1067 Replicative and Non-Replicative Nightmares in the Development of Posttraumatic Stress Disorder

Introduction

Nightmares are a frequent and disturbing symptom of posttraumatic stress disorder (PTSD). They are associated with sleep disruption and increased psychopathology. There is growing evidence that different types of nightmares may differ in their effects on psychopathology. Previous findings suggest that nightmares that are close replications of the experienced traumatic event might be especially important in the development of PTSD. This study investigated trauma-related (replicative) and non-trauma-related (non-replicative) nightmares as predictors of PTSD in a civilian sample.

Methods

Participants were recruited from the general public of the greater Boston area. The sample consisted of 108 participants who had experienced a psychological trauma in the past 2 years (e.g. sexual or physical assaults and accidents). The criteria for PTSD were met by 49% of participants. PTSD diagnosis was assessed using the Structured Clinical Interview for DSM-IV-TR Axis I Disorders-Non-Patient Edition. Participants received an Actiwatch 2 (Philips Respironics, Bend, OR) and a sleep diary for sleep measurements over an average of 14 consecutive nights. The diary included a prospective nightmare assessment and an item assessing the relatedness of each nightmare to traumatic events. Logistic regression analyses were performed with PTSD as the categorical outcome variable.

Results

Our analyses showed that replicative nightmares were the only statistically significant predictor of PTSD (OR = 1.2, p = .027), while controlling for age, sex, time since the traumatic event, and actigraphy total sleep time and minutes awake after sleep onset. All of these variables, including non-replicative nightmares, did not significantly predict PTSD in our analyses.

Conclusion

This study confirms and adds to the existing knowledge of nightmares and the importance of the degree to which they replicate the trauma in the development of PTSD. These findings underline the potential role of specific nightmare treatments after traumatic events, with a special focus on replicative nightmares.

Support

R01MH109638

1116 Interactions Among Interoceptive Sensitivity And Insomnia Symptoms In Individuals With Generalized Anxiety Disorder

Introduction

Abnormal interoception is believed to contribute to anxiety disorders as well as possibly to Insomnia Disorder. We therefore hypothesized that interoceptive sensitivity in persons with Generalized Anxiety Disorder (GAD) would vary with sleep quality and would differ between GAD patients with and without insomnia.

Methods

29 subjects (86% female) who reported GAD-7 scores ≥ 10 underwent psychiatric and sleep-disorders interviews and met DSM-5 criteria for GAD. Participants were assigned into an insomnia group if the Insomnia Severity Index (ISI) score exceeded 12 (N=18) or a non-insomnia group if ISI score < 11 (N=11). Participants completed approximately 2 weeks of actigraphy and sleep diaries (mean=14.09 days; SD=4.85) as well as online surveys of sleep quality including the Pittsburg Sleep Quality Index (PSQI) and interoceptive sensitivity including the Multidimensional Assessment of Interoceptive Awareness, the somatic anxiety scale of the State-Trait Inventory for Cognitive and Somatic Anxiety-Trait (STICSA-T) and the Anxiety Sensitivity Index. In addition, 28 participants underwent the Schandry heartbeat-counting task, an objective measure of interoceptive sensitivity.

Results

GAD-7 scores did not differ between insomnia and non-insomnia groups (t(18.36)=0.074, p = 0.941). However, insomniacs reported significantly higher scores on the somatic anxiety scale of the STICSA than non-insomniacs (t(17.71)=2.094, p=0.051). STICSA-T somatic anxiety scores positively correlated with greater PSQI scores (R=0.588; p=0.002) and trended negatively with actigraphy sleep-efficiency (R = -0.384; p=0.085). Non-insomniacs outperformed insomniacs in the Schandry task (t(25.9)=2.21, p=0.036). Higher Schandry scores were positively correlated with greater total sleep time (R=0.515; p=0.014) and lower ISI scores (R= -0.387; p=0.042).

Conclusion

Results suggest that GAD patients with insomnia have lesser objective interoceptive sensitivity and more subjective somatic anxiety than those without insomnia. Greater somatic anxiety predicted poorer subjective and objective sleep quality. Hence interoceptive sensitivity may vary with sleep disturbance in GAD.

Support

R21MH115279

0065 Initial Findings on Associations Between Sleep and Clinical Measures with Neural Activations Accompanying Fear Conditioning and Extinction in Trauma-Exposed Individuals

Introduction

We examined associations of sleep and hyperarousal with neural responses to a fear conditioning and extinction protocol in trauma-exposed individuals. We hypothesized, greater hyperarousal, poorer sleep quality and more nightmares would accompany greater activation of the salience network (associated with fear) and lesser activation of the prefrontal cortex (PFC; associated with fear regulation) throughout this protocol.

Methods

Persons exposed to trauma within the past 2 years (N=119, 43 with Post-traumatic stress disorder; PTSD) completed the PTSD Checklist-5 (PCL-5), two weeks of actigraphy and sleep/nightmare diaries, and a 2-day fear-conditioning and extinction protocol during fMRI. Hyperarousal items from PCL-5, sleep quality, and nightmare frequency were used to predict fMRI contrasts representing (1) initial activations to reinforced conditioned stimuli (CS+) during fear conditioning, (2) change in neural activation to CS+s across extinction learning, and, (3) after 24 hours (extinction recall), selective activation to an extinguished (CS+E) versus an un-extinguished CS+.

