing (and The authors thank Luke Hinsenkamp, Ben Hushek, Ayla for stimuli that appear during recovery from emotional stimuli), a Kruis, Phoebe Marquardt, Karthik Aroor, Nathan Vack, possibility that should be more fully explored in future work. Michael Anderle, Ron Fisher, and Scott Mikkelson for assist- Left lateral PFC and dorsomedial PFC regions also showed ance with data collection, processing, and analysis. enhanced activation to neutral faces following negative vs posi- tive pictures. These same regions are consistently engaged dur- ing the voluntary regulation of emotional responses (Buhle Funding et al., 2013, Figure 1), although in this study participants were This work was supported by the National Institutes of not given explicit emotion regulatory instructions. Notably, the Health (P01-AG020166 and R01-MH043454). magnitude of these regions’ inverse coupling with the amygdala was recently linked to short-term reductions in affective color- Conflict of interest. R.J.D. serves on the board of directors for ing behavior in the aftermath of fearful face viewing (Lapate the non-profit organization Healthy Minds Innovations. The et al., 2016). Moreover, these regions show significant age- other authors report no perceived or real conflicts of related declines in cortical thickness (Fjell and Walhovd, 2010) interest. and associated cognitive function (MacPherson et al., 2002). Further testing is warranted to explore whether age-associated changes in the structure and functionality of these prefrontal References regions may underlie the pronounced affective coloring effect Alloy, L.B., Abramson, L.Y. (1979). Judgment of contingency in observed in the amygdala of older participants. depressed and nondepressed students: sadder but wiser? Similar valence-modulated activation of neutral faces was Journal of Experimental Psychology, 108(4), 441–85. observed in the lateral fusiform gyrus and posterior STS, which Anderson, E., Siegel, E., White, D., Barrett, L.F. (2012). Out of sight compose a ‘core system’ for face processing (Haxby et al., 2000). but not out of mind: unseen affective faces influence evalu- Previous studies have demonstrated that these regions are ations and social impressions. Emotion, 12(6), 1210–21. involved in representing distinct facial expressions of emotion Beaupre, M.G., Hess, U. (2005). Cross-cultural emotion recogni- (Vuilleumier and Pourtois, 2007; Fox et al., 2009; Harry et al., tion among Canadian ethnic groups. Journal of Cross-Cultural 2013), thought to reflect the downstream influence of emotional Psychology, 36(3), 355–70. processing conducted in the amygdala (Vuilleumier et al., 2004; Blackford, J.U., Avery, S.N., Shelton, R.C., Zald, D.H. (2009). Haxby and Gobbini, 2011). The current findings support and ex- Amygdala temporal dynamics: temperamental differences in tend these previous studies and theoretical model, suggesting the timing of amygdala response to familiar and novel faces. that affective information may spill over from preceding and BMC Neurosciences, 10(1), 145. unrelated emotional events in a way that can color the percep- Brosschot, J.F., Gerin, W., Thayer, J.F., et al. (2006). The persevera- tion or evaluation of unrelated social stimuli. Future work tive cognition hypothesis: a review of worry, prolonged should explore the real-word impact of differential engagement stress-related physiological activation, and health. Journal of of this social-affective neural circuitry with regard to biased so- Psychosomatic Research, 60(2), 113–24. cial interactions or other behavioral outcomes. Buhle, J.T., Silvers, J.A., Wager, T.D., et al. (2013). Cognitive re- In summary, we have presented a novel fMRI paradigm that appraisal of emotion: a meta-analysis of human neuroimaging reveals the impact of affective pictures on subsequent neural studies. Cerebral Cortex, 24, 2981–90. responses to neutral social stimuli, and uncovers individual dif- Burke, H.M., Davis, M.C., Otte, C., et al. (2005). Depression and cor- ferences in affective style and amygdala activation that are tisol responses to psychological stress: a meta-analysis. related to enduring behavioral indices of affective coloring. In Psychoneuroendocrinology, 30(9), 846–56. light of relationships between trait-like differences in affective Burt, D.B., Zembar, M.J., Niederehe, G. (1995). Depression and style and risk for developing psychopathology (Fredrickson memory impairment: a meta-analysis of the association, its et al., 2003; Troy et al., 2010; Shackman et al., 2016), and alter- pattern, and specificity. Psychological Bulletin, 117(2), 285–305. ations in emotional memory and amygdala activation in condi- Cahill, L., Haier, R.J., Fallon, J., et al. (1996). Amygdala activity at tions such as depression and posttraumatic stress disorder encoding correlated with long-term, free recall of emotional (Mather et al., 2006; Hayes et al., 2011), these results suggest that information. Proceedings of the National Academy of Sciences of the an affective coloring framework may provide novel insight into United States of America, 93(15), 8016–21. neural and behavioral mechanisms of affective psychopath- Canli, T., Zhao, Z., Brewer, J., Gabrieli, J.D., Cahill, L. (2000). ology. This work paves the way for subsequent investigations in Event-related activation in the human amygdala associates affective disorders marked by alterations in affective chronom- with later memory for individual emotional experience. etry, such as truncated experiences of positive affect and slower Journal of Neuroscience, 20, RC99. [pii]. recovery from negative challenges, which will allow us to relate Cohen, L.H., Gunthert, K.C., Butler, A.C., Parrish, B.P., Wenze, S.J., these temporal dynamic differences to behavioral indices of Beck, J.S. (2008). Negative affective spillover from daily events emotional coloring and different symptom measures. predicts early response to cognitive therapy for depression. Journal of Consulting and Clinical Psychology, 76(6), 955–65. Author contributions Davidson, R.J. (1998). Affective style and affective disorders: per- spectives from affective neuroscience. Cognition and Emotion, S.M.S., R.C.L., and R.J.D. developed the study concept and con- tributed to study design. Testing and data collection was per- 12(3), 307–30. Davidson, R.J. (2004). Well-being and affective style: neural sub- formed by A.J.S. and L.K.G. D.W.G., S.M.S., R.C.L., and J.A.M. performed the data analysis and interpretation. D.W.G. drafted strates and biobehavioural correlates. Philosophical Transactions the manuscript, and all authors provided critical revisions and of the Royal Society of London Series B: Biological Sciences, approved the final version of the manuscript for submission. 359(1449), 1395–411. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 319 Knight, M., Mather, M. (2009). Reconciling findings of Desikan, R.S., Se ´ gonne, F., Fischl, B., et al. (2006). An automated labeling system for subdividing the human cerebral cortex on emotion-induced memory enhancement and impairment of MRI scans into gyral based regions of interest. Neuroimage, preceding items. Emotion, 9(6), 763–81. 31(3), 968–80. Kovacs, M., Bylsma, L.M., Yaroslavsky, I., et al. (2016). Positive af- Dichter, G.S., Tomarken, A.J. (2008). The chronometry of affective fectivity is dampened in youths with histories of major depres- startle modulation in unipolar depression. Journal of Abnormal sion and their never-depressed adolescent siblings. Clinical Psychology, 117(1), 1–15. Psychological Science, 4(4), 661–74. Fjell, A.M., Walhovd, K.B. (2010). Structural brain changes in LaBar, K.S., Phelps, E.A. (1998). Arousal-mediated memory con- aging: courses, causes and cognitive consequences. Reviews in solidation: role of the medial temporal lobe in humans. the Neurosciences, 21, 187–221. Psychological Sciences, 9(6), 490–3. Fox, C., Moon, S., Iaria, G., Barton, J. (2009). The correlates of sub- Lang, P.J., Bradley, M.M., Cuthbert, B.N. (2008). International jective perception of identity and expression in the face net- Affective Picture System (IAPS): Affective Ratings of Pictures and work: an fMRI adaptation study. Neuroimage, 44(2), 569–80. Instruction Manual. Technical Report A–8. University of Florida, Fredrickson, B.L., Tugade, M.M., Waugh, C.E., Larkin, G.R. (2003). Gainesville, FL. What good are positive emotions in crisis? A prospective study Lapate, R.C., Rokers, B., Tromp, D.P.M., et al. (2016). Awareness of of resilience and emotions following the terrorist attacks on emotional stimuli determines the behavioral consequences of