This outcome tested whether amygdala connectivity with regions of the mPFC was changed following treatment using psychophysiological interaction (PPI) analysis for this contrast/task. Regions of the mPFC include: dorsal anterior cingulate cortex (dACC), ventromedial prefrontal cortex (vmPFC), dorsomedial prefrontal cortex (dmPFC), subgenual anterior cingulate cortex (sgACC), pregenual anterior cingulate cortex (pACC). PPI analyses produce a beta weight for each participant at each timepoint, and represents the degree to which the connectivity of the amygdala and mPFC is modulated by task conditions. A positive value means average connectivity increases in the task-contrast, and a positive value for the change score means an increase in average connectivity following CBT-I treatment. It is theorized that higher amygdala connectivity is associated with better outcomes.

This measure is of the Beck Depression Inventory-II total score after excluding one sleep item.

The BDI-II is a 21-item self-report scale with high validity and reliability that assesses the severity of depression symptoms. The depression items consist of: sadness, pessimism, past failure, loss of pleasure, guilty feelings, punishment feelings, self-dislike, self-criticalness, suicidal thoughts or wishes, crying, agitation, loss of interest, indecisiveness, worthlessness, loss of energy, irritability, changes in appetite, concentration difficulty, tiredness or fatigue, and loss of interest in sex. Items are scored from 0 to 3, and summed to create an overall score of 0 to 63. higher scores indicate greater levels of severity. The ranges for depression are: 0-13 minimal, 14-19 mild, 20-28 moderate, and 29-63 severe. A negative change score means that average depression symptom severity was reduced following CBT-I treatment.

Sleep efficiency (SE) is the percentage of total time in bed actually spent sleeping. Based on the overnight PSG sleep recording, SE will be calculated as the total time (minutes) spent asleep (sum of Stages N1, N2, N3, and REM) divided by the total time (minutes) in bed, and multiplied by 100. A positive change score means average sleep efficiency increased following CBT-I treatment.

The Columbia Suicide Severity Rating Scale (CSSRS) is a 12-item checklist that was designed to quantify the severity of suicidal ideation and behavior. It is composed of two parts. The first six questions ask about suicidal ideation and behavior in the past month while the last six questions ask about suicidal ideation and behavior since the last visit. The CSSRS has been proven to be reliable and valid. It has also been shown to have high sensitivity and specificity to the different suicidal behavior classifications. The CSSRS does not provide a numerical score but categorizes risk levels based on responses. We report the proportions of risk at each timepoint.

Sleep Onset Latency (SOL) is the time (minutes) from "lights out" to actually falling asleep (sleep onset).

Number of Arousals is determined by number of times of awakening as seen on the actigraph data.

Wake After Sleep Onset (WASO) are periods of wakefulness occurring after sleep onset, before final awakening (sleep offset).

Total Sleep Time (TST) is the total time spent asleep, from the start of sleep onset to sleep offset subtracting any periods of wakefulness.

Sleep Efficiency (SE) is calculated as TST divided by total time spent in bed, multiplied by 100.

Sleep onset latency is the time it takes to fall asleep, specifically the amount of time in minutes from "LightsOff", which is the time at which the participant started trying to sleep, to stage 1 sleep. A negative change score means it took less time to fall asleep following CBT-I treatment.

Number of Arousals is determined by number of times of awakening by EEG changes. A negative change score means on average there were fewer overnight arousals following CBT-I treatment.

Wake After Sleep Onset (WASO) are periods of wakefulness occurring after sleep onset, before final awakening (sleep offset) measured by EEG changes. A negative change value means there was less WASO on average after CBT-I treatment.

Total Sleep Time (TST) is the total time (minutes) spent asleep, from the start of sleep onset to sleep offset, subtracting any periods of wakefulness. TST includes stages N1, N2, N3, and REM sleep. A positive change score means the average TST increased following CBT-I.

Fronto-central EEG power spectral density analysis associated with sleep stages will be calculated in the Delta (0.5-Hz), Theta (4-7Hz), Alpha (7-11Hz), Sigma (12-15Hz), Beta-1 (15-20Hz), Beta-2 (20-35Hz) and Gamma (35-45Hz) bands, according to published methods. A positive change score means there was an increase in absolute power in the specified frequency band following CBT-I treatment.

Subjective ratings of sleep disturbance and insomnia severity will be assessed with the Insomnia Severity Index. The Insomnia Severity Index (ISI) is a 7-item self-report measure of insomnia type, severity, and impact on functioning. The items consist of severity of sleep onset, sleep maintenance, early morning awakenings, sleep dissatisfaction, interference with daytime functioning, noticeability of sleep problems by others, and distress caused by sleep difficulties. Items are scored from 0 to 4 (0 = no problem, 4 = very severe problem), then summed to create an overall score of 0 to 28. Score ranges of insomnia are: 0-7 absent, 8-14 sub-threshold, 15-21 moderate, and 22-28 severe. The ISI has good validity and reliability. A negative change score means average insomnia symptoms improved following CBT-I treatment.

