Participants
Participants were recruited from Phase II of the Adult Health and Behavior project (AHAB II), which assesses a wide range of behavioral and biological traits among middle-aged community volunteers. All participants had completed both the emotion regulation fMRI task and the Pittsburgh Sleep Quality Index (PSQI) and were in good general health. The University of Pittsburgh Institutional Review Board approved the study and all participants provided informed consent in accordance with its regulations. Participants were evaluated for current DSM-IV Axis I disorders using the Mini-International Neuropsychiatric Interview (MINI [18]) and excluded only for history of psychosis. The participants were free of medical diagnoses of cancer, stroke, diabetes requiring insulin treatment, and chronic kidney or liver disease. Additional exclusion criteria included use of psychotropic, glucocorticoid, hypolipidemic, antiarrhythmic, antihypertensive, and prescription weight loss medication. Sleep medications were allowed if they were not taken more than 7 of 14 days prior to eligibility determination.
Our initial sample included 106 unselected participants, but 8 were removed for amygdala coverage less than 90%. One additional participant was removed for abnormally high motion artifact, leaving a final sample of 97 participants (48 women; mean age 42.78 ± 7.37 years, range 30–54 years old).
Measures
Pittsburgh Sleep Quality Index (PSQI)
All participants completed the PSQI, a 19-item self-rated questionnaire for evaluating general sleep patterns over the previous month [17]. The questionnaire is scored to produce seven clinically-derived component scores, each of which is converted to a 0–3 scale where higher numbers indicate more problematic sleep. Many of the component scores are based on a single item or reflect a single calculation based on two or more items. The component scores are subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency (a measure of time spent asleep to total time spent in bed), sleep disturbances, use of sleeping medication, and daytime dysfunction. The combined score is reported as Global Sleep Quality. Global scores greater than 5 indicate clinically meaningful sleep disturbance.
Center for Epidemiologic Studies Depression Scale (CES-D)
To ensure that observed relationships were not better accounted for by co-occurring symptoms of mood disorders, depression symptoms were also evaluated using the CES-D [19]. The CES-D is a 20-item measure of depression symptoms that has been evaluated and used in many studies of psychiatric symptoms and disorders. A cutoff score of 16 has been used to differentiate those who are likely to have clinically meaningful levels of depressive symptoms from those who are not [20].
Emotion regulation paradigm
The emotion regulation task in this experiment was adapted from a previously validated paradigm [8, 13]. The task consisted of 30 negative photographs and 15 neutral photographs selected from the International Affective Picture System (IAPS) database based on published norms [21]. Negative photographs depicted bodily illness and injury (21 photographs), acts of aggression (3 photographs), members of hate groups (2 photographs), transportation accidents (2 photographs) and human waste (2 photographs). Neutral photographs depicted inanimate objects (10 photographs) or neutral scenes (5 photographs).
Prior to completing the task, subjects were instructed that when cued to “look,” they were to maintain attention on the stimulus and allow their emotional reaction to occur without attempting to change it. When cued to “decrease,” they were to attempt to reduce their emotional response through cognitive reappraisal (i.e., by thinking of something that makes the photograph seem less negative). Subjects were given examples of reappraisal strategies for specific photographs and then practiced the skill outside the MRI scanner. During the MRI scan each trial consisted of a 2 second cue to either “look” or “decrease” one’s emotional response, then a 7 second presentation of either a negative or neutral picture, then a 4 second opportunity to rate the picture, followed by a 1–3 second rest period before the next cue. During the “rate picture” phase, subjects were instructed to report their emotional reaction to each photograph on a scale of 1 to 5, where 1 indicated neutral and 5 indicated feeling strongly negative. The ratings were made using a button response pad in the participant’s right hand and recorded in E-Prime software.
Fifteen negative photographs were presented with the “look” cue and 15 were presented with the “decrease” cue. All 15 neutral photographs were presented with the “look” cue (because there is nothing to regulate in response to a neutral photograph). The stimuli were presented in pseudo-random order such that no more than 2 of the same instruction (look vs. regulate) could be presented consecutively and no more than 4 negative stimuli could be presented consecutively. Total time for the task was 11:28 minutes. This design allows for an assessment of neural activation related to the emotional valence of the stimuli (look negative > look neutral) as well as activation related to reappraisal (regulate negative > look negative).
BOLD fMRI data acquisition
Each participant was scanned using a Siemens 3 T Allegra scanner (Siemens AG, Medical Solutions, Erlangen, Germany) developed specifically for advanced brain imaging applications and characterized by increased T2* sensitivity and fast gradients that minimize echo spacing, thereby reducing echo-planar imaging (EPI) geometric distortions and improving image quality. An autoshimming procedure was conducted to minimize field inhomogeneities. A series of 34 interleaved axial slices aligned with the AC-PC plane were acquired with a gradient-echo echo planar imaging sequence (TR/TE = 2000 ms/25 ms; FOV = 200 mm, matrix size 64 × 64; 3.125 × 3.125 × 3 mm voxels; interslice skip = 0). Two initial RF excitations were performed (and discarded) to achieve steady-state equilibrium. However, the first two acquired volumes were discarded during preprocessing to further ensure steady-state equilibrium. All scanning parameters were selected to optimize the quality of the BOLD signal while maintaining a sufficient number of slices to acquire whole brain data. Before the collection of fMRI data for each participant, we acquired a reference EPI scan that we visually inspected for artifacts (e.g., ghosting) as well as good signal across the entire volume of acquisition.