the United States on September 11th, 2001. Journal of Personality amygdala activation and amygdala-prefrontal connectivity. and Social Psychology, 84(2), 365–76. Science Report, 6(1), 1–16. Frijda, N.H., Mesquita, B., Sonnemans, J., van Goozen, S. (1991). Lapate, R.C., Samaha, J., Rokers, B., Hamzah, H., Postle, B.R., The duration of affective phenomena or emotions, sentiments Davidson, R.J. (2017). Inhibition of lateral prefrontal cortex pro- and passions. In: Strongman, K.T., editor. International Review of duces emotionally biased first impressions: a transcranial Studies on Emotion, Chichester, UK: Wiley, 187–225. magnetic stimulation and electroencephalography study. Gross, J.J., John, O.P. (2003). Individual differences in two emotion Psychological Sciences, 28, 942–53. regulation processes: implications for affect, relationships, Lapate, R.C., van Reekum, C.M., Schaefer, S.M., et al. (2014). and well-being. Journal of Personality and Social Psychology, 85(2), Prolonged marital stress is associated with short-lived re- 348–62. sponses to positive stimuli. Psychophysiology, 51(6), 499–509. Gunthert, K.C., Cohen, L.H., Butler, A.C., Beck, J.S. (2007). Larson, C.L., Schaefer, H.S., Siegle, G.J., Jackson, C.A.B., Anderle, Depression and next-day spillover of negative mood and de- M.J., Davidson, R.J. (2006). Fear is fast in phobic individuals: pressive cognitions following interpersonal stress. Cognitive amygdala activation in response to fear-relevant stimuli. Therapy and Research, 31(4), 521–32. Biological Psychiatry, 60(4), 410–7. Harry, B., Williams, M.A., Davis, C., Kim, J. (2013). Emotional ex- Lindquist, K.A., Satpute, A.B., Wager, T.D., Weber, J., Barrett, L.F. pressions evoke a differential response in the fusiform face (2015). The brain basis of positive and negative affect: evidence area. Frontiers in Human Neuroscience, 7, 1–6. from a meta-analysis of the human neuroimaging literature. Haxby, J.V., Gobbini, M.I. (2011). Distributed neural systems for Cerebral Cortex, 26, 1910–22. face perception. In: Calder, A.J., Rhodes, G., Johnson, M.H., MacPherson, S.E., Phillips, L.H., Della Sala, S. (2002). Age, executive Haxby, J. V., editors. Oxford Handbook of Face Perception. New function, and social decision making: a dorsolateral prefrontal York: Oxford University Press. theory of cognitive aging. Psychology and Aging, 17(4), 598–609. Haxby, J.V., Hoffman, E.A., Gobbini, M.I. (2000). The distributed Mather, M. (2016). The affective neuroscience of aging. Annual human neural system for face perception. Trends in Cognitive Review Of Psychology, 67, 213–38. Mather, M., Mitchell, K.J., Raye, C.L., Novak, D.L., Greene, E.J., Science, 4(6), 223–33. Hayes, J.P., LaBar, K.S., McCarthy, G., et al. (2011). Reduced hippo- Johnson, M.K. (2006). Emotional arousal can impair feature campal and amygdala activity predicts memory distortions for binding in working memory. Journal of Cognitive Neuroscience, trauma reminders in combat-related PTSD. Journal of 18(4), 614–25. Psychiatric Research, 45(5), 660–9. Messer, K., Matas, J., Kittler, J., Luettin, J., Maitre, G. (1999). Heller, A.S., Fox, A.S., Wing, E.K., Mcquisition, K.M., Vack, N.J., XM2VTSDB: The extended M2VTS database, in: 2nd Davidson, R.J. (2015). The neurodynamics of affect in the la- International Conference on Audio Visual Based Person boratory predicts persistence of real-world emotional re- Authentication. sponses. Journal of Neuroscience, 35(29), 10503–9. Mugler, J.P., Brookeman, J.R. (1990). Three-dimensional Heller, A.S., Johnstone, T., Shackman, A.J., et al. (2009). Reduced magnetization-prepared rapid gradient-echo imaging (3D MP capacity to sustain positive emotion in major depression re- RAGE). Magnetic Resonance in Medicine, 15(1), 152–7. flects diminished maintenance of fronto-striatal brain activa- Mumford, J.A., Poline, J.B., Poldrack, R.A. (2015). Orthogonalization tion. Proceedings of the National Academy of Sciences of the United of regressors in fMRI models. PLoS ONE, 10(4), e0126255–11. Murphy, S.T., Zajonc, R.B. (1993). Affect, cognition, and aware- States of America, 106(52), 22445–50. Heller, A.S., van Reekum, C.M., Schaefer, S.M., et al. (2013). ness: affective priming with optimal and suboptimal stimulus Sustained striatal activity predicts eudaimonic well-being and exposures. Journal of Personality and Social Psychology, 64(5), cortisol output. Psychological Science, 24(11), 2191–200. 723–39. Hurlemann, R., Hawellek, B., Matusch, A., et al. (2005). Reed, A.E., Chan, L., Mikels, J.A. (2014). Meta-analysis of the Noradrenergic modulation of emotion-induced forgetting. age-related positivity effect: age differences in preferences for Journal of Neuroscience, 25, 6343–9. positive over negative information. Psychology and Aging, 29(1), Kensinger, E.A., Corkin, S. (2004). Two routes to emotional mem- 1–15. ory: distinct neural processes for valence and arousal. Ryff, C.D., Keyes, C.L. (1995). The structure of psychological Proceedings of the National Academy of Sciences of the United States well-being revisited. Journal of Personality and Social Psychology, of America, 101(9), 3310–5. 69(4), 719–27. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 320 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 Troy, A.S., Wilhelm, F.H., Shallcross, A.J., Mauss, I.B. (2010). Schachter, S., Singer, J.E. (1962). Cognitive, social, and physio- logical determinants of emotional state. Psychological Reviews, Seeing the silver lining: cognitive reappraisal ability moder- 69(5), 379–99. ates the relationship between stress and depressive symp- Schaefer, S.M., Boylan, J.M., Van Reekum, C.M., et al. (2013). toms. Emotion, 10(6), 783–95. Purpose in life predicts better emotional recovery from nega- van Reekum, C.M., Schaefer, S.M., Lapate, R.C., Norris, C.J., tive stimuli. PLoS ONE, 8(11), e80329. Greischar, L.L., Davidson, R.J. (2011). Aging is associated with Schuyler, B.S., Kral, T.R., Jacquart, J., et al. (2014). Temporal dy- positive responding to neutral information but reduced recov- namics of emotional responding: amygdala recovery predicts ery from negative information. Social Cognitive and Affective emotional traits. Social Cognitive and Affective Neuroscience, 9(2), Neuroscience, 6(2), 177–85. 176–81. Verduyn, P., Delaveau, P., Rotge ´ , J.-Y., Fossati, P., Van Mechelen, Shackman, A.J., Tromp, D.P.M., Stockbridge, M.D., Kaplan, C.M., I. (2015). Determinants of emotion duration and underlying Tillman, R.M., Fox, A.S. (2016). Dispositional negativity: an in- psychological and neural mechanisms. Emotion Reviews, 7(4), tegrative psychological and neurobiological perspective. 330–5. Psychological Bulletin, 142, 1275–314. Vuilleumier, P., Pourtois, G. (2007). Distributed and interactive brain mechanisms during emotion face perception: evi- Siegel, J.S., Power, J.D., Dubis, J.W., et al. (2014). Statistical im- provements in functional magnetic resonance imaging ana- dence from functional neuroimaging. Neuropsychologia, 45(1), lyses produced by censoring high-motion data points. Human 174–94. Brain Mapping, 35(5), 1981–96. Vuilleumier, P., Richardson, M.P., Armony, J.L., Driver, J., Dolan, Siegle, G.J., Steinhauer, S.R., Thase, M.E., Stenger, V.A., Carter, R.J. (2004). Distant influences of amygdala lesion on visual cor- C.S. (2002). Can’t shake that feeling: event-related fMRI assess- tical activation during emotional face processing. Nature ment of sustained amygdala activity in response to emotional Neuroscience, 7(11), 1271–8. information in depressed individuals. Biological Psychiatry, Watson, D., Clark, L.A., Tellegen, A. (1988). Development and val- 51(9), 693–707. idation of brief measures of positive and negative affect: the Tambini, A., Rimmele, U., Phelps, E.A., Davachi, L. (2017). PANAS scales. Journal of Personality and Social Psychology, 54(6), Emotional brain states carry over and enhance future memory 1063–70. formation. Nature Neuroscience, 20(2), 271. Zajonc, R.B. (1968). Attitudinal effects of mere exposure. Journal Tottenham, N., Tanaka, J.W., Leon, A.C., et al. (2009). The of Personality and Social Psychology, 9(2, Pt.2), 1–27. NimStim set of facial expressions: judgments from untrained Zou, G.Y. (2007). Toward using confidence intervals to compare research participants. Psychiatry Research, 168(3), 242–9. correlations. Psychological Methods, 12(4), 399–413. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Cognitive and Affective Neuroscience Oxford University Press

Behavioral and neural indices of affective coloring for neutral social stimuli

Download PDF
11 pages

Loading next page...
