Design and validation of a computer-based sleep-scoring algorithm

A computer-based sleep scoring algorithm was devised for the real time scoring of sleep-wake state in Wistar rats. Electroencephalogram (EEG) amplitude (microV(rms)) was measured in the following frequency bands: delta (delta; 1.5-6 Hz), theta (Theta; 6-10 Hz), alpha (alpha; 10.5-15 Hz), beta (beta; 22-30 Hz), and gamma (gamma; 35-45 Hz). Electromyographic (EMG) signals (microV(rms)) were recorded from the levator auris longus (neck) muscle, as this yielded a significantly higher algorithm accuracy than the spinodeltoid (shoulder) or temporalis (head) muscle EMGs (ANOVA; P=0.009). Data were obtained using either tethers (n=10) or telemetry (n=4). We developed a simple three-step algorithm that categorizes behavioural state as wake, non-rapid eye movement (NREM) sleep, rapid eye movement (REM) sleep, based on thresholds set during a manually-scored 90-min preliminary recording. Behavioural state was assigned in 5-s epochs. EMG amplitude and ratios of EEG frequency band amplitudes were measured, and compared with empirical thresholds in each animal.STEP 1: EMG amplitude greater than threshold? Yes: "active" wake, no: sleep or "quiet" wake. STEP 2: EEG amplitude ratio (delta x alpha)/(beta x gamma) greater than threshold? Yes: NREM, no: REM or "quiet" wake. STEP 3: EEG amplitude ratio Theta(2)/(delta x alpha) greater than threshold? Yes: REM, no: "quiet" wake. The algorithm was validated with one, two and three steps. The overall accuracy in discriminating wake and sleep (NREM and REM combined) using step one alone was found to be 90.1%. Overall accuracy using the first two steps was found to be 87.5% in scoring wake, NREM and REM sleep. When all three steps were used, overall accuracy in scoring wake, NREM and REM sleep was determined to be 87.9%. All accuracies were derived from comparisons with unequivocally-scored epochs from four 90-min recordings as defined by an experienced human rater. The algorithms were as reliable as the agreement between three human scorers (88%).     

Sleep electroencephalogram changes in acute hemispheric stroke

Background/objective: Since reports of the effects of cerebral hemispheric stroke on sleep architecture are rare and contradictory, we prospectively studied 24 patients with first acute supratentorial, extra-thalamic stroke.Methods: We assessed stroke severity, topography, and volume (on brain MRI). Sleep electroencephalogram recordings were performed a mean of 12 days after stroke onset, and scored for sleep stages over the healthy hemisphere. Sleep spindles and sawtooth waves were analyzed over both hemispheres. Data were compared with those of 17 age and gender-matched patients with normal brain imaging.Results: Compared to controls, stroke patients had lower total sleep time (P<0.01), lower sleep efficiency (P=0.02), and reduced amounts of NREM sleep stages 2-4 (P=0.02). Sleep spindles and sawtooth waves were often bilaterally reduced in patients with stroke volumes >25 ml. Abnormalities of REM sleep were more common in sleep studies performed within 3 days after stroke onset. Compared to patients with poor outcome, those with good outcome had higher sleep efficiency (P<0.01), more sleep time (P=0.02), and more NREM sleep stage 2 (P<0.01).Conclusion: Acute hemispheric stroke is accompanied by sleep EEG changes over the healthy hemisphere that correlate with stroke severity. These findings support the hypothesis that the cerebral hemispheres participate in the control of sleep.     

Nocturnal complex partial seizures precipitated by REM sleep. A case report

A 16-year-old patient presenting with complex partial seizures occurring in the transition from a REM period to wakefulness is described. His baseline EEG showed generalized and symmetrical slow spike and wave complexes, on a normal background activity, activated by NREM sleep. Polygraphic and videotape recordings, carried out for several nights, showed that after nearly each REM period, he would wake up briefly, presenting eye blinking followed by a burst of generalized hypersynchronous theta to start his seizures. These were characterized by moaning and autoaggressive behaviour, the ictal EEG showing generalized slow spike and wave complexes in the midst of several movement artifacts. At the end of each fit he fell back to REM sleep. Carbamazepine treatment completely resolved his symptoms, with full normalization of EEG activity.     