The SF-36 measures health-related quality of life based on eight domains: physical activity, social activities, limitations in activities due to physical health problems, bodily pain, general mental health, limitations in activities due to emotional problems, vitality, and general health perceptions. Items are recoded then averaged together to create each an average score for all items that the respondent answered. The eight subscales are then into two component summary t-scores (Mental and Physical Component Summary t-scores), each with a mean of 50 and a standard deviation of 10. A t-score higher than 50 means better mental or physical health than the general population, and a t-score below 50 means worse than the general population. Instructions for scoring these component scores recommends that they be set to missing if any subscales are missing. A positive change score means the Component Summary Score improved following CBT-I.

The BAI is a 21-item self-report scale that assesses the severity of anxiety symptoms.

Items are scored from 0 to 3 (0 = not at all, 3 = severe), then summed to create an overall score range of 0 to 63. Higher scores indicate greater levels of severity, and the ranges for anxiety levels are: 0-9 normal to minimal, 10-18 mild to moderate, 19-29 moderate to severe, and 30-63 severe. The BAI consists of two factors: somatic and cognitive. A negative change score means that average anxiety symptom severity was reduced following CBT-I treatment.

RSA is the phenomenon of an increased heart rate during inhalation and a decreased heart rate during exhalation. Since these fluctuations are controlled mainly by vagal influences on the heart, RSA serves as a reliable metric for measuring parasympathetic activity. RSA has been proven to be a reliable measure of emotion regulation and emotional responding in numerous studies.

The Nonconscious condition of the Facial Expressions of Emotion task measures subliminal (with backward masking) emotional face processing. Amygdala activation while viewing threat-related emotional faces relative to neutral faces was quantified using fMRI as a marker of Emotion Regulation Network engagement. Blood-oxygenation level dependent (BOLD) signal change before and after CBT-I treatment was compared by modeling the activity of the amygdala while viewing emotional faces using generalized linear models, producing beta weights for each participant and timepoint. A positive beta-weight at pre-treatment means that the amygdala increased its activity in response to emotional faces, relative to neutral faces. A negative value for the change in amygdala activation means that average amygdala reactivity decreased following treatment. It is theorized that higher amygdala emotional reactivity is associated with worse outcomes.

Participants who completed pre-treatment baseline, and who had interpretable image data at the respective timepoint.

Participants are asked to "look" or "decrease" their emotional response to negative and neutral valence images taken from the International Affective Picture System. Amygdala activation while viewing emotional scenes relative to neutral scenes, and while passively viewing emotional scenes relative to down-regulating emotion, was quantified using fMRI as a marker of Emotion Regulation Network engagement. Blood-oxygenation level dependent (BOLD) signal change before and after CBT-I treatment was compared by modeling the activity of the amygdala while viewing emotional scenes or while downregulating using generalized linear models, producing beta weights for each participant and timepoint. A positive beta-weight at pre-treatment means that the amygdala increased its activity in response to the task demands, and a negative value means that average amygdala reactivity decreased following treatment. It is theorized that higher amygdala emotional reactivity is associated with worse outcomes

Participants who completed pre-treatment baseline, and who had interpretable image data at the respective timepoint.

This outcome tested whether amygdala connectivity with regions of the mPFC was changed following treatment using psychophysiological interaction (PPI) analysis for this contrast/task. Regions of the mPFC include: dorsal anterior cingulate cortex (dACC), ventromedial prefrontal cortex (vmPFC), dorsomedial prefrontal cortex (dmPFC), subgenual anterior cingulate cortex (sgACC), pregenual anterior cingulate cortex (pACC). PPI analyses produce a beta weight for each participant at each timepoint, and represents the degree to which the connectivity of the amygdala and mPFC is modulated by task conditions. A positive value means average connectivity increases in the task-contrast, and a positive value for the change score means an increase in average connectivity following CBT-I treatment. It is theorized that higher amygdala connectivity is associated with better outcomes.

Participants who completed pre-treatment baseline, and who had interpretable image data at the respective timepoint.

This measure is of the Beck Depression Inventory-II total score after excluding one sleep item.