1
 
/lp/ou_press/behavioral-and-neural-indices-of-affective-coloring-for-neutral-social-kPX2gzBgQc

References (81)

Publisher
Oxford University Press
Copyright
© The Author(s) (2018). Published by Oxford University Press.
ISSN
1749-5016
eISSN
1749-5024
DOI
10.1093/scan/nsy011
pmid
29447377
Publisher site
See Article on Publisher Site

Abstract

Emotional processing often continues beyond the presentation of emotionally evocative stimuli, which can result in affective biasing or coloring of subsequently encountered events. Here, we describe neural correlates of affective coloring and examine how individual differences in affective style impact the magnitude of affective coloring. We conducted functional magnetic resonance imaging in 117 adults who passively viewed negative, neutral and positive pictures presented 2 s prior to neutral faces. Brain responses to neutral faces were modulated by the valence of preceding pictures, with greater activation for faces following negative (vs positive) pictures in the amygdala, dorsomedial and lateral prefrontal cortex, ventral visual cortices, posterior superior temporal sulcus, and angular gyrus. Three days after the magnetic resonance imaging scan, participants rated their memory and liking of previously encountered neutral faces. Individuals higher in trait positive affect and emotional reappraisal rated faces as more likable when preceded by emotionally arousing (negative or positive) pictures. In addition, greater amygdala responses to neutral faces preceded by positively valenced pictures were associated with greater memory for these faces 3 days later. Collectively, these results reveal individual differences in how emotions spill over onto the processing of unrelated social stimuli, resulting in persistent and affectively biased evaluations of such stimuli. Key words: affective coloring; amygdala; individual differences; faces; emotion in an intractable state of negative affect, which is subsequently Introduction reflected in overly critical evaluations of the statistics exams Emotional experiences are inherently ‘sticky’. Responses to she grades that evening. On the other hand, after having a affectively significant events (whether positive or negative) are manuscript accepted in a prestigious journal, the same student rarely temporally constrained to the time of exposure, but can may experience sustained positive affect that positively colors persist long after these events have passed (Frijda et al., 1991; her evaluations of the next stack of exams. Notably, under cer- Verduyn et al., 2015). Completely unrelated events, encountered tain circumstances such affective coloring has the potential to in the wake of these evocative emotional experiences, may be result in biased long-term evaluations of, or memory for, subse- processed through a particular affective filter, consequently quent unrelated events or interactions (Tambini et al., 2017; ‘coloring’ the perception or evaluation of unrelated events Lapate et al., 2017). While recent studies have begun to illumin- (Murphy and Zajonc, 1993; Anderson et al., 2012). For example, ate the neural correlates of affective coloring (Lapate et al., 2016, harsh feedback from a graduate student’s supervisor may result 2017; Tambini et al., 2017), it is unknown whether and how Received: 21 August 2017; Revised: 17 January 2018; Accepted: 9 February 2018 V C The Author(s) (2018). Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 311 inter-individual variability in neural mechanisms or trait-like All participants were right handed and were screened to ensure indices of affective style bear on affective coloring outcomes. MRI compatibility, and all provided informed consent prior to Individual differences in the temporal dynamics of emo- all study procedures. tional responding, or ‘affective chronometry’ (Davidson, 1998), may be critically important for the extent of such affective col- MRI data collection oring. It is now well established that individual differences in MRI data were collected on a 3 T scanner (MR750 GE Healthcare, affective chronometry are critical for both positive well-being Waukesha, WI) using an 8-channel head coil. We collected three (Davidson 2004; Heller et al. 2013; Schaefer et al. 2013) and risk of sets of echo planar images during the fMRI task (231 volumes, psychopathology (Siegle et al. 2002; Burke et al. 2005; Larson et al. TR ¼ 2000, TE ¼ 20, flip angle¼ 60 , field of view ¼ 220 mm, 9664 2006; Dichter and Tomarken 2008; Blackford et al. 2009; Lapate matrix, 3 mm slice thickness with 1 mm gap, 40 interleaved sa- et al. 2014; Schuyler et al. 2014; Heller et al. 2015) in addition to gittal slices, and ASSET parallel imaging with an acceleration negative physical health outcomes (Brosschot et al., 2006). factor of 2). A three-dimensional magnetization-prepared rapid Investigating affective coloring may afford insight into one gradient-echo sequence (Mugler and Brookeman, 1990) was pathway through which the temporal dynamics of emotion can used to acquire a T1-weighted anatomical image for functional impact psychological health: Individuals who quickly recover to data registration (TR ¼ 8.2, TE ¼ 3.2, flip angle ¼ 12 , field of a baseline state may exhibit minimal behavioral or biological in- view ¼ 256 mm, 256256 matrix, 160 axial slices, inversion time- dicators of spillover of emotion from one context to the next, ¼ 450 ms). The MRI session also included the acquisition of field whereas those who are slower to recover from negative events map images, a resting-state scan, diffusion-weighted data, and (or who are more capable of ‘savoring’ positive events) will be perfusion data. more susceptible to this spillover or coloring. For example, pre- vious studies utilizing daily diary methods have shown that while individuals with depression do not necessarily show Functional MRI task greater reactivity to stressful life events, the impact of interper- During fMRI scanning, participants passively viewed pictures sonal stressors ‘spills over’ into greater experiences of negative from the International Affective Picture System (IAPS; Lang affect on subsequent days (Gunthert et al., 2007), and that the et al., 2008), presented for 4 s (Figure 1). Each of three runs con- extent of this spillover predicts poorer response to cognitive be- tained 10 pictures each from negative, neutral and positive va- havioral therapy (Cohen et al., 2008). The laboratory investiga- lence categories, for a total of 90 picture presentations. Pictures tion of individual differences in affective coloring may thus were presented using a pseudorandom trial order with the re- provide novel mechanistic insight into affective pathology and quirement that no more than two pictures from the same va- its treatment in a more carefully controlled setting. lence category be presented in a row. Within each valence To that end, our lab has developed a functional magnetic category, stimuli were randomly selected for each participant. resonance imaging (fMRI) paradigm designed to measure the Each picture was followed by a 2 s inter-stimulus interval and a impact of well-characterized affective stimuli on immediate 0.5 s neutral face presentation. Participants were instructed to (neural) and long-term (behavioral) indices of affective coloring press one of two buttons to indicate the gender of the face, a (specifically, the coloring of neutral social stimuli; Figure 1). In low-level categorization task chosen to confirm task engage- this task, participants passively view negative, neutral and posi- ment without interfering with natural processing of the faces or tive pictures that are followed by a face displaying a neutral ex- preceding IAPS pictures. Faces were followed by an inter-trial pression, and are instructed to indicate the gender of the face (a interval between 3.5 and 27.5 s (mean duration ¼ 7.5 s). A 1 s low-level categorization task). To investigate affective coloring crosshair appeared before the start of each trial to orient partici- of neural responses, we compare brain responses to neutral pants’ attention. faces as a function of the preceding picture valence. To investi- This trial timing was chosen based on simulations conducted gate enduring behavioral indices of affective coloring, we pre- using optseq2 (https://surfer.nmr.mgh.harvard.edu/fswiki/opt sent the same faces to participants 3 days later and test the seq2), with the constraint that faces be presented within 4 s of extent to which memory or liking of these faces is modulated by picture presentation. This relatively brief ISI increased statistical the type of picture that initially preceded these faces. power to observe a spillover effect by maximizing the number of trials within the scan session. Specifically, we compared model efficiency for trial timings that included (i) average ISIs of 2 and Materials and methods 3 s, (ii) jittered ISIs of 0, 0.5 and 1 s and (iii) ITIs that varied (on Participants average) between 4.5 and 8 s. Based on these simulations, the fixed 2 s interval provided the greatest power to detect effects of Participants for this study were drawn from Midlife in the United States (MIDUS), a national longitudinal study of health interest during the face period. Because this fixed ISI may raise concerns regarding collinearity of picture and face regressors, we and well-being across the lifespan (http://www.midus.wisc. edu). For this study, we included participants from the MIDUS calculated the variance inflation factor (VIF) for each regressor. A ‘refresher’ sample who were enrolled beginning in 2012 in an