Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Apr 24.
Published in final edited form as: Mol Psychiatry. 2016 Sep 6;22(5):774–783. doi: 10.1038/mp.2016.143

Genome-wide association study of positive emotion identifies a genetic variant and a role for microRNAs

Aliza P Wingo 1,2,#, Lynn M Almli 2,#, Jennifer S Stevens 2,#, Tanja Jovanovic 2, Thomas S Wingo 1,3,4, Gregory Tharp 5, Yujing Li 4, Adriana Lori 2, Maria Briscione 1,2, Peng Jin 4, Elisabeth B Binder 6, Bekh Bradley 1,2, Greg Gibson 7, Kerry J Ressler 2,8,*
PMCID: PMC5339071  NIHMSID: NIHMS800485  PMID: 27595594

Abstract

Positive affect denotes a state of pleasurable engagement with the environment eliciting positive emotion such as contentment, enthusiasm, or happiness. Positive affect is associated with favorable psychological, physical, and economic outcomes in many longitudinal studies. With a heritability of ≤64%, positive affect is substantially influenced by genetic factors; however, our understanding of genetic pathways underlying individual differences in positive affect is still limited. Here, through a genome-wide association study (GWAS) of positive affect in African American participants, we identify a single nucleotide polymorphism (SNP), rs322931, significantly associated with positive affect at p<5×10−8, and replicate this association in another cohort. Furthermore, we show that the minor allele of rs322931 predicts expression of microRNAs miR-181a and miR-181b in human brain and blood, greater nucleus accumbens reactivity to positive emotional stimuli, and enhanced fear inhibition. Prior studies have suggested that miR-181a is part of the reward neurocircuitry. Taken together, we identify a novel genetic variant for further elucidation of genetic underpinning of positive affect that mediates positive emotionality potentially via the nucleus accumbens and miR-181.

Positive affect and negative affect constitute two broad general factors consistently identified as the dominant dimensions of emotional experience in inter-individual and intra-individual analyses across diverse descriptor sets, time frames, languages, and cultures.1–4 Positive affect, i.e. a state of pleasurable engagement with the environment eliciting positive emotions5, 6, has been shown to be associated with favorable psychological, physical, and economic outcomes in many longitudinal studies5–9. For instance, positive affect was a source of resilience for women having chronic pain7, buffer against risk for depression8, and protect against psychiatric symptoms in children exposed to natural disaster9. Further, a systematic review of longitudinal studies in healthy populations and populations with medical illness found that positive affect was associated with reduced mortality in both populations, independently of negative affect5. Moreover, a large U.S. representative longitudinal study found that young adults who reported higher positive affect went on to earn significantly higher income a decade later, even after the adjustment of potential confounders6.

Despite the many important impacts of positive affect, we still have very limited knowledge of its underlying biological mechanisms. Positive affect has a heritability of ≤64%10, 11, suggesting that genetic factors moderately contribute to this construct; however, the specific molecular loci contributing to this heritability remain to be identified. Hence, to elucidate genetic variants contributing to this phenotype, we performed a genome-wide association study (GWAS) of positive affect and a quasi-replication of our findings in positive affect in the realm of spiritual wellbeing in an independent cohort. Subsequently, we investigated functional significance of the identified genetic variants using brain and blood gene expression data and behavioral and neuroimaging data.

METHODS

The discovery sample

Participants in the discovery sample were recruited by the Grady Trauma Project (GTP) using the inclusion criteria of age ≥18, understanding English, and being able to give informed consent. Exclusion criteria included being acutely suicidal, psychotic, or having acute medical problems. Participants gave informed consent and the Institutional Review Boards of Emory University and Grady Memorial Hospital approved the study.

We assessed the trait-level positive affect with the Positive And Negative Affect Schedule (PANAS)12. Positive affect was represented by total score of the 10-item scale for positive emotional state (i.e. feeling interested, excited, strong, enthusiastic, proud, inspired, determined, attentive, and active). Each item was rated by participants using a 5-point scale ranging from 1=very slightly to not at all to 4=quite a bit, and 5=extremely for the extent that they feel a particular positive emotion in general, that is, on average. The score ranges from 0–50, with higher score reflecting more positive affect. The PANAS has been validated in college students, other adults, and clinical populations and showed high internal consistency (Cronbach’s α of 0.88), good stability over a two-month period (test-retest correlation of 0.6812), and substantial temporal stability over seven years (test-retest correlation of 0.42)13.

Childhood maltreatment was assessed with the well-validated Childhood Trauma Questionnaire14.

Genotypes and principal components in the GTP discovery sample

DNA was extracted from saliva or blood and genotyped on Illumina’s Omni1-Quad BeadChip. We used Illumina’s GenomeStudio to call genotypes and PLINK15 to perform quality-control (QC) analyses (see Supplementary Figure 1). We removed SNPs with call rates ≤98% or minor allele frequency (MAF) <0.01, and individuals with >2% missing data or having heterozygosity ≥2.5 standard deviations (SDs) away from the mean. We used PLINK to estimate the proportion of identity by descent (IBD) for each pair of individuals. Among pairs of individuals with an IBD >0.12 (indicating cousins or a closer relation), we removed the individual in each pair with the higher rate of missing genotype data. Using autosomal data pruned in PLINK, we performed principal-component analysis (PCA) to infer axes of ancestry and remove outlier subjects. Based on PCA, we retained those African-American individuals who fell within three SDs of the medians of the first and second principal components (PCs) in our sample. After completion of QC and PCA, our sample consisted of 3728 African-American individuals genotyped for 883,511 SNPs. Of the individuals included in the genetic sample, 68% had positive affect data and were included in the GWAS (N=2522). We compared the N=2522 group with the N=3728 group and found that they had very similar distributions of the MAFs, demographic, and psychological characteristics (Supplementary Table 1) suggesting that the discovery sample was representative of the overall N=3728 sample.

GWAS of positive affect in the discovery sample

The positive affect score was transformed using the one-parameter Box-Cox transformation with a lambda value of 2 to increase normality16, 17. Using PLINK, we regressed the positive affect score on allele count assuming an additive model and included gender, childhood maltreatment total score, and the top 10 PCs as covariates. We use the commonly accepted genome-wide significance level of p<5×10e-818, 19.

cis-eQTL analysis for rs322931 using brain dataset

The publicly accessible dataset from the UK Human Brain Expression Consortium (BRAINEAC; braineac.org), consists of post-mortem brain samples from 134 neuropathologically normal individuals of European descent20, 21. We examined the average gene expression across the ten available regions in this dataset: cerebellar cortex, frontal cortex, hippocampus, medulla, occipital cortex, putamen, substantia nigra, temporal cortex, thalamus, and intralobular white matter. Affymetrix GeneChip Human Exon 1.0 ST arrays were used to obtain whole transcriptome profiling20. The QC process was previously described in detail20, 21. Expression data were corrected for gender and batch effects of hybridization date and brain bank prior to eQTL analyses.