The BDI-II is a 21-item self-report scale with high validity and reliability that assesses the severity of depression symptoms. The depression items consist of: sadness, pessimism, past failure, loss of pleasure, guilty feelings, punishment feelings, self-dislike, self-criticalness, suicidal thoughts or wishes, crying, agitation, loss of interest, indecisiveness, worthlessness, loss of energy, irritability, changes in appetite, concentration difficulty, tiredness or fatigue, and loss of interest in sex. Items are scored from 0 to 3, and summed to create an overall score of 0 to 63. higher scores indicate greater levels of severity. The ranges for depression are: 0-13 minimal, 14-19 mild, 20-28 moderate, and 29-63 severe. A negative change score means that average depression symptom severity was reduced following CBT-I treatment.

All participants who competed pre-treatment baseline and with interpretable data at the respective timepoint.

Sleep efficiency (SE) is the percentage of total time in bed actually spent sleeping. Based on the overnight PSG sleep recording, SE will be calculated as the total time (minutes) spent asleep (sum of Stages N1, N2, N3, and REM) divided by the total time (minutes) in bed, and multiplied by 100. A positive change score means average sleep efficiency increased following CBT-I treatment.

Participants who completed pre-treatment baseline, and who had interpretable image data at the respective timepoint.

The Beck Scale of Suicidal Ideation (BSSI) is designed to assess the severity of suicidal ideation over the past week. The total score is derived from the sum of the first 19 items, creating an overall score ranging from 0 to 38. Scores of 0 are interpreted as no suicidal ideation, 1-8 as low levels, 9-16 as moderate levels, and 17-38 as high levels. A negative change score means an average reduction in suicidal ideation following CBT-I treatment.

Participants who completed pre-treatment baseline, and who had interpretable data at the respective timepoint.

The Columbia Suicide Severity Rating Scale (CSSRS) is a 12-item checklist that was designed to quantify the severity of suicidal ideation and behavior. It is composed of two parts. The first six questions ask about suicidal ideation and behavior in the past month while the last six questions ask about suicidal ideation and behavior since the last visit. The CSSRS has been proven to be reliable and valid. It has also been shown to have high sensitivity and specificity to the different suicidal behavior classifications. The CSSRS does not provide a numerical score but categorizes risk levels based on responses. We report the proportions of risk at each timepoint.

Participants who completed pre-treatment baseline, and who had interpretable CSSRS data at the respective timepoint.

Sleep Onset Latency (SOL) is the time (minutes) from "lights out" to actually falling asleep (sleep onset).

Number of Arousals is determined by number of times of awakening as seen on the actigraph data.

Wake After Sleep Onset (WASO) are periods of wakefulness occurring after sleep onset, before final awakening (sleep offset).

Total Sleep Time (TST) is the total time spent asleep, from the start of sleep onset to sleep offset subtracting any periods of wakefulness.

Sleep Efficiency (SE) is calculated as TST divided by total time spent in bed, multiplied by 100.

Sleep onset latency is the time it takes to fall asleep, specifically the amount of time in minutes from "LightsOff", which is the time at which the participant started trying to sleep, to stage 1 sleep. A negative change score means it took less time to fall asleep following CBT-I treatment.

Participants who completed pre-treatment baseline, and who had interpretable PSG data at the respective timepoint.

Number of Arousals is determined by number of times of awakening by EEG changes. A negative change score means on average there were fewer overnight arousals following CBT-I treatment.

Participants who completed pre-treatment baseline, and who had interpretable PSG data at the respective timepoint.

Wake After Sleep Onset (WASO) are periods of wakefulness occurring after sleep onset, before final awakening (sleep offset) measured by EEG changes. A negative change value means there was less WASO on average after CBT-I treatment.

Participants who completed pre-treatment baseline, and who had interpretable PSG data at the respective timepoint.

Total Sleep Time (TST) is the total time (minutes) spent asleep, from the start of sleep onset to sleep offset, subtracting any periods of wakefulness. TST includes stages N1, N2, N3, and REM sleep. A positive change score means the average TST increased following CBT-I.

Participants who completed pre-treatment baseline, and who had interpretable PSG data at the respective timepoint.

Fronto-central EEG power spectral density analysis associated with sleep stages will be calculated in the Delta (0.5-Hz), Theta (4-7Hz), Alpha (7-11Hz), Sigma (12-15Hz), Beta-1 (15-20Hz), Beta-2 (20-35Hz) and Gamma (35-45Hz) bands, according to published methods. A positive change score means there was an increase in absolute power in the specified frequency band following CBT-I treatment.

All participants who competed pre-treatment baseline and with interpretable PSG data at the respective timepoint.

Subjective ratings of sleep disturbance and insomnia severity will be assessed with the Insomnia Severity Index. The Insomnia Severity Index (ISI) is a 7-item self-report measure of insomnia type, severity, and impact on functioning. The items consist of severity of sleep onset, sleep maintenance, early morning awakenings, sleep dissatisfaction, interference with daytime functioning, noticeability of sleep problems by others, and distress caused by sleep difficulties. Items are scored from 0 to 4 (0 = no problem, 4 = very severe problem), then summed to create an overall score of 0 to 28. Score ranges of insomnia are: 0-7 absent, 8-14 sub-threshold, 15-21 moderate, and 22-28 severe. The ISI has good validity and reliability. A negative change score means average insomnia symptoms improved following CBT-I treatment.