ef- VIF cutoff of either 5 or 10 is typically used to indicate problem- atic levels of collinearity, and VIF values for the face regressors fort to expand and refresh the original MIDUS sample with younger age cohorts. The majority of the MIDUS refresher sam- fell close to 5 (mean for picture regressors¼ 4.22–4.72; mean for ple was recruited through random digit dialing of adults face regressors¼ 4.82–5.45). Although the VIF values for face throughout the U.S. The refresher sample also includes an over- regressors in some cases exceeded this more conservative sampling of African American participants from Milwaukee, WI threshold, collinearity generally does not affect the Type I error rate, but rather can lead to more variable estimates and thus who were recruited by door-to-door solicitation. fMRI task data were collected on a subset of 122 MIDUS refresher participants negatively impact statistical power (Mumford et al., 2015). In par- living in the Midwest who were able to travel to our laboratory, ticular, more variable estimates at the level of individual subjects including 35 participants from the Milwaukee subsample. will average out across subjects, and should thus stabilize with a Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 312 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 Fig. 1. (A) Schematic of fMRI trial structure. On each trial, participants passively viewed a 4 s picture from the IAPS (Lang et al., 2008). Following a 2 s delay, a neutral face was presented and participants pressed a button to indicate the gender of the face. (B) Three days following the scan, participants used an online or paper rating scale to provide memory then likability ratings for 45 neutral faces viewed during the fMRI scan and 45 novel neutral faces. larger sample size. In contrast, collinearity in higher-level ana- participants and two runs from one participant; for the other lyses (which was not the case in this study) may result in invalid 106 participants, we included data from all three runs in group- inferences regarding the magnitude of effect sizes (for discussion, level analyses. The first-level general linear model included six regressors see Mumford et al., 2015). Average normative arousal ratings were equivalent for nega- of interest, consisting of separate picture and face regressors for tive (5.466 0.66) and positive (5.476 0.53) IAPS pictures, which each picture valence (Negative/Neutral/Positive). Regressors of were both greater than those for neutral pictures (3.166 0.42; no interest included mean-centered response times to face ts> 16.1, Ps< 0.001). Pictures within each valence category were presentations, six motion parameters with their first- and matched on luminosity and visual complexity, and for the num- second-order derivatives, and a confound regressor for each ber of pictures rated to be social in content. Faces were drawn volume with> .9 mm framewise displacement (Siegel et al., from the XM2VTSDB multi-modal face database (Messer et al., 2014). Additional processing steps included correction of time 1999), the NimStim database (Tottenham et al., 2009), and the series autocorrelation, resampling of functional data to 2 mm Montreal Set of Facial Displays of Emotion (Beaupre et al., 2005). isotropic voxels, and registration to Montreal Neurological In light of our diverse subject population, these faces included Institute template space. equal proportions of male and females with a broad range of As a complementary approach, we modeled amygdala time ages and ethnicities. Faces were cropped just above the hair and series data using a set of seven finite impulse response (FIR) below the chin, converted to black and white, and edited (e.g. to basis functions. This allowed for model-free estimation of the remove distinctive facial hair and eyeglasses). A total of 45 neu- hemodynamic response from the time of picture onset through tral faces were presented, and each face was presented follow- the first 14 s of each trial (corresponding to the minimum trial ing two randomly selected pictures from the same valence length). We then extracted empirically derived activation esti- category. mates for each trial type separately. For group-level analyses we averaged these estimates across individuals at each volume, calculated valence and arousal contrast estimates at each time FMRI data processing and analysis point, and conducted paired-sample t tests for these contrasts. FMRI data processing was carried out using FEAT (FMRI Expert Analysis Tool) Version 6.00, part of FSL (FMRIB’s Software Group analysis of fMRI data Library, www.fmrib.ox.ac.uk/fsl). Preprocessing steps included removal of the first four volumes, motion correction using Primary group analyses for the fMRI data were conducted using MCFLIRT, removal of non-brain regions using BET, spatial two orthogonal contrasts for both the picture and face epochs: a smoothing using a Gaussian kernel with 5 mm full-width at ‘Valence’ contrast (Negative vs Positive) and an ‘Arousal’ con- half-maximum, grand-mean intensity normalization, and high- trast [0.5*(Negative þ Positive) Neutral]. For significant effects pass temporal filtering. We excluded five participants following involving the Arousal contrast, follow-up analyses examined whether effects were driven more by the Negative or Positive first-level data processing (1 for excessive movement [>1mm mean framewise displacement across the entire run], 2 for fail- condition. For primary analyses focused on the amygdala, we ing to provide behavioral responses, and 2 for distorted func- extracted mean parameter estimates for each of these contrasts tional data), resulting in a final sample size of 117 for fMRI from voxels with  50% probability of being assigned to the bi- analyses (64 female/53 male; Milwaukee subsample N¼ 35; lateral amygdala from the Harvard–Oxford probabilistic atlas mean age6 s.d. ¼ 48.36 11.8; range ¼ 26–76). In addition, due to (volume ¼ 4040 mm ; Desikan et al., 2006; Figure 2A), which was excessive motion and/or failure to provide behavioral responses chosen as an a priori ROI prior to data analysis. In addition to on at least 25% of trials, we excluded one run from 10 this a priori region of interest, we conducted whole-brain Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 313 Fig. 2. (A) Anatomically defined amygdala region of interest (ROI). (B) Mean percent signal change values across the amygdala ROI for neutral faces presented following negative, neutral and positive pictures. (C) Estimated signal from the amygdala ROI using a set of finite impulse response basis functions revealed complementary evi- dence for greater amygdala activation to faces following negative vs positive pictures. Notes: Error bars represent 95% confidence intervals. *P < 0.05. voxelwise analyses on the same two contrasts for the picture with higher values indicating greater liking of, or greater confi- and face periods. Cluster threshold correction was applied using dence of having previously seen, each face. These ratings were a Z-threshold of 3.09 and a cluster-corrected significance value completed exactly 3 days after the scan in 85 out of 89 subjects of P< 0.05. who completed ratings online (two participants completed rat- ings 4 days after the scan, and one each completed ratings 7 and 9 days after the scan). All 16 participants completing paper rat- Behavioral data collection and analysis ings reported completing these exactly 3 days after the scan. Participants were asked to log onto the internet 3 days after the Analysis of these behavioral data tested the extent to which MRI session, and complete ratings of the 45 faces they viewed each participant’s long-term memory or liking of neutral faces during the MRI task, which were interspersed with 45 novel was colored by the content of preceding IAPS pictures from the faces. Each of these 90 faces was presented for an unlimited scan session. We conducted paired t tests on these behavioral duration with one of two continuous rating scales that asked data using the same contrasts as with the fMRI data in order to participants to rate their memory and liking of these faces. The test whether the valence or arousal of pictures had a persistent memory scale had equidistant anchor points labeled ‘com- impact on behavioral evaluations of subsequently presented pletely certain new’, ‘somewhat certain new’, ‘unsure’, ‘some- faces. We predicted that (i) memory for neutral faces would be what certain old’ and ‘completely certain old’; the liking scale enhanced for faces following emotionally arousing pictures had equidistant anchor points labeled ‘really dislike’, ‘some- (Negative or Positive, relative to Neutral), and (ii) liking of neu- what dislike’, ‘unsure’, ‘somewhat like’ and ‘really like’. Memory tral faces would be enhanced for faces following Positive rela- ratings were always completed prior to liking ratings. tive to Negative pictures. Memory and liking ratings were obtained from 105 and 104 participants, respectively. Of these participants, 16 were not Individual differences analyses comfortable using the online rating scale or did not have access to a computer; these individuals instead completed ratings We conducted analyses to explore the impact of individual dif- using paper packets. Ratings collected electronically and on ferences in both affective style and amygdala activation on the paper were both converted to values ranging between 1 and 1, affective coloring of behavior, as assayed by 3-day liking and Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 314 