Genomic DNA was extracted from brain tissues and genotyped on Illumina’s Infinium Omni1-Quad BeadChip and ImmunoChip. After standard QC, both datasets were combined and only SNPs with good post-imputation quality (r2>0.50) and MAF≥0.05 were included for analyses20, 21. For rs322931, only one subject had imputed genotype in this dataset.

We obtained exon-level and transcript-level probes for the genes located within 1 MB of rs322931 for cis-eQTL analysis21, in which the expression level was the outcome and rs322931 genotype the independent variable (using additive model). Bonferroni adjustment was used for multiple testing correction.

rs322931 vs. miR-181 in blood

RNA was extracted from whole blood of 68 GTP participants. Small RNA was size-fractionated from total RNA and then used as the starting material to generate small RNA libraries following the Illumina TruSeq® Small RNA protocol. Quality of the sequencing data was inspected with FastQC. Adapters were trimmed with Trimmomatic22. After trimming, sequences <14 nucleotides were excluded. Trimmed reads were aligned to the miRBase v.2123 with SHRiMP aligner24, and only reads with a PHRED score ≥10 were considered for alignment. Percent reads mapped to mature miRNAs ranged from 81% to 98%, indicating high quality libraries and sequencing data. Among the 68 samples, one was an outlier with raw read counts 2.4 SDs below the mean and was excluded from the analysis. Counts of aligned miRNAs were normalized with regularized log-transformation using DESeq225. The following miR-181a/b was extracted from this dataset: miR-181a-2-3p, miR-181a-3p, miR-181a-5p, miR-181b-2-3p, miR-181b-3p, and miR-181b-5p. Above 90% of the 67 samples had 0 abundance count for miR-181b-2-3p and miR-181b-3p, and 90% of the samples had counts ≤ 7 for miR-181a-3p; thus these three miRNAs were excluded from the eQTL analysis based on the threshold criteria for calling expressed miRNAs with high confidence23. eQTL analysis for the remaining three miR-181a and miR-181b was performed using linear regression, where the normalized expression of the miRNA was the outcome, genotype was the independent variable (additive model), and gender, the first genotypic PC, and batch were the covariates. Multiple testing was addressed with Bonferroni correction.

Functional MRI study

Procedure

55 women from the GTP cohort (2 TT, 16 TC, and 37 CC for rs322931) completed an MRI scan, as illustrated in Figure 3. Two participants were excluded for falx calcification and eight due to excessive head motion (see below). The final sample for analysis included N=45: 2 TT, 12 TC, and 31 CC (demographics in Supplementary Table 4). Participants viewed scene stimuli from the International Affective Picture Series26. Trials began with a white fixation cross, centered on a black background for a jittered inter-trial interval (ITI) of 1.5 – 2.5 seconds. The scene was then presented for 1.5 seconds, followed by a rating period of 1.5 seconds. Scene stimuli were displayed in color full-screen at a resolution of 1024 × 768. The rating screen included a black background with “like / neutral / dislike” centered in white, 48pt Helvetica font. Thirty-six positive, 36 negative, and 36 neutral scenes were presented in a semirandom order such that no more than two pictures of the same valence preceded one another. Following scanning, participants reported subjective emotional arousal responses to each scene on a 1–5 scale (1-very little or no arousal, 5 - high arousal). Item order in the viewing and rating tasks was counterbalanced across participants.

Figure 3. FMRI task measuring response to emotional scene stimuli.

Figure 3

Open in a new tab

a) Example trial structure during the scene-viewing task. b) Subjective arousal ratings, shown as the mean across participants for each emotion condition, by rs322931 genotype group. Emotion condition significantly influenced arousal ratings, F[2,60]=40.9, p<0.001), and there was no effect of genotype or genotype*emotion interaction (p-values >0.10). Error bars represent 1 SE. c) Significant clusters of fMRI activation for positive relative to neutral scenes across the whole sample, p<.05, corrected at the whole-brain level. d) Significant clusters of fMRI activation for negative relative to neutral scenes across the whole sample, p<.05, corrected at the whole-brain level. FMRI activation clusters are overlaid on a representative single subject in MNI space, and axial slices are shown in neurological orientation. Positive and negative stimuli elicited greater fMRI activation than neutral stimuli in brain regions consistent with emotional arousal and regulation, including the bilateral amygdala, insula, medial prefrontal cortex (mPFC), and occipital cortex. Abbreviations: AMY=amygdala; OCC=occipital cortex; INS=insula, mPFC=medial prefrontal cortex.

MRI acquisition

Scanning took place on a 3.0T Siemens Trio with echo-planar imaging (EPI) (Siemens, Malvern, PA). T1-weighted scans were collected using 3D MP-RAGE with 176 contiguous 1 mm axial slices (TR/TE/TI = 2600/3.02/900ms, 1mm3 voxel size). Functional images were gathered using 37 3mm axial slices in a descending interleaved sequence (TR/TE =2000/3.00ms, 3mm3 voxel size).

MRI preprocessing and analysis

Initial data quality checks were performed using ArtRepair27. Slices containing spike artifacts were identified and replaced using linear interpolation, with no more than 4% of slices repaired per participant. Volumes affected by motion artifact were repaired using linear interpolation, with no more than 5% of volumes repaired per participant. Participants with head motion > 1.5mm/TR in more than 5% of volumes were excluded from further analyses. Additional image preprocessing steps and statistical analyses were implemented in SPM8 (Wellcome Trust Centre for Neuroimaging). Volumes were slice-timing corrected to the middle slice in time, and spatially realigned to the first image of the run. A 128Hz high-pass filter removed low-frequency noise28. Segmented T1s and co-registered functional images were normalized to the Montreal Neurological Institute (MNI) template. Functional images were visually examined to verify that no participant had dropout in any substantial portion of the nucleus accumbens (NAc) or amygdala. Images were smoothed with a 6mm Gaussian kernel.