Localizing value of rapid eye movement sleep in temporal lobe epilepsy

Sleep state is a physiological modulator of epilepsy. Non rapid-eye-movement (NREM) sleep generally increases interictal epileptiform discharges (IEDs) and expands their field of distribution. In contrast, rapid eye movement (REM) sleep tends to suppress IEDs and may limit their spread outside of the region of primary seizure origin. The ability of REM sleep to restrict IEDs may have localizing value in temporal lobe epilepsy patients undergoing evaluations for epilepsy surgery. We present the case of a woman with medically-refractory seizures secondary to a mesial temporal glioma. Although scalp and intracranial electroencephalographic (EEG) seizure recordings supported bilateral epileptic foci, the IEDs recorded during REM sleep were restricted to the region of the glioma, and the patient had a successful surgical outcome. Our findings support the usefulness of combining sleep recordings with EEG monitoring in the evaluation of candidates for epilepsy surgery.     

Visual and spectral analysis of sleep EEG in acute hemispheric stroke

BACKGROUND: Reports on the effects of focal hemispheric damage on sleep EEG are rare and contradictory. PATIENTS AND METHODS: Twenty patients (mean age +/- SD 53 +/- 14 years) with a first acute hemispheric stroke and no sleep apnea were studied. Stroke severity [National Institute of Health Stroke Scale (NIHSS)], volume (diffusion-weighted brain MRI), and short-term outcome (Rankin score) were assessed. Within the first 8 days after stroke onset, 1-3 sleep EEG recordings per patient were performed. Sleep scoring and spectral analysis were based on the central derivation of the healthy hemisphere. Data were compared with those of 10 age-matched and gender-matched hospitalized controls with no brain damage and no sleep apnea. RESULTS: Stroke patients had higher amounts of wakefulness after sleep onset (112 +/- 53 min vs. 60 +/- 38 min, p < 0.05) and a lower sleep efficiency (76 +/- 10% vs. 86 +/- 8%, p < 0.05) than controls. Time spent in slow-wave sleep (SWS) and rapid eye movement (REM) sleep and total sleep time were lower in stroke patients, but differences were not significant. A positive correlation was found between the amount of SWS and stroke volume (r = 0.79). The slow-wave activity (SWA) ratio NREM sleep/wakefulness was lower in patients than in controls (p < 0.05), and correlated with NIHSS (r = -0.47). CONCLUSION: Acute hemispheric stroke is accompanied by alterations of sleep EEG over the healthy hemisphere that correlate with stroke volume and outcome. The increased SWA during wakefulness and SWS over the healthy hemisphere contralaterally to large strokes may reflect neuronal hypometabolism induced transhemispherically (diaschisis).     

Deep sleep after social stress: NREM sleep slow-wave activity is enhanced in both winners and losers of a conflict

Sleep is considered to be a recovery process of prior wakefulness. Not only duration of the waking period affects sleep architecture and sleep EEG, the quality of wakefulness is also highly important. Studies in rats have shown that social defeat stress, in which experimental animals are attacked and defeated by a dominant conspecific, is followed by an acute increase in NREM sleep EEG slow wave activity (SWA). However, it is not known whether this effect is specific for the stress of social defeat or a result of the conflict per se. In the present experiment, we examined how sleep is affected in both the winners and losers of a social conflict. Sleep–wake patterns and sleep EEG were recorded in male wild-type Groningen rats that were subjected to 1 h of social conflict in the middle of the light phase. All animals were confronted with a conspecific of similar aggression level and the conflict took place in a neutral arena where both individuals had an equal chance to either win or lose the conflict. NREM sleep SWA was significantly increased after the social conflict compared to baseline values and a gentle stimulation control condition. REM sleep was significantly suppressed in the first hours after the conflict. Winners and losers did not differ significantly in NREM sleep time, NREM sleep SWA and REM sleep time immediately after the conflict. Losers tended to have slightly more NREM sleep later in the recovery period. This study shows that in rats a social conflict with an unpredictable outcome has quantitatively and qualitatively largely similar acute effects on subsequent sleep in winners and losers.     

