A High Density EEG Comparison of Sleep Patterns in Insomnia

Completed

Conditions: Primary Insomnia

Interventions: Behavioral: Serial awakenings

Registration Number: NCT01960452

Lead Sponsor: University of Wisconsin, Madison

Brief Summary: Insomnia, defined as a subjective report of difficulty initiating sleep, maintaining sleep, and/or non-restorative sleep, leads to significant daytime dysfunction and increased health risks. A commonly held hypothesis is that insomnia is caused by a state of hyperarousal, but the neurobiological mechanisms of hyperarousal in insomnia are poorly understood, in part because of limitations in our ability to image the brain during normal human sleep with sufficient temporal resolution. Furthermore, the efficacy of insomnia treatment is judged by subjective report of the patient and demonstration of changes in sleep latency and/or sleep amount which are generally small in magnitude; there are currently no data to demonstrate that insomnia treatments correct any functional abnormalities in the sleep process that likely contribute to neurobehavioral abnormalities and health risks. The goals of the proposed study are to use high density EEG to define abnormalities in specific aspects of sleep in insomnia patients compared to healthy sleeping control subjects to define biomarkers that will both increase our understanding of the pathophysiology of insomnia as well as provide targets to assess treatments for insomnia.

Detailed Description: Recent advances in electroencephalographic recording techniques have produced new ways to probe the process and function of sleep. Through the use of high-density EEG (hdEEG, up to 256 channels), it is possible to approach the spatial resolution of other brain imaging modalities while affording the millisecond temporal resolution of EEG and providing a direct measure of the underlying brain activity, unlike the indirect and/or secondary biophysical signals of brain hemodynamics/metabolism obtained with PET or SPECT that are suboptimal for exploring the short-lived spatio-temporal dynamics of many brain processes.

Here we used hdEEG to try to characterize topographic changes in sleep EEG expression in individuals with insomnia compared to normal controls. We further used serial awakenings to determine if individuals with insomnia were more likely to subjectively report being awake when they were sleeping, and study instances where a direct confirmation of sleep was followed by a subjective report of wakefulness to see if they are characterized by changes in EEG oscillations.

Recruitment & Eligibility

Status: COMPLETED

Sex: All

Target Recruitment: 45

Inclusion Criteria

18-45 years old
English speaking, reading, and writing
For control subjects: Insomnia Severity Index (ISI) less than or equal to 6 and does not meet criteria for insomnia
For insomnia subjects: ISI greater than or equal to 7, meets criteria for insomnia, and reports insomnia symptoms for at least 6 months

Exclusion Criteria

Imminent danger to self or others
Clinical diagnosis of dementia
Active Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV Axis I disorder or alcohol or drug dependence or abuse
Other sleep disorders aside from insomnia
History of significant head trauma or loss of consciousness over 30 minutes
Regular use of psychotropic medications in past 4 weeks
Regular tobacco use
Drinking more than 3 caffeinated beverages per day
Significant neurological or medical illness
Pregnant, less than 6 months post-partum, or planning to become pregnant during the study
Left-handedness
Body Mass Index (BMI) greater than 40
Apnea Hypopnea Index (AHI) greater than 10 on Apnea Link
Mini mental status exam score less than 27

Study & Design

Study Type: OBSERVATIONAL

Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
Primary insomnia	Serial awakenings	-
Healthy sleeping controls	Serial awakenings	-

Primary Outcome Measures

Name	Time	Method
EEG Power During Sleep	Individual night of sleep recorded on average within 4 weeks of enrollment (Night 1, Baseline EEG)	The difference in spontaneous NREM sleep (stage N2 or N3) EEG power between subjects with insomnia and good sleeping controls measured with 256 channel high-density EEG equipment. Reported here is the median spectral power estimated using Welch's method (MATLAB function pwelch) across all 6 second epochs staged by a registered sleep technician as NREM sleep stage N2 or N3 sleep. The median spectral estimate across all NREM epochs was used to estimate spectral power for each subject as this value is robust to outliers from individual sleep epochs. The Welch's method done in this way estimates frequency content in .33 Hz bins which are then averaged across the spindle frequency band range (12 - 16Hz) from the approximate CP1 electrode (channel 89). The value reported below is the MEAN of the median spectral estimate. Higher values represent more spindle band activity. This frequency range and location showed the most consistent difference between groups across outcome measures.
EEG Power Examined Before Arousal From Sleep on Serial Awakening Night	30 seconds of NREM sleep prior to serial awakening (Night 2, Serial Awakening EEG)	The difference in 30 seconds of spontaneous NREM (stage N2 or N3) sleep EEG power prior to serial awakening between subjects with insomnia and good sleeping controls measured with 256 channel high-density EEG equipment. Reported here is the median spectral power estimated using Welch's method (2 second epochs) of the 30 seconds of NREM sleep stage N2 or N3 sleep immediately prior to the 40dB tone played to initiate a serial awakening sequence. The median spectral estimate across epochs was used to estimate spectral power for each subject as this value is robust to outliers from individual epochs. The Welch's method done in this way estimates frequency content in .5 Hz bins which are then averaged across the spindle frequency band range (12 - 16Hz) from the approximate CP1 electrode (channel 89). Each subject had between 4 - 10 serial awakenings from NREM sleep. The value reported below is the MEAN of the median spectral estimate. Higher values represent more spindle band activity.
EEG Power Examined as Subjects Fall Asleep on Serial Awakening Night	30 seconds of falling asleep EEG immediately after the first non-waking epoch following a serial awakening and proceeding 5 minutes of stable sleep (Night 2, Serial Awakening EEG)	The difference in 30 seconds of spontaneous falling asleep EEG power between subjects with insomnia and good sleeping controls measured with 256 channel high-density EEG equipment. Reported here is the median spectral power estimated using Welch's method with 2 second epochs from the 30 seconds of sleep immediately following a serial awakening (starting after the last waking epoch) that resulted in the subject reaching 5 minutes of stable sleep averaged across the spindle frequency band range 12 - 16Hz from the approximate CP1 electrode (channel 89) and across all serial awakening falling asleep periods for each subject. Each subject had between 4 - 8 falling asleep periods that resulted in 5 minutes of stable sleep within 30 minutes of the serial awakening attempt. The value reported below is the MEAN of the median spectral estimate. Higher values represent more spindle band activity.