Task-related activity was modeled by convolving the onset times of scene stimuli in the positive, negative, and neutral emotion conditions with a canonical hemodynamic response function. First-level models included the 3 emotion conditions, and covariates for 6 rigid-body motion parameters. Contrasts of the Negative > Neutral conditions, and Positive > Neutral conditions from each subject were used in group-level random effects analyses. Genotype effects on emotion-related activation were investigated for rs322931. A dominance model was used to compare emotion-related activation in the CC group (N=31) to activation in T-carriers (TC, TT; N=14). Regions of interest (ROIs) were defined anatomically, using the masks from the IBASPM MaxPro MNI atlas29 (bilateral NAc), and probabilistic maps following Amunts et al30 (bilateral amygdala). Statistical significance levels for ROIs and exploratory whole-brain analyses were height-extent-corrected to a threshold of p<0.05, using Alphasim Monte Carlo simulation with 1000 iterations for voxels within the NAc and amygdala ROIs, and a gray-matter mask created from the segmented MNI 152-subject average T1 for whole-brain analysis. With a cluster-forming threshold of p<0.05, extents of k=5 for the NAc (voxel-wise p<0.008) and k=18 for the amygdala (voxel-wise p<0.003) were required for a corrected threshold of p<0.05. Exploratory whole-brain analyses with a cluster-forming threshold of p<0.01 required an extent threshold of k=25 to reach a corrected threshold of p<0.05 (voxel-wise p<0.0004). Subjective emotional arousal ratings were tested using a mixed-effects 3×2 ANOVA with emotion as a within-subjects predictor (positive, negative, neutral), and rs322931 group as between-subjects predictor (TT/TC vs. CC).

Fear-potentiated startle

Demographic data on the 248 GTP participants (14 TT, 87 TC, 147 CC for rs322931) who completed the fear-potentiated startle (FPS) task are shown in Supplementary Table 7. The FPS protocol was based off previous work31 shown to produce robust conditioning and consisted of a habitual phase, in which the conditioned stimuli (CSs) were presented without any reinforcement and a conditioning phase, which consisted of three blocks, each containing four presentations of each trial type, a reinforced CS (CS+), nonreinforced CS (CS−), and noise probe alone (NA). A response keypad was used during fear conditioning to record expectancies of the US on each CS presentation.

As previously described32, we measured the eyeblink component of the acoustic startle response by electromyography recordings of the right orbicularis oculi muscle. Data were acquired using Biopac MP150 for Windows and were filtered, rectified, and smoothed using MindWare software (MindWare Technologies, Ltd., Gahanna, OH). Startle magnitude was measured as the maximum amplitude of the eyeblink muscle contraction 20–200 ms after presentation of the startle probe. FPS was calculated by subtracting the mean startle magnitude for NA trials from the CS+ or CS− trials. A dominance model was used to compare the FPS to the CS+ (threat condition) and FPS to CS− (safety condition) in the CC group (N=147) to T-carriers (TC, TT; N=101).

Center for Health Discovery and Well Being (CHDWB) Replication Sample

The CHDWB was established to evaluate effectiveness and utility of a health and prevention-focused rather than disease-focused care setting. Inclusion criteria and exclusion criteria have been detailed previously33. Spiritual wellbeing was assessed with the FACIT Spiritual Well-Being scale, a well validated, 12-item scale measuring sense of inner peace, strength from spiritual beliefs, and sense of purpose in life in the past seven days34. Its score ranges from 0–48, with the higher score reflecting more spiritual wellbeing. The scale has been translated and validated in 15 different languages and used in many studies34. We included the two main ethnic groups, African American (N=103) and Caucasian (N=323) in the analysis, and excluded one American-Indian and 24 Asian-American individuals. Per dbSNP, rs322931 has similar MAF in African Americans (MAF=0.184) and Caucasians (MAF=0.183).

Genotypes and PCs in the CHDWB: DNA was extracted from blood and typed either on Illumina’s OmniQuad or Core+Exome arrays to obtain genome-wide genotypes. SNPs with MAF <0.01, with missing data in >5% of the samples, and with Hardy-Weinberg Equilibrium (HWE) p-value <10−4 were removed. Using PLINK15, we pruned the genome-wide autosome SNPs in 50-SNP-windows to remove all SNPs with linkage disequilibrium (LD) >0.2, then we computed pairwise identity-by-state metrics, from which the multiple dimensional scaling was computed.

Association between rs322931 and spiritual wellbeing was examined with linear regression, in which genotype for rs322931 was the independent variable (using additive model), total score of spiritual wellbeing was the outcome, gender, and first four PCs as covariates.

RESULTS

2522 African-American participants, whose socio-demographic characteristics are presented in Supplementary Table 2, were included in the GWAS. We found no evidence to suggest inflation of the association test statistics on the quantile-quantile plot after co-varying for 10 PCs, gender, and childhood maltreatment (genomic inflation factor λ=1.00; Supplementary Figure 2). We found two SNPs significantly associated with positive affect after gender, childhood maltreatment, and 10 PCs were adjusted for: rs322931 (β (SE)=60.3 (10.8), p=2.59×10−8) and rs7550394 (β (SE)=62.7 (11.4), p=3.84×10−8; Figure 1A). The SNPs reside on chromosome 1 in a locus for LINC01221 and are in high LD with each other (D’=0.988 and r2 = 0.844). rs322931 and rs7550394 had a MAF of 0.19 and 0.17, and HWE p-value of 0.8 and 1, respectively. The SNPs are closest to miR-181a and miR-181b (Figure 1B). The minor alleles of both SNPs were associated with having more positive affect. These two SNPs also met genome-wide significance level when we only adjusted for gender and 10 PCs (rs322931: β=61.6; p=3.86 × 10−8; rs7550394: β=64.4; p=4.73 × 10−8). Given their high degree of LD with each other, in follow-up studies, we focused on rs322931 for further analyses.

Figure 1. Genome-wide Association Study of Positive Affect Identifies SNP rs322931.

Figure 1

Open in a new tab

a) Manhattan Plot of the GWAS for positive affect (rs322931, p=2.59×10−8); the blue line denotes standard genome-wide corrected significance18 at the negative log10(5×10−8) = 7.3. b) Regional plot of SNP associations within 1 MB of rs322931 (purple triangle) on chromosome 1; plot generated with LocusZoom.

rs322931 is a brain cis-eQTL for miR-181a and miR-181b

We next investigated if rs322931 is a cis-eQTL in the brain. There are six genes within 1 MB of rs322931: MIR181A, MIR181B, PTPRC, ATP6V1G3, NEK7, and NR5A2 (Figure 1B). The six genes were represented by eight transcripts in the BRAINEAC dataset (Table 1). Among these transcripts, rs322931 was significantly associated with all four of the transcripts for miR-181a1 and miR-181b1, even after multiple testing correction (uncorrected p-values = [1.30×10−5 to 6.60×10−4]; Bonferroni adjusted p-values = [1.04×−4 to 5.28×10−3], (Table 1)).