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 memory ratings. We examined four self-report measures of af- Amygdala responses to neutral faces were not modulated by fective style: state Positive and Negative Affect from the Positive the arousal contrast for preceding pictures [t(116)¼0.33, and Negative Affect Schedule (Watson et al., 1988); the P> 0.7]. However, amygdala responses to neutral faces were Reappraisal subscale of the Emotion Regulation Questionnaire modulated by the valence of preceding pictures, with greater ac- tivation when the preceding picture was negative vs positive (Gross and John, 2003); and total score on the Psychological [t(116) ¼ 2.39, P ¼ 0.027, d ¼ 0.22; Figure 2B]. We also estimated Well-Being Scale (Ryff and Keyes, 1995). For each of these four indices we calculated Pearson correlation coefficients with sub- amygdala responses directly from the data using a set of FIR basis functions (Figure 2C). This model-free approach revealed sequent ratings of face liking and memory, both as a function of clearly distinct hemodynamic responses in the amygdala to the preceding pictures’ Valence and Arousal. We corrected for mul- picture and face epochs, with peak responses three volumes tiple comparisons testing using a conservative Bonferroni- (6 s) after picture onset and two to three volumes (4–6 s) after corrected P value of 0.003125 [4 measures2 behavioral ratings face onset. Consistent with our primary analysis, we observed (liking/memory)2 contrasts (Valence/Arousal)]. greater amygdala activation for the negative vs positive condi- We also tested two specific hypotheses regarding relation- tion at volumes 5 [t(116) ¼ 2.76, P ¼ 0.0067, d ¼ 0.27] and 7 ships between individual differences in affectively-modulated [t(116) ¼ 3.68, P< 0.001, d ¼ 0.39], i.e. corresponding to 4 and 8 s, amygdala responses to neutral faces and long-lasting behav- respectively, after face onset. Results of the whole-brain, voxel- ioral evaluations of these faces. First, as an extension of find- wise FIR analysis for all volumes are available at https://neuro ings that the amygdala mediates arousal-enhanced memory for vault.org/collections/3154/. emotional stimuli (Canli et al., 2000; Kensinger and Corkin, Notably, when including age as a covariate, we no longer 2004), we hypothesized that amygdala responses to faces follow- observed a significant group effect of Valence [t(112) ¼ 1.24, ing arousing pictures would be associated with greater memory P ¼ 0.22]. Indeed, age showed a marginally significant correl- for those faces 3 days later. Second, we correlated amygdala re- ation with amygdala activation for the Valence contrast sponses to faces following Negative vs Positive pictures [r(115) ¼ 0.17, P ¼ 0.064], with greater (Negative> Positive) amyg- (Valence contrast) with liking ratings, with the prediction that dala activation in older adults. To visualize this age effect, we more negative affective coloring of faces (as manifest in stron- conducted a post hoc median split of the sample into adults ger amygdala responses) would be associated with relatively younger than 50 (N¼ 58) and older than 50 (N ¼ 59). Whereas lower levels of liking 3 days later. In addition to these a priori older adults had significant amygdala modulation by preceding hypotheses, tested by calculating bivariate Pearson correlation picture valence [t(58) ¼ 2.75, P ¼ 0.008, d ¼ 0.36], younger adults coefficients, we also conducted exploratory whole-brain, voxel- showed no statistical difference [t(57) ¼ 0.79, P ¼ 0.44, d ¼ 0.10; wise correlation analyses to identify other regions in which acti- Figure 3]. Valence modulation of the amygdala was also reduced vation was associated with subsequent behavioral evaluations to non-significance when controlling for gender [t(115) ¼ 1.62, of neutral faces. P ¼ 0.11], although there were no gender differences in amyg- For each of these individual differences analyses, partici- dala signal [t(115)¼0.02, P ¼ 0.98]. The amygdala effect re- pants with behavioral data> 3 s.d. from the mean were mained significant when controlling for Sample status excluded as univariate outliers. In addition, for all effects identi- [t(115) ¼ 2.20, P ¼ 0.03]. fied as significant, the primary analysis (t test or correlation) was repeated as a linear regression model, controlling separ- Whole-brain responses to faces modulated by the ately for effects of age, gender, or Sample (Milwaukee/main valence of preceding pictures sample). Statistical maps for all whole-brain, voxelwise ana- lyses are available at http://neurovault.org/collections/2803/. We conducted whole-brain-corrected voxelwise searches to identify whether regions beyond the amygdala showed differ- ential responses to neutral faces based on preceding pictures’ Results valence or arousal. For the valence contrast, we observed Amygdala responses to neutral faces are modulated by greater activation for neutral faces following negative vs posi- the valence of preceding pictures tive pictures in the dorsomedial prefrontal cortex (dmPFC), left inferior frontal gyrus, left cerebellum, and regions typically During picture viewing, amygdala activation was modulated by involved in face processing including lateral fusiform gyri bilat- emotionally arousing pictures [0.5(Negativeþ Positive) > Neutral]; erally, right posterior superior temporal sulcus (STS), and bilat- t(116)¼ 5.94, P< 0.001, d¼ 0.55), and also by the valence of these eral angular gyri (Figure 4A and https://neurovault.org/images/ pictures [Negative> Positive; t(116)¼ 3.33, P¼ 0.0012, d¼ 0.31]. 52492/). The arousal contrast revealed enhanced activation in Whole-brain activation during picture viewing was generally con- occipital cortex for faces following emotionally arousing (nega- sistent with previous studies investigating affective picture viewing tive or positive vs neutral) pictures (Figure 4B and https://neuro (e.g. Lindquist et al.,2015). For example, the arousal contrast yielded vault.org/images/52494/). All results held when controlling for activation in ventromedial and dorsomedial PFC, lateral orbitofron- covariates of age, gender, or Sample, and no regions were iden- tal cortex and inferior frontal gyrus, amygdala, midbrain, pre- tified that showed effects in the opposite directions. cuneus and occipital cortices; and deactivation in lateral frontal poles, precuneus, and lateral occipital cortex (https://neurovault. Individual differences in affective style and liking of org/images/52490/). Greater activation for negative vs positive pic- neutral faces 3 days later tures was seen in lateral and inferior frontal gyri, midbrain, occipi- tal cortices into the ventral visual stream and cerebellum; greater On an average of 3 days after the fMRI session, participants pro- activation for positive vs negative pictures was observed in the vided memory and then liking ratings of the 45 faces viewed dur- middle frontal gyrus and midline regions, including medial PFC, ing the fMRI session, intermixed with 45 face foils. Participants posterior cingulate, precuneus and cuneus (https://neurovault.org/ were able to successfully discriminate previously seen from novel images/52489/). foil faces [t(104)¼ 12.48, P< 0.001, d¼ 1.22], and they rated Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 315 Fig. 3. Mean percent signal change values across the amygdala ROI for neutral faces presented following negative, neutral and positive pictures, for younger and older adults (based on a post hoc median split of the data). Older but not younger adults showed significant modulation of amygdala activation to faces based on the preced- ing valence of pictures, driven by a (non-significant) attenuation of amygdala activation to faces following positive pictures in older vs younger adults [t(115) ¼ 1.43, P ¼ 0.16]. Notes: Error bars represent 95% confidence intervals. *P < 0.01. Fig. 4. (A) Whole-brain voxelwise analysis for the Valence contrast revealed greater activation to neutral faces following negative vs positive pictures in the dorsomedial pre- frontal cortex, left inferior frontal gyrus and lateral fusiform gyri bilaterally, in addition to the right posterior STS, bilateral angular gyri and left cerebellum. Activation esti- mates from a set of finite impulse response basis functions are presented for these functionally defined clusters for illustrative purposes only. (B) Whole-brain voxelwise analysis for the Arousal contrast revealed greater activation to neutral faces following negative and positive (relative to neutral) pictures only in the occipital cortex. Maps are presented at P< 0.05, using cluster threshold correction. Full statistical maps can be viewed at http://neurovault.org/collections/2803/. previously seen faces as more likable than novel faces [Zajonc, participants completing paper ratings [mean6 s.d.¼ 0.066 0.08; 1968; t(103)¼ 7.33, P< 0.001, d¼ 0.72].There was a significant ef- t(103)¼ 4.21, P< 0.001]. No such effect was observed for likability fect of rating modality on memory ratings: the 89 participants ratings [t(102)¼ 0.53, P¼ 0.60]. Contrary to hypotheses, there were who completed ratings online had greater memory for old vs no main effects of preceding pictures’ valence or arousal on long- foil faces (mean difference6 s.d.¼ 0.256 0.18) than the 16 term memory or liking ratings [jtsj< 0.9, Ps> 0.4]. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 316 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 Fig. 5. (A) Individuals higher in trait positive affect on the Positive and Negative Affect Schedule (PANAS) showed greater arousal-modulated liking of faces viewed dur- ing the MRI session, such that faces initially preceded by negative or positive (relative to neutral) faces were rated as more likable when viewed 3 days later. (B) A similar relationship was observed for the Reappraisal subscale of the Emotion Regulation Questionnaire (ERQ). We next asked whether individual differences in self-report evaluations (liking or memory) of those faces 3 days later measures of affective style were associated with long-lasting af- (rsj< 0.11, Ps> 0.3). Exploratory analyses did not reveal any fective coloring of neutral faces. Faces preceded by emotionally other brain regions in which affectively colored activation to arousing (vs neutral) pictures were rated 3 days later as more like- neutral faces were associated with face ratings 3 days later able by individuals higher in trait positive affect [r(100)¼ 0.25, (https://neurovault.org/images/52495/ and https://neurovault. P¼ 0.012, uncorrected; Figure 5A] and emotional reappraisal org/images/52496/). [r(99)¼ 0.34, P¼< 0.001, Bonferroni-corrected a< 0.05; Figure 5B]. Collectively, these results show that short-term (neural) re- The relationship between reappraisal and liking still survived sponse to neutral faces are affectively colored by preceding Bonferroni correction (a< 0.05) when controlling for age, gender, emotional pictures, and that long-term (behavioral) indices of Sample and rating modality (paper vs online). The relationship affective coloring vary as a function of individual differences in with positive affect remained significant (not correcting for mul- trait affective style as well as emotionally modulated amygdala tiple comparisons) when controlling for age, gender, Sample and activation. rating modality (ts> 2.3, Ps< 0.03). Positive affect and reappraisal were significantly correlated with one another [r(113)¼ 0.23, Discussion P¼ 0.012], and only reappraisal accounted for unique variance in face liking in a simultaneous regression model [reappraisal: In a large and diverse sample of healthy adults, we have dem- t(98)¼ 3.01, P¼ 0.004; positive affect: t(99)¼ 1.73, P¼ 0.09]. onstrated an important role for individual differences in emo- Follow-up analyses showed that these effects were present tional responding on determining enduring behavioral effects for faces following both positive and negative (relative to neu- of affective coloring. We found that self-report measures of tral) pictures (positive affect: rs> 0.20, Ps< 0.05; reappraisal: positive affective style were associated with more positive ap- rs> 0.25, Ps< 0.02). No correlations were observed between lik- praisals of neutral faces that 3 days earlier had been presented ing ratings and trait negative affect or psychological well-being, in the wake of either positive or negative (relative to neutral) or between memory ratings and any of these four self-report pictures. We also found that greater amygdala responsivity for measures (jrsj< 0.14, Ps> 0.1). neutral faces appearing after emotional pictures was associated with increased memory for these faces 3 days later, an effect Individual differences in amygdala activation and that somewhat unexpectedly was present for positive but not memory for neutral faces 3 days later negative affective stimuli. Collectively, our results underscore the critical role of individual differences in determining how Consistent with predictions, individuals with increased amyg- emotions spill over onto unrelated social stimuli, thus coloring dala activation to faces presented following emotionally arous- our subsequent preferences and memory for these stimuli. ing pictures had greater memory for those same faces when Contrary to predictions, we observed no main effects of tested 3 days later [r(100) ¼ 0.21, P ¼ 0.039; Figure 6A]. Planned emotional pictures’ valence or arousal on behavioral evalu- follow-up analyses showed that this effect was driven by faces ations of the following neutral faces, assayed 3 days after the following positive vs neutral pictures [r(99) ¼ 0.21, P ¼ 0.036; fMRI scan, across the sample. Instead, our results indicate that Figure 6B], with no relationship for faces following negative vs certain individuals are more prone than others to the influence neutral pictures [r(100)¼0.03, P ¼ 0.78; Figure 6C]. The confi- of emotion stimuli on subsequent coloring. We found that indi- dence interval for the difference between these correlations vidual differences in questionnaire measures of positive affect- included zero (0.01, 0.48), indicating that the relationships be- ive style (emotional reappraisal and trait positive affect) were tween positive and negative conditions were not significantly robustly related to arousal-modulated liking of neutral faces different (Zou, 2007). Amygdala responses to faces following viewed 3 days after the fMRI scan. For individuals high in these positive pictures were associated with enhanced memory for dimensions, the arousal elicited by pictures may be appraised those faces when controlling for age, gender, Sample or rating modality (ts> 2.1, Ps< 0.04). In contrast, there was no relation- in a more positive manner due to differences in baseline emo- ship between differential amygdala responses to faces following tional traits (Schachter and Singer, 1962), and may thus negative vs positive pictures and subsequent behavioral have differential consequences for how other stimuli in the Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 317 Fig. 6. (A) Individuals with stronger amygdala responses to neutral faces following negative or positive (relative to neutral) pictures showed arousal-enhanced memory for neutral faces 3 days after the MRI session. This effect was specific to the positive valence condition (B), as amygdala responses for faces following positive (vs neu- tral) but not negative pictures (C) predicted greater memory for these same faces 3 days later. environment are processed—in this case, leading individuals to consider the possibility that altered affective coloring in psycho- form persistent positive appraisals of neutral faces encountered pathology may not reflect excessive coloring by negative events, in the wake of emotionally arousing pictures. but rather deficient coloring by positive events, or the absence Neurally, greater amygdala responses to neutral faces fol- of what speculatively may be a protective ‘positivity bias’ (cf. lowing more arousing pictures (in particular, positive pictures) Alloy and Abramson, 1979). were associated with increased memory for those faces when In addition to the long-term behavioral consequences of af- viewed 3 days later. This result provides a novel contribution to fective coloring, our experimental paradigm allows for the studies on the role of the amygdala in emotional memory, examination of short-term neural correlates of affective color- which have previously shown that amygdala activation at ing. We provide a novel theoretical contribution by demonstrat- encoding mediates the relationship between emotionally arous- ing that amygdala responses to neutral social stimuli were ing stimuli and increased memory encoding or retrieval for differentially modulated by the valence of preceding emotional those stimuli (Cahill et al., 1996; LaBar and Phelps, 1998; Canli pictures, with greater responses following negative vs positive et al., 2000; Kensinger and Corkin, 2004). Here, we demonstrate pictures. Results from our model-free approach corroborated that memory for neutral faces, presented shortly after positive those of primary analyses, and provided visual evidence that affective stimuli, is enhanced specifically in individuals who this effect resulted from differential increases in amygdala acti- show elevated amygdala responses to those neutral faces (see vation following neutral face onset (as opposed to a sustained also Tambini et al., 2017). It is notable that emotionally modu- response to emotional pictures themselves). This coloring of lated amygdala activation was linked to greater memory for amygdala responses was observed over a very short timescale, positive and not for negative stimuli. Similarly, an earlier report with neutral faces appearing only 2 s after picture offset. It found memory enhancement for neutral stimuli presented prior would be informative to probe the system at later time points to to positive stimuli, as opposed to a retrograde amnesic effect for track the persistence of this effect, and to vary the picture-face neutral stimuli presented before negative stimuli (Hurlemann ISI to probe individual differences in this persistence. This amygdala effect was no longer significant when control- et al., 2005). Indeed, emotionally enhanced memory of neutral stimuli is more frequently observed when these neutral stimuli ling for age, due to a (non-significant) increase in amygdala acti- precede rather than follow emotional stimuli (Knight and vation to faces following negative–positive pictures with Mather, 2009). increasing age. These findings may seem inconsistent with the Notably, we did not identify a main effect of memory en- literature on the positivity effect, which refers to biased informa- hancement or impairment for emotional relative to neutral tion processing in older adults for positive vs negative informa- stimuli; again, as with findings related to affective style, this tion. The positivity effect is consistently reflected in increased memory effect emerged purely as a function of our individual attention or memory for positive vs negative information (Reed differences analytical approach. If replicated and extended in et al., 2014); importantly, the