Participants who completed pre-treatment baseline, and who had ISI data at the respective timepoint.

The SF-36 measures health-related quality of life based on eight domains: physical activity, social activities, limitations in activities due to physical health problems, bodily pain, general mental health, limitations in activities due to emotional problems, vitality, and general health perceptions. Items are recoded then averaged together to create each an average score for all items that the respondent answered. The eight subscales are then into two component summary t-scores (Mental and Physical Component Summary t-scores), each with a mean of 50 and a standard deviation of 10. A t-score higher than 50 means better mental or physical health than the general population, and a t-score below 50 means worse than the general population. Instructions for scoring these component scores recommends that they be set to missing if any subscales are missing. A positive change score means the Component Summary Score improved following CBT-I.

Participants who completed the pre-treatment baseline, and who had interpretable SF-36 data at the respective timepoint.

The BAI is a 21-item self-report scale that assesses the severity of anxiety symptoms.

Items are scored from 0 to 3 (0 = not at all, 3 = severe), then summed to create an overall score range of 0 to 63. Higher scores indicate greater levels of severity, and the ranges for anxiety levels are: 0-9 normal to minimal, 10-18 mild to moderate, 19-29 moderate to severe, and 30-63 severe. The BAI consists of two factors: somatic and cognitive. A negative change score means that average anxiety symptom severity was reduced following CBT-I treatment.

All participants who competed pre-treatment baseline and with data at the respective timepoint.

RSA is the phenomenon of an increased heart rate during inhalation and a decreased heart rate during exhalation. Since these fluctuations are controlled mainly by vagal influences on the heart, RSA serves as a reliable metric for measuring parasympathetic activity. RSA has been proven to be a reliable measure of emotion regulation and emotional responding in numerous studies.

Analysis of Nonconscious Threat > Neutral amygdala reactivity was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values from linear mixed effects models were False Discovery Rate-adjusted (Benjamini-Hochberg method) for the number of contrasts (3) in the Nonconscious Faces task. The a priori threshold for statistical significance was < 0.05.

Analysis of Emotion Regulation Scenes task Look Negative > Decrease Negative amygdala reactivity was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values from linear mixed effects models were False Discovery Rate-adjusted (Benjamini-Hochberg method) for the number of contrasts (2) in the Emotion Regulation Scenes task. The a priori threshold for statistical significance was < 0.05.

Analysis of amygdala-dmPFC connectivity from the Conscious Fear > Neutral task/contrast was done by applying linear mixed effects models testing the effect of treatment-time on connectivity. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values from linear mixed effects models were False Discovery Rate-adjusted (Benjamini-Hochberg method) for the number of mPFC regions (5). The a priori threshold for statistical significance was < 0.05.

Analysis of amygdala-pACC connectivity from the Conscious Fear > Neutral task/contrast was done by applying linear mixed effects models testing the effect of treatment-time on connectivity. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values from linear mixed effects models were False Discovery Rate-adjusted (Benjamini-Hochberg method) for the number of mPFC regions (5). The a priori threshold for statistical significance was < 0.05.

Analysis of amygdala-sgACC connectivity from the Conscious Fear > Neutral task/contrast was done by applying linear mixed effects models testing the effect of treatment-time on connectivity. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values from linear mixed effects models were False Discovery Rate-adjusted (Benjamini-Hochberg method) for the number of mPFC regions (5). The a priori threshold for statistical significance was < 0.05.

Analysis of amygdala-vmPFC connectivity from the Conscious Fear > Neutral task/contrast was done by applying linear mixed effects models testing the effect of treatment-time on connectivity. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values from linear mixed effects models were False Discovery Rate-adjusted (Benjamini-Hochberg method) for the number of mPFC regions (5). The a priori threshold for statistical significance was < 0.05.

Analysis of change in PSG Theta absolute power was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.

Analysis of change in PSG Alpha absolute power was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.

Analysis of change in PSG Sigma absolute power was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.

Analysis of change in PSG Beta-1 absolute power was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.

Analysis of change in PSG Beta-2 absolute power was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.

Analysis of change in PSG Gamma absolute power was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.

Analysis of change in SF-36 Physical Component Score was done by applying linear mixed effects models testing the effect of treatment-time. Random intercepts at the individual participant level are included to account for the clustering of observations within individuals across time. All statistical models include age and sex as covariates of non-interest.

P-values are uncorrected from linear mixed effects models. The a priori threshold for statistical significance was < 0.05.