Table 1.

Association between rs322931 and average gene expression across the ten brain regions for the six genes located within 1 MB of rs322931 (cis-eQTL) at the transcript-level ID.

geneSymbol marker rsid exprID chr start stop p-value
MIR-181A1 chr1:199019855 rs322931 t2450056 chr1 198868041 198868274 1.30E-05
MIR-181A1 chr1:199019855 rs322931 t2450058 chr1 198868449 198868844 1.40E-04
MIR-181A1 chr1:199019855 rs322931 t2450054 chr1 198867434 198868019 2.30E-04
MIR181A1, MIR181B1 chr1:199019855 rs322931 t2450000 chr1 198776723 198906548 6.60E-04
NR5A2 chr1:199019855 rs322931 t2374126 chr1 199845845 200146534 6.30E-03
NEK7 chr1:199019855 rs322931 t2373736 chr1 197958385 198292204 1.20E-01
ATP6V1G3 chr1:199019855 rs322931 t2449922 chr1 198492257 198510075 2.00E-01
PTPRC chr1:199019855 rs322931 t2373842 chr1 198519065 198726605 2.10E-01
Open in a new tab

LOC100131234 is identified as miR-181A in pubmed.org

At the exon-level, there were eight probes for miR-181a1/b1 (Supplementary Table 3), of which six were significantly associated with rs322931 at uncorrected p-values ranging from 1.30×10−5 to 1.90×10−3. The minor allele of rs322931 was significantly associated with lower expression of miR-181a1 and miR-181b1 for all of these six probes. The top three of these six probes were exprID 2450057 (β=0.29±0.06; p=1.30×10−5; Figure 2A), exprID 2450024 (β=0.22±0.06, p=1.40×10−4; Figure 2B), and exprID 2450059 (β=0.20±0.05, p=1.40×10−4; Figure 2C).

Figure 2. SNP rs322931 associates with levels of miR-181 in brain and blood.

Figure 2

Open in a new tab

The minor allele of rs322931 is significantly associated with lower expression of miR-181a1 and miR-181b1 in the brain for all six of their exon-level probes. Shown on the x-axis are the number of minor allele of rs322931 and on the y-axis log2 of level of expression for each of the top three exon-level probes of miR-181a1 and miR-181b1: a) exprID 2450057 (p=1.30×10−5); b) exprID 2450024 (p=1.40×10−4); and c) exprID 2450059 (p=1.40×10−4); d) The minor allele of rs322931 is significantly associated with higher expression of miR-181-5p in blood. Shown on the x-axis is the number of minor alleles of rs322931 and on the y-axis level of RNA expression of miR-181b-5p. Outliers are depicted as black dots and the median is the middle line in the box.

In our secondary analysis we found that rs7550394 was also significantly associated with all four of the transcripts for miR-181a1/b1, with p-values ranging from 9.10×10−5 to 3.30×10−3. Likewise, its minor allele was associated with decreased expression of miR-181a1 and miR-181b1.

rs322931 is a blood eQTL for miR-181b

Since we found that rs322931 is a brain cis-eQTL for miR-181a/b, we examined the relationship between this SNP and miR-181a/b in peripheral whole blood. We found that rs322931 was significantly associated with miR-181b-5p after gender, age, population substructure, and batch effects were simultaneously adjusted for (β=0.19±0.07, p=0.0121, Bonferroni adjusted p=0.0363; Figure 2D). The minor allele of rs322931 was associated with higher expression of miR-181b-5p (Figure 2D).

Effects of rs322931 on neuroimaging and psychophysiological measures of emotion

To probe potential effects on systems-level neurobiology, we examined the rs322931 polymorphism in relation to brain and psychophysiological markers of emotional reactivity. In the fMRI task comparing positive versus neutral stimuli, T-allele carriers showed greater responses than those with CC genotype in the NAc (Figure 4A) and amygdala (Figure 4B) ROIs. No group differences were observed in the NAc and amygdala for negative versus neutral stimuli. Whole-brain analyses of T-carriers relative to the CC showed greater activation in the left putamen and caudate, left superior and middle frontal gyri, and left occipital cortex in response to both positive and negative stimuli, relative to neutral (Figure 4C–D, Supplementary Tables 5–6). T-carriers showed less activation than CC for negative versus neutral stimuli in the left thalamus and in midbrain areas including the periaqueductal gray and substantia nigra. In sum, the minor allele of rs322931 was associated with greater fMRI activation of brain regions involved in emotion and reward.

Figure 4. Effect of SNP rs322931 genotype on fMRI responses to emotional scene stimuli.

Figure 4

Open in a new tab

In both the NAc (a) and amygdala (b) ROIs, rs322931 T-carriers showed a greater response than the CC group to positive stimuli (positive>neutral; NAc: xyz=15,5,−14, Z=2.6, k=12; amygdala: xyz=36,−1,−32, Z=3.3, k=21). Bar graphs show mean contrast estimates across voxels within the bilateral anatomical ROI masks; error bars show one SE. Genotype did not influence the response to negative stimuli (negative>neutral) in either ROI. Additional regions that showed group differences in the whole-brain analysis are shown in (c) for the positive>neutral contrast, and (d) for the negative>neutral contrast. For all fMRI results, significant clusters are overlaid on slices from a representative template brain in MNI space, in neurological orientation. e) Effect of rs322931 genotype (TT/TC vs. CC) on fear-potentiated startle (p<0.05 marked with *).

In fear conditioning paradigm we found that T-carriers had less fear to the safety cue than CC group (p=0.025; Figure 4E). The groups did not differ in their responses to the fear cue. This finding suggests that the minor allele of rs322931 is associated with better fear inhibition in response to safety cue.

Quasi-replication of rs322931 in positive affect in an independent cohort

As spiritual wellbeing is a dimension of positive affect, we performed a quasi-replication of our GWAS finding in spiritual wellbeing in the CHDWB cohort. Among the 426 participants (295 women and 131 men), we found a significant association between rs322931 and spiritual wellbeing after gender and PCs were adjusted for (s (SE) = 1.28 (0.63); p=0.044). The minor allele of rs322931 was associated with having more spiritual wellbeing, consistent with our finding in the discovery sample.