Ultradian rhythms of alternating cerebral hemispheric EEG dominance are coupled to rapid eye movement and non-rapid eye movement stage 4 sleep in humans

Objective: To replicate the left minus right (L−R) hemisphere EEG power shifts coupled to rapid eye movement (REM) and non-rapid eye movement (NREM) sleep observed in 1972 by Goldstein (Physiol Behav (1972) 811), and to characterize the L−R EEG power spectra for total EEG, delta, theta, alpha and beta bands.Background: Ultradian alternating cerebral hemispheric dominance rhythms are observed using EEG during both waking and sleep, and with waking cognition. The question of whether this cerebral rhythm is coupled to the REM–NREM sleep cycle and the basic rest–activity cycle (BRAC) deserves attention.Methods: L−R EEG signals for ten young, normal adult males were converted to powers and the means were normalized, smoothed and subtracted. Sleep hypnograms were compared with L−R EEGs, and spectra were computed for C3, C4 and L−R EEG powers.Results: Significant peaks were found for all C3, C4 and L−R frequency bands at the 280–300, 75–125, 55–70 and 25–50 min bins, with power dominating in the 75–125 min bin. L−R EEG rhythms were observed for all bands. Greater right hemisphere EEG dominance was found during NREM stage 4 sleep, and greater left during REM for total EEG, delta and alpha bands (Chi-squares, P<0.001). Theta was similar, but not significant (P=0.163), and beta was equivocal.Conclusions: Earlier ultradian studies show that lateral EEG and L−R EEG power have a common pacemaker, or a mutually entrained pacemaker with the autonomic, cardiovascular, neuroendocrine and fuel-regulatory hormone systems. These results for L−R EEG coupling to sleep stages and multi-variate relations may present a new perspective for Kleitman's BRAC and for diagnosing variants of pathopsychophysiological states.     

Modifications of sleep EEG induced by chronic vagus nerve stimulation in patients affected by refractory epilepsy.

OBJECTIVE: The aim of this study was to evaluate the impact of chronic vagus nerve stimulation (VNS) on sleep/wake background EEG and interictal epileptiform activity (IEA) of patients with medically refractory epilepsy. METHODS: From a broader sample of 10 patients subjected to baseline and treatment polysomnographies, spectral analysis and IEA count have been performed on 6 subjects' recordings, comparing the results by means of statistical analysis. RESULTS: An overall increase in EEG total power after VNS has been observed, more marked in NREM sleep; collapsing EEG power spectra into 5 frequency bands, we have found a statistically significant increase in delta and theta in NREM sleep, and of alpha in wakefulness and REM sleep. The incidence of IEA is diminished, although not significantly; only the duration of discharges is significantly diminished. CONCLUSIONS AND SIGNIFICANCE: Long-term VNS produces an enhancement in sleep EEG power of medically refractory epileptic patients. These results may be related to a better structured composition of EEG, and it is possible that chronic VNS may have a major role in enhancing the brain's ability to generate an electrical activity.     

Benign EEG variants in the sleep–wake cycle: A prospective observational study using the 10–20 system and additional electrodes

ObjectivesTo study the prevalence of benign EEG variants (BEVs) in the sleep–wake cycle among 1163 consecutive patients.MethodsProspective, observational EEG study using the 10–20 system with systematically two additional anterior-temporal electrodes. Depending on clinical indications, other electrodes were added. REM sleep identification was based on its characteristic EEG grapho-elements and rapid eye movements, clearly detectable with the additional anterior-temporal and fronto-polar electrodes due to eye proximity. The video-EEG monitoring duration was between 24 hours and eight days.ResultsWe identified 710 patients (61%) with BEVs. Positive occipital sharp transients of sleep (POSTs) were observed in 36.4% of participants, mu rhythm in 22.4%, lambda waves in 16.7%, wicket spikes (WS) in 15%, 14- and 6-Hz positive bursts in 8.3%, benign sporadic sleep spikes (BSSS) in 3.3%, rhythmic mid-temporal theta burst of drowsiness (RMTD) in 2.15%, midline theta rhythm in 2.1% and six-Hz spike and wave (SW) bursts in 0.1%. WS and RMTD were present during wakefulness, NREM (14.1%, 1.3%, respectively) and REM sleep (3.3%, 1.1%, respectively). Mu rhythm was also observed during NREM (1.5%) and REM sleep (7.7%). Fourteen- and 6-Hz positive bursts were present during NREM (4.5%) and REM sleep (6.5%). BSSS and six-Hz SW bursts were only observed during NREM sleep.ConclusionsThe prevalence of BEVs is much higher than current estimates. POSTs and WS can no longer be considered as unusual patterns but physiological patterns of NREM sleep. RMTD and mu rhythm may be observed during NREM and REM sleep.     