BOLD fMRI data analysis
Whole-brain image analysis of all fMRI data was conducted at the Laboratory of NeuroGenetics at Duke University using the general linear model (GLM) of SPM8 (http://www.fil.ion.ucl.ac.uk/spm). Images for each participant were realigned to the first volume in the time series to correct for head motion, spatially normalized into a standard stereotactic space (Montreal Neurological Institute template) using a 12-parameter affine model (final resolution of functional images = 2 mm isotropic voxels), and smoothed to minimize noise and residual difference in gyral anatomy with a Gaussian filter, set at 6-mm full-width at half-maximum. Preprocessed data sets were analyzed using second-level random-effects models that account for both scan-to-scan and participant-to-participant variability to determine task-specific regional responses.
Variability in single-subject whole-brain functional volumes was determined using the Artifact Recognition Toolbox (http://www.nitrc.org/projects/artifact_detect). Individual whole-brain BOLD fMRI volumes meeting at least one of the following two criteria were identified as artifacts during the determination of task-specific effects: 1) significant mean-volume signal intensity variation 2(i.e., within volume mean signal greater or less than 4 standard deviations of mean signal of all volumes in time series), and 2) individual volumes where scan-to-scan movement exceeded 2 mm translation or 2° rotation in any direction. Artifacts were then treated as regressors of no interest in subsequent preprocessing steps. One participant was removed from further analyses due to abnormally high artifact (37% of whole-brain BOLD fMRI volumes). The remaining participants had, on average, 2.64% of all volumes identified as artifacts, thus we believe this approach enhanced our capacity to determine task-specific effects by minimizing the influence of volumes with substantial variability without compromising our power to detect task-specific effects by excluding a large number of volumes.
Following preprocessing, linear contrasts employing canonical hemodynamic response functions were used to estimate condition-specific (i.e. negative > neutral and regulate > look) BOLD responses for each individual. Individual contrast images (i.e., weighted sum of the beta images) were then used in second-level random effects models accounting for scan-to-scan and participant-to-participant variability to determine mean condition-specific regional responses using one-sample t-tests. A voxel-level statistical threshold of p < 0.05, FWE corrected for multiple comparisons (across the anatomical amygdala ROIs for negative > neutral, and across whole brain for regulate > look) was applied. An additional extent threshold of 10 contiguous voxels was applied to both ROI and whole brain analyses.
Based on previous findings [7, 13, 22] we selected the amygdala as a region of interest (ROI) where we expected to find effects of emotional reactivity (negative > neutral contrast). A bilateral amygdala ROI mask was created from the Automated Anatomical Labeling (AAL) atlas [23]. Because of the potential for signal loss and noise often observed in the amygdala and adjacent regions, single-subject BOLD fMRI data were included in subsequent analyses only if there was a minimum of 90% signal coverage in the amygdala masks bilaterally. Although we anticipated finding main effects of task in regions previously identified [8, 13], these analyses were completed using whole brain analyses. No additional ROI masks were created.
BOLD parameter estimates from clusters exhibiting condition-effects (negative > neutral; regulate > look) were extracted using the VOI tool in SPM8 and exported for analyses in R and SPSS (v.18). Extracting parameter estimates from functional clusters activated by our fMRI paradigm, rather than clusters specifically correlated with our independent variables of interest, precludes the possibility of any correlation coefficient inflation that may result when an explanatory covariate is used to select a region of interest [24]. We have used this more conservative and rigorous analytic strategy in recent studies [25, 26].
Statistical analyses
The influence of sleep duration on emotional reactivity and emotion regulation was investigated using regression analyses. Because sleep duration was measured prior to the emotion regulation task and because it is believed to play a causal role in influencing subjective responses and neural responses during emotional reactivity and regulation, it was treated as an independent variable in all regression analyses. Dependent variables included subjective emotional responses to photographs as well as extracted BOLD parameter estimates from max voxels of clusters showing significant condition-effects and ratings of emotional reactions during each condition. Dependent variables were analyzed to ensure they were approximately normally distributed prior to entry into regression analyses. Relationships among subjective emotional reactions and extracted BOLD parameter estimates were investigated using Pearson correlations. Because we made a priori predictions that poorer sleep would predict less brain activation during emotion regulation and less subjective regulatory success, one-tailed significance tests were used in these analyses. For all other analyses, two-tailed significance tests were used.