Percent variance of positive affect explained by genome-wide common SNPs

Lastly, using GCTA software35 we estimated the proportion of variance of positive affect explained by genome-wide common SNPs (MAF ≥0.01) in the GTP discovery sample. We found that the common autosomal SNPs together explained 21.4% ± 14.6% of the variance of positive affect when 10 PCs were included as covariates at p-value of 0.07, and 21.5% ± 14.6% of the variance of positive affect when 10 PCs and gender were included as covariates at p-value of 0.07. Our large standard deviations and nearly significant p-values for the estimates are likely due to our relatively small sample size.

DISCUSSION

In a GWAS of positive affect we identified a novel SNP on chromosome 1, rs322931, whose minor allele was significantly associated with having more positive affect in the discovery sample and more spiritual wellbeing in an independent replication sample. Consistently, we found the minor allele of rs322931 significantly associated with increased fMRI activation of the NAc, a key brain region for pleasure, reward, and motivation, to positive emotional stimuli, and with better inhibition of fear during presentation of a safety cue in a human fear conditioning paradigm. Furthermore, rs322931 influences expression of miR-181a and miR-181b in the brain and miR-181b in blood, which is notable for several reasons. First, prior studies suggest that miR-181a is part of the reward neurocircuitry since its expression in hippocampal neurons is induced by cocaine and amphetamine and dopamine signaling36, and its expression in the NAc influences cocaine-induced additive behavior37. We postulate that rs322931 mediates positive affect via the reward neurocircuitry by influencing miR-181a expression. Second, miR-181a is strongly enriched in the synaptodendritic compartment of the NAc and influences synaptic plasticity through regulating glutamate receptor 2 subunit of AMPA-type glutamate receptors36. Hence, rs322931 can be inferred to mediate synaptic plasticity, a dynamic process that is important for learning and memory38, and its decline is associated with cognitive aging and age-related neurological disorders39. Such effects on plasticity may explain the finding that the minor allele of rs322931 was associated with better learning of the safety signal in the fear conditioning experiment. Third, miR-181a/b are intrinsic regulators of T-cell receptor sensitivity and signaling strength40, 41; thus, by influencing expression of miR-181a/b, the SNP rs322931 may have some influence on the immune system. Taken together, we postulate that rs322931 influences positive emotion via miR-181 and the nucleus accumbens and may also influence synaptic plasticity and immune functioning, consistent with the psychosomatic interplay among affect, neuroplasticity, and physical health observed in prospective epidemiological studies.

The minor allele of rs322931 was associated with having more positive affect and decreased expression of miR-181a/b in the brain but increased expression of miR-181b in the blood, which is consistent with adaptive effects of down-regulation of miR-181a/b in the brain and upregulation of miR-181b in the blood. For instance, patients with schizophrenia had up-regulated expression of miR-181b in the temporal cortex42, 43. Similarly, depressed mice had upregulated expression of miR-181a in the prefrontal cortex versus controls44. Likewise, overexpression of miR-181 in rat hippocampal neurons inhibited dendritic growth45 and decreased the size and number of dendritic spines36. In peripheral blood, PTSD patients had reduced expression of miR-181a/b versus controls46.

It is notable that rs322931 genotype influenced NAc and amygdala responses to positive stimuli but had no effect on responses to negative stimuli. These findings suggest potential differences in the genetic pathways contributing to positive and negative emotion traits. For disease models, the findings highlight that positive emotional traits do not simply reflect decreased negative emotion but may confer resilience to co-occurring negative mood or stress. However, given the observation that rs322931 minor allele predicted greater reactivity to both positive and negative stimuli in striatal clusters outside the NAc, additional research is needed to specifically probe the effects of positive versus negative stimuli on various aspects of reward-related circuitry. The minor allele of rs322931was also associated with lower activation to negative stimuli in the left thalamus, and in midbrain nuclei that play major roles in regulating arousal responses, whose lesion can block arousal-related potentiation of the acoustic startle responses47,48, 49. Down-regulation of midbrain activation in minor allele carriers may contribute to the decreased startle response to the CS− observed in the fear conditioning experiment. Moreover, conditioned safety signals such as the one used in this study have been shown to activate the NAc due to their having rewarding properties50.

We observed that the amygdala responded similarly to the NAc to positive stimuli but not to negative stimuli in the rs322931 T-carriers vs. CC group (Figure 4). Specifically, the amygdala showed increased activation to negative stimuli while the NAc did not, perhaps because a 1:1 association between these regions is not expected. Indeed, previous fMRI studies have often shown some disconnect in the responses of the amygdala versus NAc to positive and negative outcomes51, and a meta-analysis observes a consistent co-activation of the NAc and amygdala across studies, but much stronger associations between NAc and regions including anterior cingulate, medial frontal gyrus, parahippocampal gyrus, and other aspects of the striatum52. Both the amygdala and NAc are directly connected with many cortical and subcortical regions, and it is likely that they may receive different regulatory inputs from regions such as the prefrontal cortex and hippocampus depending upon the task or context.

For the GWAS of positive affect, we included childhood maltreatment as a covariate in the regression model for two reasons. First, childhood is a sensitive developmental window period when upbringing experiences can powerfully shape emotional development and psychological functioning in adulthood53, 54. In our dataset, there is a significant inverse correlation between childhood maltreatment total score and positive affect score (p<0.0001). Second, studies have shown that genetic variants can moderate one’s susceptibility to environmental insults55, 56.

We examined rs322931 in relation to depression, PTSD symptoms, alcohol use, drug use, negative affect, and resilience in the GTP sample, adjusting for gender and population substructure. Rs322931 was not significantly associated with depression, PTSD, alcohol use, or drug use. However, its minor allele was significantly associated with having less negative affect (p=0.002) and having more resilience (p=0.0018), consistent with our positive affect finding (Supplementary Table 8).