Results

During fear conditioning, hyperarousal was positively correlated with activation to the CS+ in the right lateral PFC, whereas nightmare frequency was negatively correlated with activations in bilateral orbitofrontal cortex (OFC). Across extinction learning, sleep onset latency (SOL) was negatively correlated with increased activation to CS+ in bilateral insular and dorsal and middle anterior cingulate cortices (salience regions). At extinction recall, nightmare frequency was negatively correlated with selective activation to the CS+E in the left insular cortex.

Conclusion

Except that fewer nightmares predicted greater OFC activation during fear conditioning, results did not support hypothesized relationships of hyperarousal, poor sleep and nightmares with increased salience network and decreased PFC activation to fear-related stimuli.

Support

Funding: R01MH109638

1081 Associations Among REM Density And Parasympathetic Activity, Nightmares, And Hyperarousal In Trauma-exposed Individuals

Introduction

Individuals with posttraumatic stress disorder (PTSD) exhibit autonomic hyperarousal and nightmares. We hypothesized that REM density (REMD) and REM heart rate variability would predict self-reported hyperarousal, nightmares, and PTSD diagnosis in trauma-exposed individuals.

Methods

Ninety-nine individuals (aged 18-40, 68 females) exposed to a DSM-5 PTSD criterion-A trauma within the past two years (48 meeting PTSD criteria) completed a night of ambulatory polysomnography (PSG) preceded by an acclimation night. REMD in scored sleep recordings were computed using the Matlab program written by Benjamin Yetton. Indices of parasympathetic tone during REM were computed using Kubios software and included Average Root Mean Square of the Successive Differences (RMSSD) and High Frequency power (HFpower). Participants completed two weeks of sleep diaries with nightmare questionnaire and completed the Clinician-Administered PTSD Scale (CAPS-5) and the PTSD Checklist for DSM-5 (PCL-5). Hyperarousal-item scores were computed from the PCL-5 without the sleep item (PCLhyp) and from the CAPS-5 (CAPShyp), and these scores (with their sleep items) were combined into a Composite Hyperarousal Index (CHI). Nightmare rate was the proportion of sleep diaries reporting a nightmare. Simple regressions measured associations among REMD, REM parasympathetic indices, hyperarousal measures, and nightmare rate.

Results

REMD did not significantly predict PTSD diagnosis or hyperarousal scores but did predict decreased parasympathetic activity for both RMSSD (p= 0.002, R= -0.316) and HFpower (p= 0.016 R= -0.250). REMD predicted increased nightmare rate (p= 0.011 R= 0.262). Parasympathetic tone was negatively correlated with CAPShyp, PCLhyp, and CHI for both RMSSD (p= 0.04, 0.011, <0.000, respectively) and HFpower (p= 0.051, 0.021, 0.010, respectively). Lower parasympathetic tone also predicted PTSD diagnosis with both RMSSD (p=0.012, t=2.559) and HFpower (p=0.010, t=2.627), but did not predict nightmare rate.

Conclusion

REMD predicted decreased parasympathetic tone and higher nightmare rate. Parasympathetic tone, but not REMD, predicted hyperarousal and PTSD diagnosis.

Support

R01MH109638

1071 Subjective Measures Of Hyperarousal Predict Subjective Longitudinal And Retrospective Measures Of Sleep Quality In Individuals Exposed To Trauma

Introduction

Hyperarousal and disturbed sleep are intrinsic symptoms of posttraumatic stress disorder (PTSD). We explored whether self-reported indices of hyperarousal predict longitudinally measured objective, subjective, and retrospective evaluations of sleep quality in trauma-exposed individuals.

Methods

Individuals exposed to a DSM-5 PTSD Criterion-A traumatic event within the past two years (N=130, 91 females), aged 18-40 (mean 24.43, SD 5.30), 51.54% of whom met DSM-5 criteria for PTSD, completed 14 days of actigraphy and sleep diaries. Participants also completed the PTSD Checklist for DSM-5 (PCL-5), the Clinician-Administered PTSD Scale (CAPS-5), published Hyperarousal (HAS) and Hypervigilance (HVQ) scales, and the Pittsburgh Sleep Quality Index (PSQI) (N=108-125 for different scales). Mean total sleep time (TST), sleep onset latency (SOL), sleep efficiency (SE) and sleep midpoint were calculated from actigraphy and subjective SOL, SE, number of awakenings, and time spent awake from diaries. Simple regressions were used to predict associations of the PCL-5, HAS, and HVQ scores with measures of sleep quality.

Results

Hyperarousal indices predicted diary but not actigraphy measures of sleep quality. Longer diary-reported SOL was predicted by higher scores for: PCL-5 total score (R=0.290, p=0.001), PCL-5 hyperarousal items without the sleep item (R=0.261, p=0.004), and HAS without sleep items (R=0.220, p=0.016). Diary-reported number of awakenings and wake time after sleep onset were predicted by higher HAS scores without the sleep question: (R=0.373, p<0.001; r=0.352, p<0.001). Similarly, all hyperarousal indices significantly predicted PSQI global score (PCL-5: R=0.482, p<0.001; PCL-5 hyperarousal: R=0.389, p<0.001; HVQ: R=0.214, p=0.017; HAS without sleep question: R=0.415, p<0.001).

Conclusion

Self-reported hyperarousal measures predict subjective longitudinal (especially SOL) and retrospective measures, but not objective measurements of sleep quality. Similar discrepancies between self-reported and objective measures of sleep quality have been reported in patients with insomnia disorder. Cognitive-behavioral therapy for insomnia may be especially effective in treating post-traumatic sleep disturbances.

Support

R01MH109638