relationship we observed between other samples, this finding may have important real-world con- amygdala responses to faces following positive pictures and sub- sequences: those individuals who show prolonged amygdala ac- sequent memory for these faces remained significant while con- tivation in the immediate aftermath of emotionally arousing trolling for age. In addition to biased information processing, events may be more apt to process unrelated (but temporally aging is also associated with relatively greater amygdala activa- proximal) neutral stimuli in the environment in a way that en- tion for positive vs negative stimuli (Mather, 2016). This literature hances encoding or later retrieval. Research into the temporal has focused primarily on amygdala activation during the process- dynamics of emotional responses may thus bear on our under- ing of affective stimuli, and does not speak to post-stimulus acti- standing of emotional memory and its disruption in conditions vation or the temporal dynamics of amygdala responding. such as posttraumatic stress disorder (Hayes et al., 2011) or de- Indeed, in a study using corrugator EMG to assess affective chron- pression (Burt et al., 1995; Mather et al., 2006). Depression in par- ometry, we previously identified slower recovery from negative ticular is increasingly linked to deficiencies in the experience of, IAPS stimuli in older women relative to younger women and men or ability to sustain, positive affect (Heller et al., 2009, 2013; (van Reekum et al., 2011). This underscores the possibility that the Kovacs et al., 2016). Based on our results, it is important to relationship between aging and differential responses to positive Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 318 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 vs negative information may be qualitatively different during the Acknowledgements processing of emotional stimuli vs after stimulus processing (and The authors thank Luke Hinsenkamp, Ben Hushek, Ayla for stimuli that appear during recovery from emotional stimuli), a Kruis, Phoebe Marquardt, Karthik Aroor, Nathan Vack, possibility that should be more fully explored in future work. Michael Anderle, Ron Fisher, and Scott Mikkelson for assist- Left lateral PFC and dorsomedial PFC regions also showed ance with data collection, processing, and analysis. enhanced activation to neutral faces following negative vs posi- tive pictures. These same regions are consistently engaged dur- ing the voluntary regulation of emotional responses (Buhle Funding et al., 2013, Figure 1), although in this study participants were This work was supported by the National Institutes of not given explicit emotion regulatory instructions. Notably, the Health (P01-AG020166 and R01-MH043454). magnitude of these regions’ inverse coupling with the amygdala was recently linked to short-term reductions in affective color- Conflict of interest. R.J.D. serves on the board of directors for ing behavior in the aftermath of fearful face viewing (Lapate the non-profit organization Healthy Minds Innovations. The et al., 2016). Moreover, these regions show significant age- other authors report no perceived or real conflicts of related declines in cortical thickness (Fjell and Walhovd, 2010) interest. and associated cognitive function (MacPherson et al., 2002). Further testing is warranted to explore whether age-associated changes in the structure and functionality of these prefrontal References regions may underlie the pronounced affective coloring effect Alloy, L.B., Abramson, L.Y. (1979). Judgment of contingency in observed in the amygdala of older participants. depressed and nondepressed students: sadder but wiser? Similar valence-modulated activation of neutral faces was Journal of Experimental Psychology, 108(4), 441–85. observed in the lateral fusiform gyrus and posterior STS, which Anderson, E., Siegel, E., White, D., Barrett, L.F. (2012). Out of sight compose a ‘core system’ for face processing (Haxby et al., 2000). but not out of mind: unseen affective faces influence evalu- Previous studies have demonstrated that these regions are ations and social impressions. Emotion, 12(6), 1210–21. involved in representing distinct facial expressions of emotion Beaupre, M.G., Hess, U. (2005). Cross-cultural emotion recogni- (Vuilleumier and Pourtois, 2007; Fox et al., 2009; Harry et al., tion among Canadian ethnic groups. Journal of Cross-Cultural 2013), thought to reflect the downstream influence of emotional Psychology, 36(3), 355–70. processing conducted in the amygdala (Vuilleumier et al., 2004; Blackford, J.U., Avery, S.N., Shelton, R.C., Zald, D.H. (2009). Haxby and Gobbini, 2011). The current findings support and ex- Amygdala temporal dynamics: temperamental differences in tend these previous studies and theoretical model, suggesting the timing of amygdala response to familiar and novel faces. that affective information may spill over from preceding and BMC Neurosciences, 10(1), 145. unrelated emotional events in a way that can color the percep- Brosschot, J.F., Gerin, W., Thayer, J.F., et al. (2006). The persevera- tion or evaluation of unrelated social stimuli. Future work tive cognition hypothesis: a review of worry, prolonged should explore the real-word impact of differential engagement stress-related physiological activation, and health. Journal of of this social-affective neural circuitry with regard to biased so- Psychosomatic Research, 60(2), 113–24. cial interactions or other behavioral outcomes. Buhle, J.T., Silvers, J.A., Wager, T.D., et al. (2013). Cognitive re- In summary, we have presented a novel fMRI paradigm that appraisal of emotion: a meta-analysis of human neuroimaging reveals the impact of affective pictures on subsequent neural studies. Cerebral Cortex, 24, 2981–90. responses to neutral social stimuli, and uncovers individual dif- Burke, H.M., Davis, M.C., Otte, C., et al. (2005). Depression and cor- ferences in affective style and amygdala activation that are tisol responses to psychological stress: a meta-analysis. related to enduring behavioral indices of affective coloring. In Psychoneuroendocrinology, 30(9), 846–56. light of relationships between trait-like differences in affective Burt, D.B., Zembar, M.J., Niederehe, G. (1995). Depression and style and risk for developing psychopathology (Fredrickson memory impairment: a meta-analysis of the association, its et al., 2003; Troy et al., 2010; Shackman et al., 2016), and alter- pattern, and specificity. Psychological Bulletin, 117(2), 285–305. ations in emotional memory and amygdala activation in condi- Cahill, L., Haier, R.J., Fallon, J., et al. (1996). Amygdala activity at tions such as depression and posttraumatic stress disorder encoding correlated with long-term, free recall of emotional (Mather et al., 2006; Hayes et al., 2011), these results suggest that information. Proceedings of the National Academy of Sciences of the an affective coloring framework may provide novel insight into United States of America, 93(15), 8016–21. neural and behavioral mechanisms of affective psychopath- Canli, T., Zhao, Z., Brewer, J., Gabrieli, J.D., Cahill, L. (2000). ology. This work paves the way for subsequent investigations in Event-related activation in the human amygdala associates affective disorders marked by alterations in affective chronom- with later memory for individual emotional experience. etry, such as truncated experiences of positive affect and slower Journal of Neuroscience, 20, RC99. [pii]. recovery from negative challenges, which will allow us to relate Cohen, L.H., Gunthert, K.C., Butler, A.C., Parrish, B.P., Wenze, S.J., these temporal dynamic differences to behavioral indices of Beck, J.S. (2008). Negative affective spillover from daily events emotional coloring and different symptom measures. predicts early response to cognitive therapy for depression. Journal of Consulting and Clinical Psychology, 76(6), 955–65. Author contributions Davidson, R.J. (1998). Affective style and affective disorders: per- spectives from affective neuroscience. Cognition and Emotion, S.M.S., R.C.L., and R.J.D. developed the study concept and con- tributed to study design. Testing and data collection was per- 12(3), 307–30. Davidson, R.J. (2004). Well-being and affective style: neural sub- formed by A.J.S. and L.K.G. D.W.G., S.M.S., R.C.L., and J.A.M. performed the data analysis and interpretation. D.W.G. drafted strates and biobehavioural correlates. Philosophical Transactions the manuscript, and all authors provided critical revisions and of the Royal Society of London Series B: Biological Sciences, approved the final version of the manuscript for submission. 359(1449), 1395–411. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 D. W. Grupe et al. | 319 Knight, M., Mather, M. (2009). Reconciling findings of Desikan, R.S., Se ´ gonne, F., Fischl, B., et al. (2006). An automated labeling system for subdividing the human cerebral cortex on emotion-induced memory enhancement and impairment of MRI scans into gyral based regions of interest. Neuroimage, preceding items. Emotion, 9(6), 763–81. 