Okbay and colleagues recently published a very large GWAS of subjective well-being, operationally defined as having life satisfaction or positive affect, in which they identified three SNPs meeting genome-wide significance level for subjective well-being57. Additionally, the authors listed the top 44 SNPs for subjective wellbeing and the top 30 SNPs for positive affect in particular from their meta-analyses of original, unimputed datasets, as well as the top 76 SNPs for subjective wellbeing from a meta-analysis of cohorts with imputed genotypes. Our GTP dataset has original, unimputed genotypes. We examined the top SNPs from Okbay meta-analysis of GWAS of unimputed datasets in our GTP sample. Of their top 30 positive affect SNPs, 15 were found in the GTP sample. Of these 15 SNPs, one, rs6581971 (on chromosome 12, Okbay p= 4.84×10−6), was significantly associated with positive affect in our GTP sample at uncorrected p-value of 0.007. Of the Okbay’s top 44 SNPs for subjective wellbeing, nine SNPs were found in the GTP sample. Of these nine, one SNP, rs13102973 (on chromosome 4, Okbay p= 3.05×10−6) was significantly associated with positive affect at the uncorrected p of value of 0.0487. Next we examined the top 76 SNPs from Okbay’s meta-analysis of imputed datasets in our GTP sample. Of Okbay’s 76 top SNPs for subjective wellbeing, 5 SNPs were found in the GTP dataset but none was associated with positive affect at uncorrected p-value<0.05. These observations may be due to three factors. First, our GTP participants were all African Americans compared to the majority of Caucasian participants in Okbay’s GWAS. Second, the majority of the studies included in the Okbay’s meta-analysis used one item, usually extracted from various scales, to assess subjective wellbeing or positive affect while ours used a 10-item, psychometrically validated positive affect scale. While the Okbay’s GWAS has great power from their enormous sample size, our study may have more comprehensive and fine-grained assessment of positive affect. Third, our GTP discovery sample consisted of inner-city participants with high levels of stress and trauma exposure. Hence, positive affect in this context reflects positive emotions in the face of environmental adversity, which arguably allowed us to observe adaptive effects of polymorphisms that may not be observable in populations with lower trauma or stress exposure.

Our results should be interpreted in light of their limitations. First, the sample sizes for both the discovery and replication samples are relatively small. With relatively few studies on genetic underpinnings of positive emotions, we face the challenges of a nascent research area. Second, our replication is a quasi-replication since trait-level positive affect was assessed in the discovery sample and state-level spiritual wellbeing was used in the replication sample. Nevertheless, spiritual wellbeing represents positive affect specific to the dimensions of inner peace, sense of purpose in life, and strength from spiritual beliefs. Similar to positive affect, spiritual wellbeing has been shown to be associated with better quality of life (r=0.48), better life satisfaction (r = 0.318), and better physical wellbeing (r=0.47), and inversely associated with hopelessness (r=−0.55), anxious preoccupation (r=−0.49) and depression (r=−0.34)34. Notably, genetic factors influencing the expression of the different dimensions of psychological wellbeing have been shown to be largely shared, as reflected by the high estimates of genetic correlations among the different dimensions (ranging from 0.77–0.99)11.

Future studies are needed to replicate our findings and to deeply sequence the region around rs322931 to elucidate the causal genetic variants for positive emotions. Understanding the neurobiology of positive affect and emotional wellbeing may provide a complementary and powerful approach to the current predominant focus on the biology of psychopathology.

Supplementary Material

1

Acknowledgments

We appreciate the technical support of all of the staff and volunteers of the Grady Trauma Project, particularly Kimberly Kerley, Jordan Laird, Allen W. Graham, Angelo Brown, and Rebecca Roffman. We thank C.F. Gillespie, A Schwartz, and T. Weiss for medical support, A. Lott for maintaining the database, and Biao Zeng and Jing Zhao for genetic analytical support. We thank the participants of the Grady Trauma Project for their time and effort. We gratefully acknowledge the support of the Center for Health Discovery and Well Being and the Atlanta Clinical and Translational Science Institute for the clinical profiling of the CHDWB and the Georgia Tech Research Institute for start-up support to GG for genomic profiling. Funding: This study was supported in part by the Department of Veterans Affairs Career Development Award IK2CX000601 and the NARSAD Young Investigator Award (to APW). This work was primarily supported by the National Institutes of Mental Health (MH096764 and MH071537 to KJR; F32-MH101976 to JSS). Support was also received from Emory and Grady Memorial Hospital General Clinical Research Center, National Institutes of Health (NIH). TSW was supported by the Veterans Health Administration (BX001820) and NIH/NIA AG025688. The contents do not represent the views of the Department of Veterans Affairs or the United States Government.

Footnotes

Author contributions: All authors edited and commented on the manuscript. KJR obtained funding for the GWAS of positive affect, oversees the GTP study, and revised the manuscript. APW, LMA, and JSS wrote the first draft of the manuscript. APW and LMA performed the GWAS of positive affect. APW performed the eQTL analyses in brain and blood. JSS designed the fMRI study and analyzed the fMRI data. GG provided the genetic data for the CHDWB cohort. GG and APW performed the genetic replication study. TJ designed the study of fear conditioning and analyzed the neurophysiology data. APW and TSW performed the GCTA heritability analysis. GT contributed to the miRNA data analysis. YL and PJ contributed to the generation of miRNA data. EBB contributed to the generation of GWAS data. BB contributed to the generation of positive affect data.

Competing interests: The authors have no conflicts of interest to declare.