Evidence for 2-stage models of sleep and memory: Learning-dependent changes in spindles and theta in rats

What processes are involved in the formation of enduring memory traces? Sleep has been proposed to play a role in memory consolidation and the present study provides evidence to support 2-stage models of sleep and memory including both non-rapid eye movement (NREM) and rapid eye movement (REM) sleep. Previous research has shown REM sleep increases following avoidance learning and memory is impaired if REM deprivation occurs during these post-training periods indicating that REM sleep may have a role in memory consolidation processes. These discrete post-training periods have been termed REM sleep windows (RSWs). It is not known whether the electroencephalogram has unique characteristics during the RSW. Further investigation of the RSW was one of the primary goals of this study. We investigated the epidural-recorded electrophysiological learning-related changes following avoidance training in rats. Theta power increased in the learning group during the RSW, suggesting that theta is involved in memory consolidation during this period. Sleep spindles subsequently increased in slow wave sleep (SWS). The results suggest that both NREM and REM sleep are involved in sleep-dependent memory consolidation, and provide support for existing 2-stage models. Perhaps first theta increases to organize and consolidate material via hippocampal–neocortical dialogue, followed by subsequent refinement in the cortex by spindles during SWS.     

INFLUENCE OF AGE ON THE INTERRELATION BETWEEN EEG FREQUENCY BANDS DURING NREM AND REM SLEEP

The age-dependence of temporal interrelations between distinct frequency bands of sleep EEG was investigated in a group of 59 healthy young and middle-aged males via cross correlation analysis. Based on global evaluation throughout the entire night, a highly significant decline of the delta/theta correlation with increasing age was found. A separate analysis for non-rapid eye movement (NREM) and rapid eye movement (REM) sleep revealed different changes with aging. During NREM sleep, the correlation between the delta and theta frequency bands decreased with increasing age. In contrast, during REM sleep, a stronger correlation became obvious between the theta, alpha, and beta frequency bands with increasing age, whereas the lower frequency components were not affected. These findings indicate that aging processes seem to interact with sleep EEG rhythms in a complex manner, where most conspicuous is a disintegration of the activities in the lower frequency range, both concerning the successive sleep cycles across the night and the microstructure of NREM sleep     

Localized pontine lesion: nearly total absence of REM sleep

Eight whole-night polysomnographic recordings were conducted in a 33-year-old man with a localized pontine lesion inflicted by a shrapnel fragment. Sleep recordings revealed no rapid eye movement (REM) sleep in 3 nights, and markedly reduced REM sleep in 5 nights; non-rapid eye movement (NREM) sleep was normal. In spite of marked reduction of REM sleep, the patient conducted a normal life and had none of the typical symptoms of REM-sleep deprivation.     

Baseline and post-deprivation recovery sleep in SCN-lesioned rats

In humans, advancing age alters sleep patterns, reducing high voltage NREM sleep, sleep bout length, and delta power during NREM sleep. Although the mechanism by which these alterations occur is unknown, age-related changes in normal circadian processes may play a role. Increased age produces histological and functional changes in the suprachiasmatic nucleus (SCN), and alters the amplitude and phase of circadian rhythms. To examine the relationship between SCN function and age-related changes in sleep, we produced radiofrequency (RF) lesions of the SCN in rats of different ages and examined sleep behavior before and after sleep deprivation. Three-, 12- and 18-month-old rats received RF or sham lesions of the SCN. After verifying loss of circadian rhythm, 24-h EEG/EMG/temperature recordings were made in dim light before and after 24 h of sleep deprivation using the disk-over-water method. Age-related changes in NREM sleep, sleep bout length, and delta EEG power persisted despite SCN lesions. SCN lesions in all age groups increased baseline NREM sleep by 4% and NREM delta power by 15%, and decreased REM sleep by 10%. Although SCN lesions initially produced more REM and NREM sleep during recovery, 24-h values did not differ. Deteriorating SCN function is unlikely to cause the characteristic changes in sleep that occur with age. Our data also imply that an intact SCN slightly inhibits NREM sleep in the rat. Changes in NREM sleep and delta EEG power during recovery in lesioned rats suggest that the SCN may influence homeostatic regulation.     