31(3), 968–80. Kovacs, M., Bylsma, L.M., Yaroslavsky, I., et al. (2016). Positive af- Dichter, G.S., Tomarken, A.J. (2008). The chronometry of affective fectivity is dampened in youths with histories of major depres- startle modulation in unipolar depression. Journal of Abnormal sion and their never-depressed adolescent siblings. Clinical Psychology, 117(1), 1–15. Psychological Science, 4(4), 661–74. Fjell, A.M., Walhovd, K.B. (2010). Structural brain changes in LaBar, K.S., Phelps, E.A. (1998). Arousal-mediated memory con- aging: courses, causes and cognitive consequences. Reviews in solidation: role of the medial temporal lobe in humans. the Neurosciences, 21, 187–221. Psychological Sciences, 9(6), 490–3. Fox, C., Moon, S., Iaria, G., Barton, J. (2009). The correlates of sub- Lang, P.J., Bradley, M.M., Cuthbert, B.N. (2008). International jective perception of identity and expression in the face net- Affective Picture System (IAPS): Affective Ratings of Pictures and work: an fMRI adaptation study. Neuroimage, 44(2), 569–80. Instruction Manual. Technical Report A–8. University of Florida, Fredrickson, B.L., Tugade, M.M., Waugh, C.E., Larkin, G.R. (2003). Gainesville, FL. What good are positive emotions in crisis? A prospective study Lapate, R.C., Rokers, B., Tromp, D.P.M., et al. (2016). Awareness of of resilience and emotions following the terrorist attacks on emotional stimuli determines the behavioral consequences of the United States on September 11th, 2001. Journal of Personality amygdala activation and amygdala-prefrontal connectivity. and Social Psychology, 84(2), 365–76. Science Report, 6(1), 1–16. Frijda, N.H., Mesquita, B., Sonnemans, J., van Goozen, S. (1991). Lapate, R.C., Samaha, J., Rokers, B., Hamzah, H., Postle, B.R., The duration of affective phenomena or emotions, sentiments Davidson, R.J. (2017). Inhibition of lateral prefrontal cortex pro- and passions. In: Strongman, K.T., editor. International Review of duces emotionally biased first impressions: a transcranial Studies on Emotion, Chichester, UK: Wiley, 187–225. magnetic stimulation and electroencephalography study. Gross, J.J., John, O.P. (2003). Individual differences in two emotion Psychological Sciences, 28, 942–53. regulation processes: implications for affect, relationships, Lapate, R.C., van Reekum, C.M., Schaefer, S.M., et al. (2014). and well-being. Journal of Personality and Social Psychology, 85(2), Prolonged marital stress is associated with short-lived re- 348–62. sponses to positive stimuli. Psychophysiology, 51(6), 499–509. Gunthert, K.C., Cohen, L.H., Butler, A.C., Beck, J.S. (2007). Larson, C.L., Schaefer, H.S., Siegle, G.J., Jackson, C.A.B., Anderle, Depression and next-day spillover of negative mood and de- M.J., Davidson, R.J. (2006). Fear is fast in phobic individuals: pressive cognitions following interpersonal stress. Cognitive amygdala activation in response to fear-relevant stimuli. Therapy and Research, 31(4), 521–32. Biological Psychiatry, 60(4), 410–7. Harry, B., Williams, M.A., Davis, C., Kim, J. (2013). Emotional ex- Lindquist, K.A., Satpute, A.B., Wager, T.D., Weber, J., Barrett, L.F. pressions evoke a differential response in the fusiform face (2015). The brain basis of positive and negative affect: evidence area. Frontiers in Human Neuroscience, 7, 1–6. from a meta-analysis of the human neuroimaging literature. Haxby, J.V., Gobbini, M.I. (2011). Distributed neural systems for Cerebral Cortex, 26, 1910–22. face perception. In: Calder, A.J., Rhodes, G., Johnson, M.H., MacPherson, S.E., Phillips, L.H., Della Sala, S. (2002). Age, executive Haxby, J. V., editors. Oxford Handbook of Face Perception. New function, and social decision making: a dorsolateral prefrontal York: Oxford University Press. theory of cognitive aging. Psychology and Aging, 17(4), 598–609. Haxby, J.V., Hoffman, E.A., Gobbini, M.I. (2000). The distributed Mather, M. (2016). The affective neuroscience of aging. Annual human neural system for face perception. Trends in Cognitive Review Of Psychology, 67, 213–38. Mather, M., Mitchell, K.J., Raye, C.L., Novak, D.L., Greene, E.J., Science, 4(6), 223–33. Hayes, J.P., LaBar, K.S., McCarthy, G., et al. (2011). Reduced hippo- Johnson, M.K. (2006). Emotional arousal can impair feature campal and amygdala activity predicts memory distortions for binding in working memory. Journal of Cognitive Neuroscience, trauma reminders in combat-related PTSD. Journal of 18(4), 614–25. Psychiatric Research, 45(5), 660–9. Messer, K., Matas, J., Kittler, J., Luettin, J., Maitre, G. (1999). Heller, A.S., Fox, A.S., Wing, E.K., Mcquisition, K.M., Vack, N.J., XM2VTSDB: The extended M2VTS database, in: 2nd Davidson, R.J. (2015). The neurodynamics of affect in the la- International Conference on Audio Visual Based Person boratory predicts persistence of real-world emotional re- Authentication. sponses. Journal of Neuroscience, 35(29), 10503–9. Mugler, J.P., Brookeman, J.R. (1990). Three-dimensional Heller, A.S., Johnstone, T., Shackman, A.J., et al. (2009). Reduced magnetization-prepared rapid gradient-echo imaging (3D MP capacity to sustain positive emotion in major depression re- RAGE). Magnetic Resonance in Medicine, 15(1), 152–7. flects diminished maintenance of fronto-striatal brain activa- Mumford, J.A., Poline, J.B., Poldrack, R.A. (2015). Orthogonalization tion. Proceedings of the National Academy of Sciences of the United of regressors in fMRI models. PLoS ONE, 10(4), e0126255–11. Murphy, S.T., Zajonc, R.B. (1993). Affect, cognition, and aware- States of America, 106(52), 22445–50. Heller, A.S., van Reekum, C.M., Schaefer, S.M., et al. (2013). ness: affective priming with optimal and suboptimal stimulus Sustained striatal activity predicts eudaimonic well-being and exposures. Journal of Personality and Social Psychology, 64(5), cortisol output. Psychological Science, 24(11), 2191–200. 723–39. Hurlemann, R., Hawellek, B., Matusch, A., et al. (2005). Reed, A.E., Chan, L., Mikels, J.A. (2014). Meta-analysis of the Noradrenergic modulation of emotion-induced forgetting. age-related positivity effect: age differences in preferences for Journal of Neuroscience, 25, 6343–9. positive over negative information. Psychology and Aging, 29(1), Kensinger, E.A., Corkin, S. (2004). Two routes to emotional mem- 1–15. ory: distinct neural processes for valence and arousal. Ryff, C.D., Keyes, C.L. (1995). The structure of psychological Proceedings of the National Academy of Sciences of the United States well-being revisited. Journal of Personality and Social Psychology, of America, 101(9), 3310–5. 69(4), 719–27. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018 320 | Social Cognitive and Affective Neuroscience, 2018, Vol. 13, No. 3 Troy, A.S., Wilhelm, F.H., Shallcross, A.J., Mauss, I.B. (2010). Schachter, S., Singer, J.E. (1962). Cognitive, social, and physio- logical determinants of emotional state. Psychological Reviews, Seeing the silver lining: cognitive reappraisal ability moder- 69(5), 379–99. ates the relationship between stress and depressive symp- Schaefer, S.M., Boylan, J.M., Van Reekum, C.M., et al. (2013). toms. Emotion, 10(6), 783–95. Purpose in life predicts better emotional recovery from nega- van Reekum, C.M., Schaefer, S.M., Lapate, R.C., Norris, C.J., tive stimuli. PLoS ONE, 8(11), e80329. Greischar, L.L., Davidson, R.J. (2011). Aging is associated with Schuyler, B.S., Kral, T.R., Jacquart, J., et al. (2014). Temporal dy- positive responding to neutral information but reduced recov- namics of emotional responding: amygdala recovery predicts ery from negative information. Social Cognitive and Affective emotional traits. Social Cognitive and Affective Neuroscience, 9(2), Neuroscience, 6(2), 177–85. 176–81. Verduyn, P., Delaveau, P., Rotge ´ , J.-Y., Fossati, P., Van Mechelen, Shackman, A.J., Tromp, D.P.M., Stockbridge, M.D., Kaplan, C.M., I. (2015). Determinants of emotion duration and underlying Tillman, R.M., Fox, A.S. (2016). Dispositional negativity: an in- psychological and neural mechanisms. Emotion Reviews, 7(4), tegrative psychological and neurobiological perspective. 330–5. Psychological Bulletin, 142, 1275–314. Vuilleumier, P., Pourtois, G. (2007). Distributed and interactive brain mechanisms during emotion face perception: evi- Siegel, J.S., Power, J.D., Dubis, J.W., et al. (2014). Statistical im- provements in functional magnetic resonance imaging ana- dence from functional neuroimaging. Neuropsychologia, 45(1), lyses produced by censoring high-motion data points. Human 174–94. Brain Mapping, 35(5), 1981–96. Vuilleumier, P., Richardson, M.P., Armony, J.L., Driver, J., Dolan, Siegle, G.J., Steinhauer, S.R., Thase, M.E., Stenger, V.A., Carter, R.J. (2004). Distant influences of amygdala lesion on visual cor- C.S. (2002). Can’t shake that feeling: event-related fMRI assess- tical activation during emotional face processing. Nature ment of sustained amygdala activity in response to emotional Neuroscience, 7(11), 1271–8. information in depressed individuals. Biological Psychiatry, Watson, D., Clark, L.A., Tellegen, A. (1988). Development and val- 51(9), 693–707. idation of brief measures of positive and negative affect: the Tambini, A., Rimmele, U., Phelps, E.A., Davachi, L. (2017). PANAS scales. Journal of Personality and Social Psychology, 54(6), Emotional brain states carry over and enhance future memory 1063–70. formation. Nature Neuroscience, 20(2), 271. Zajonc, R.B. (1968). Attitudinal effects of mere exposure. Journal Tottenham, N., Tanaka, J.W., Leon, A.C., et al. (2009). The of Personality and Social Psychology, 9(2, Pt.2), 1–27. NimStim set of facial expressions: judgments from untrained Zou, G.Y. (2007). Toward using confidence intervals to compare research participants. Psychiatry Research, 168(3), 242–9. correlations. Psychological Methods, 12(4), 399–413. Downloaded from https://academic.oup.com/scan/article-abstract/13/3/310/4855279 by Ed 'DeepDyve' Gillespie user on 16 March 2018

Journal

Social Cognitive and Affective Neuroscience – Oxford University Press

Published: Mar 1, 2018

There are no references for this article.