REFERENCES

  • 1.Watson D, Clark LA. Manual for the Positive and Negative Affect Schedule - Expanded Form. The University of Iowa; 1999. [Google Scholar]
  • 2.Mayer JD, Gaschke YN. The experience and meta-experience of mood. J Pers Soc Psychol. 1988;55(1):102–111. doi: 10.1037//0022-3514.55.1.102. [DOI] [PubMed] [Google Scholar]
  • 3.Watson D, Clark LA, Tellegen A. Cross-cultural convergence in the structure of mood: A Japanese replication and a comparison with U.S. findings. Journal of Personality and Social Psychology. 1984;47(1):127–144. [Google Scholar]
  • 4.Watson D, Tellegen A. Toward a consensual structure of mood. Psychol Bull. 1985;98(2):219–235. doi: 10.1037//0033-2909.98.2.219. [DOI] [PubMed] [Google Scholar]
  • 5.Chida Y, Steptoe A. Positive psychological well-being and mortality: a quantitative review of prospective observational studies. Psychosom Med. 2008;70(7):741–756. doi: 10.1097/PSY.0b013e31818105ba. [DOI] [PubMed] [Google Scholar]
  • 6.De Neve JE, Oswald AJ. Estimating the influence of life satisfaction and positive affect on later income using sibling fixed effects. Proc Natl Acad Sci U S A. 2012;109(49):19953–19958. doi: 10.1073/pnas.1211437109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zautra AJ, Johnson LM, Davis MC. Positive affect as a source of resilience for women in chronic pain. J Consult Clin Psychol. 2005;73(2):212–220. doi: 10.1037/0022-006X.73.2.212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Wichers MC, Myin-Germeys I, Jacobs N, Peeters F, Kenis G, Derom C, et al. Evidence that moment-to-moment variation in positive emotions buffer genetic risk for depression: a momentary assessment twin study. Acta Psychiatr Scand. 2007;115(6):451–457. doi: 10.1111/j.1600-0447.2006.00924.x. [DOI] [PubMed] [Google Scholar]
  • 9.Kujawa A, Hajcak G, Danzig AP, Black SR, Bromet EJ, Carlson GA, et al. Neural Reactivity to Emotional Stimuli Prospectively Predicts the Impact of a Natural Disaster on Psychiatric Symptoms in Children. Biol Psychiatry. 2015 doi: 10.1016/j.biopsych.2015.09.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Boardman J, Blalock C, Button T. Sex differences in the heritability of resilience. Twin Res Hum Genet. 2008;11:12–27. doi: 10.1375/twin.11.1.12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Gigantesco A, Stazi MA, Alessandri G, Medda E, Tarolla E, Fagnani C. Psychological well-being (PWB): a natural life outlook? An Italian twin study on heritability of PWB in young adults. Psychological Medicine. 2011;41(12):2637–2649. doi: 10.1017/S0033291711000663. [DOI] [PubMed] [Google Scholar]
  • 12.Watson D, Clark LA, Tellegen A. Development and validation of brief measures of positive and negative affect: the PANAS scales. J Pers Soc Psychol. 1988;54(6):1063–1070. doi: 10.1037//0022-3514.54.6.1063. [DOI] [PubMed] [Google Scholar]
  • 13.Watson D, Walker LM. The long-term stability and predictive validity of trait measures of affect. J Pers Soc Psychol. 1996;70(3):567–577. doi: 10.1037//0022-3514.70.3.567. [DOI] [PubMed] [Google Scholar]
  • 14.Bernstein DP, Stein JA, Newcomb MD, Walker E, Pogge D, Ahluvalia T, et al. Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse & Neglect. 2003;27(2):169–190. doi: 10.1016/s0145-2134(02)00541-0. [DOI] [PubMed] [Google Scholar]
  • 15.Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira M, Bender D, et al. PLINK: a toolset for whole-genome association and population-based linkage analysis. American Journal of Human Genetics. 2007;81:559–575. doi: 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.LaLonde SM. Transforming Variables for Normality and Linearity - When, How, Why and Why Not's. In: 2012 SGF, editor. Statistics and Data Analysis. SAS; 2012. pp. 430–436. [Google Scholar]
  • 17.Box GEP, Cox DR. An analysis of transformations. Journal of the Royal Statistical Society, Series B. 1964;26:211–252. [Google Scholar]
  • 18.Barsh GS, Copenhaver GP, Gibson G, Williams SM. Guidelines for genome-wide association studies. PLoS Genet. 2012;8(7):e1002812. doi: 10.1371/journal.pgen.1002812. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Identification of risk loci with shared effects on five major psychiatric disorders: a genome-wide analysis. The Lancet. 381(9875):1371–1379. doi: 10.1016/S0140-6736(12)62129-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Trabzuni D, Ryten M, Walker R, Smith C, Imran S, Ramasamy A, et al. Quality control parameters on a large dataset of regionally dissected human control brains for whole genome expression studies. Journal of neurochemistry. 2011;119(2):275–282. doi: 10.1111/j.1471-4159.2011.07432.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Ramasamy A, Trabzuni D, Guelfi S, Varghese V, Smith C, Walker R, et al. Genetic variability in the regulation of gene expression in ten regions of the human brain. Nat Neurosci. 2014;17(10):1418–1428. doi: 10.1038/nn.3801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30(15):2114–2120. doi: 10.1093/bioinformatics/btu170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kozomara A, Griffiths-Jones S. miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Res. 2014;42(Database issue):D68–D73. doi: 10.1093/nar/gkt1181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Rumble SM, Lacroute P, Dalca AV, Fiume M, Sidow A, Brudno M. SHRiMP: accurate mapping of short color-space reads. PLoS computational biology. 2009;5(5):e1000386. doi: 10.1371/journal.pcbi.1000386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Lang P, Bradley MM, Cuthbert BN. International affective picture system (IAPS): Affective ratings of pictures and instruction manual. Technical Report A-8. Gainesville, FL: University of Florida; 2008. [Google Scholar]
  • 27.Mazaika P, Whitfield-Gabrieli S, Reiss A. Artifact Repair for fMRI Data from High Motion Clinical Subjects. Human Brain Mapping. 2007 [Google Scholar]
  • 28.Holmes A, Josephs O, Buchel C, Friston K. Statistical modelling of low-frequency confounds in fMRI. NeuroImage. 1997;5:S480. [Google Scholar]
  • 29.Proceedings of the Nuclear Science Symposium Conference Record, 2008. NSS'08. IEEE. IEEE; 2008. Use of IBASPM atlas-based automatic segmentation toolbox in pathological brains: effect of template selection. [Google Scholar]
  • 30.Amunts K, Kedo O, Kindler M, Pieperhoff P, Mohlberg H, Shah NJ, et al. Cytoarchitectonic mapping of the human amygdala, hippocampal region and entorhinal cortex: intersubject variability and probability maps. Anatomy and Embryology. 2005;210(5):343–352. doi: 10.1007/s00429-005-0025-5. [DOI] [PubMed] [Google Scholar]
  • 31.Norrholm SD, Jovanovic T, Olin IW, Sands LA, Karapanou I, Bradley B, et al. Fear Extinction in Traumatized Civilians with Posttraumatic Stress Disorder: Relation to Symptom Severity. Biological Psychiatry. 2011;69(6):556–563. doi: 10.1016/j.biopsych.2010.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Jovanovic T, Norrholm SD, Blanding NQ, Davis M, Duncan E, Bradley B, et al. Impaired fear inhibition is a biomarker of PTSD but not depression. Depress Anxiety. 2010;27(3):244–251. doi: 10.1002/da.20663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Ferranti EP, Dunbar SB, Higgins M, Dai J, Ziegler TR, Frediani JK, et al. Psychosocial factors associated with diet quality in a working adult population. Research in nursing & health. 2013;36(3):242–256. doi: 10.1002/nur.21532. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Bredle J, Salsman J, Debb S, Arnold B, Cella D. Spiritual well-being as a component of health-related quality of life: The Functional Assessment of Chronic Illness Therapy-Spiritual Well-Being Scale (FACIT-Sp) Religions. 2011;2:77–94. [Google Scholar]
  • 35.Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: A Tool for Genome-wide Complex Trait Analysis. The American Journal of Human Genetics. 2011;88(1):76–82. doi: 10.1016/j.ajhg.2010.11.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Saba R, Storchel PH, Aksoy-Aksel A, Kepura F, Lippi G, Plant TD, et al. Dopamine-regulated microRNA MiR-181a controls GluA2 surface expression in hippocampal neurons. Molecular and cellular biology. 2012;32(3):619–632. doi: 10.1128/MCB.05896-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Chandrasekar V, Dreyer JL. Regulation of MiR-124, Let-7d, and MiR-181a in the accumbens affects the expression, extinction, and reinstatement of cocaine-induced conditioned place preference. Neuropsychopharmacology. 2011;36(6):1149–1164. doi: 10.1038/npp.2010.250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Martin SJ, Grimwood PD, Morris RG. Synaptic plasticity and memory: an evaluation of the hypothesis. Annual review of neuroscience. 2000;23:649–711. doi: 10.1146/annurev.neuro.23.1.649. [DOI] [PubMed] [Google Scholar]
  • 39.Henley JM, Wilkinson KA. AMPA receptor trafficking and the mechanisms underlying synaptic plasticity and cognitive aging. Dialogues in clinical neuroscience. 2013;15(1):11–27. doi: 10.31887/DCNS.2013.15.1/jhenley. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Li QJ, Chau J, Ebert PJ, Sylvester G, Min H, Liu G, et al. miR-181a is an intrinsic modulator of T cell sensitivity and selection. Cell. 2007;129(1):147–161. doi: 10.1016/j.cell.2007.03.008. [DOI] [PubMed] [Google Scholar]
  • 41.Zietara N, Lyszkiewicz M, Witzlau K, Naumann R, Hurwitz R, Langemeier J, et al. Critical role for miR-181a/b-1 in agonist selection of invariant natural killer T cells. Proc Natl Acad Sci U S A. 2013;110(18):7407–7412. doi: 10.1073/pnas.1221984110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Beveridge NJ, Tooney PA, Carroll AP, Gardiner E, Bowden N, Scott RJ, et al. Dysregulation of miRNA 181b in the temporal cortex in schizophrenia. Hum Mol Genet. 2008;17(8):1156–1168. doi: 10.1093/hmg/ddn005. [DOI] [PubMed] [Google Scholar]
  • 43.Beveridge NJ, Gardiner E, Carroll AP, Tooney PA, Cairns MJ. Schizophrenia is associated with an increase in cortical microRNA biogenesis. Mol Psychiatry. 2010;15(12):1176–1189. doi: 10.1038/mp.2009.84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Dwivedi Y, Roy B, Lugli G, Rizavi H, Zhang H, Smalheiser NR. Chronic corticosterone-mediated dysregulation of microRNA network in prefrontal cortex of rats: relevance to depression pathophysiology. Translational psychiatry. 2015;5:e682. doi: 10.1038/tp.2015.175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Liu Y, Zhao Z, Yang F, Gao Y, Song J, Wan Y. microRNA-181a is involved in insulin-like growth factor-1-mediated regulation of the transcription factor CREB1. Journal of neurochemistry. 2013;126(6):771–780. doi: 10.1111/jnc.12370. [DOI] [PubMed] [Google Scholar]
  • 46.Zhou J, Nagarkatti P, Zhong Y, Ginsberg JP, Singh NP, Zhang J, et al. Dysregulation in microRNA expression is associated with alterations in immune functions in combat veterans with post-traumatic stress disorder. PLoS One. 2014;9(4):e94075. doi: 10.1371/journal.pone.0094075. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Bunzeck N, Düzel E. Absolute coding of stimulus novelty in the human substantia nigra/VTA. Neuron. 2006;51(3):369–379. doi: 10.1016/j.neuron.2006.06.021. [DOI] [PubMed] [Google Scholar]
  • 48.Buhle JT, Kober H, Ochsner KN, Mende-Siedlecki P, Weber J, Hughes BL, et al. Common representation of pain and negative emotion in the midbrain periaqueductal gray. Social cognitive and affective neuroscience. 2013;8(6):609–616. doi: 10.1093/scan/nss038. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Samuels E, Szabadi E. Functional neuroanatomy of the noradrenergic locus coeruleus: its roles in the regulation of arousal and autonomic function part II: physiological and pharmacological manipulations and pathological alterations of locus coeruleus activity in humans. Current neuropharmacology. 2008;6(3):254. doi: 10.2174/157015908785777193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Leknes S, Lee M, Berna C, Andersson J, Tracey I. Relief as a reward: hedonic and neural responses to safety from pain. PLoS One. 2011;6(4):e17870. doi: 10.1371/journal.pone.0017870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Ernst M, Nelson EE, Jazbec S, McClure EB, Monk CS, Leibenluft E, et al. Amygdala and nucleus accumbens in responses to receipt and omission of gains in adults and adolescents. Neuroimage. 2005;25(4):1279–1291. doi: 10.1016/j.neuroimage.2004.12.038. [DOI] [PubMed] [Google Scholar]
  • 52.Cauda F, Cavanna AE, D'Agata F, Sacco K, Duca S, Geminiani GC. Functional connectivity and coactivation of the nucleus accumbens: a combined functional connectivity and structure-based meta-analysis. J Cogn Neurosci. 2011;23(10):2864–2877. doi: 10.1162/jocn.2011.21624. [DOI] [PubMed] [Google Scholar]
  • 53.Andersen SL, Teicher MH. Stress, sensitive periods and maturational events in adolescent depression. Trends in neurosciences. 2008;31(4):183–191. doi: 10.1016/j.tins.2008.01.004. [DOI] [PubMed] [Google Scholar]
  • 54.Young JC, Widom CS. Long-term effects of child abuse and neglect on emotion processing in adulthood. Child Abuse Negl. 2014;38(8):1369–1381. doi: 10.1016/j.chiabu.2014.03.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Caspi A, McClay J, Moffitt TE, Mill J, Martin J, Craig IW, et al. Role of genotype in the cycle of violence in maltreated children. Science. 2002;297(5582):851–854. doi: 10.1126/science.1072290. [DOI] [PubMed] [Google Scholar]
  • 56.Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science. 2003;301(5631):386–389. doi: 10.1126/science.1083968. [see comment] [DOI] [PubMed] [Google Scholar]
  • 57.Okbay A, Baselmans BM, De Neve JE, Turley P, Nivard MG, Fontana MA, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nat Genet. 2016 doi: 10.1038/ng.3552. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS800485-supplement-1.pdf (278.7KB, pdf)

RESOURCES