Electroencephalographic correlates of cataplectic attacks in narcoleptic canines

Cataplectic attacks were monitored behaviorally and polygraphically in 4 narcoleptic dogs, of which three inherited the disorder. The recorded EEG signals were evaluated by power spectral analysis. We found 3 distinct stages of cataplexy: an initial stage which resembled wakefulness with tonic suppression of EMG activity, a later stage which was highly similar to REM sleep, and a final transitional stage to wakefulness or NREM sleep. The first stage of cataplexy was characterized by full postural collapse, a waking-like EEG spectrum, visual tracking, and a hypotonic EMG. The second stage of cataplexy differed electrographically from the previous stage by the onset of hypersynchronous hippocampal theta activity, a REM-like EEG spectrum, larger amplitude EEG signals, and a higher peak theta frequency. Glazed eyes, sporadic rapid eye movements and muscle twitches were also present. The final stage of cataplexy was characterized by mixed amplitude, mixed frequency EEG activity, and by the absence of rapid eye movements, visual tracking, directed movements, and muscle twitches. The EEG spectra of two other narcoleptic phenomena, sleep-onset REM periods and NREM sleep onsets from cataplexy, were nearly identical to the spectra of the normally occurring REM and NREM sleep periods.     

Estradiol treatment modulates spontaneous sleep and recovery after sleep deprivation in castrated male rats

Exogenous estradiol (E) is used occasionally to treat the side effects associated with androgen-deprivation in men, but its effects on sleep patterns have received little attention. We examined whether E modulates sleep patterns and recovery from sleep loss in castrated male rats. Adult male rats were castrated and implanted subcutaneously with Silastic tubes containing either oil (Cast + Oil) or E (Cast + E). Sham-operated male rats (Intact) were implanted with oil-filled tubes. All rats were also implanted with EEG and EMG electrodes for sleep/wake recordings. After two weeks, polysomnographic recordings were made before, during, and following 6 h of sleep deprivation (SD). At baseline, the Cast + Oil group showed sleep and EEG patterns similar to those in the Intact group. Compared to these groups, the Cast + E group spent more time awake during the dark (active) phase, and showed higher EEG theta power (a measure of cortical activation) during wake and rapid eye movement (REM) sleep in both the light and dark phases. Following SD, the Cast + E group showed a larger increase from baseline in REM sleep amount, compared to the Cast + Oil group. The Cast + Oil group showed prolonged rebound in non-REM sleep and EEG delta power, and reduced REM sleep rebound, compared to the other two groups. These results indicate that E treatment in castrated male rats promotes baseline wakefulness during the active phase, and facilitates recovery of REM sleep after acute sleep loss. The possible benefit of E treatment for improving sleep quality in androgen-deprived men remains to be investigated.     

Reduced fronto-cortical brain connectivity during NREM sleep in Asperger syndrome: An EEG spectral and phase coherence study

Objective

To investigate whether sleep macrostructure and EEG power spectral density and coherence during NREM sleep are different in Asperger syndrome (AS) compared to typically developing children and adolescents.

Methods

Standard all night EEG sleep parameters were obtained from 18 un-medicated subjects with AS and 14 controls (age range: 7.5–21.5 years) after one adaptation night. Spectral, and phase coherence measures were computed for multiple frequency bands during NREM sleep.

Results

Sleep latency and wake after sleep onset were increased in AS. Absolute power spectrum density (PSD) was significantly reduced in AS in the alpha, sigma, beta and gamma bands and in all 10 EEG derivations. Relative PSD showed a significant increase in delta and a decrease in the sigma band for frontal, and in beta for centro-temporal derivations. Intrahemispheric coherence measures were markedly lower in AS in the frontal areas, and the right hemisphere over all EEG channels. The most prominent reduction in intrahemispheric coherence was observed over the fronto-central areas in delta, theta, alpha and sigma EEG frequency bands.

Conclusion

EEG power spectra and coherence during NREM sleep, in particular in fronto-cortical derivations are different in AS compared to typically developing children and adolescents.

Significance

Quantitative analysis of the EEG during NREM sleep supports the hypothesis of frontal dysfunction in AS.     

The sleep EEG's microstructure in depression: alterations of the phase relations between EEG rhythms during REM and NREM sleep

OBJECTIVE: We investigated the microstructure of sleep electroencephalograms (EEGs) of 13 unmedicated depressive inpatients and 13 healthy controls matched in sex and age, hypothesizing that depressives depict an alteration of certain EEG oscillations across the night. METHODS: We digitized the sleep EEGs with a sampling rate of 100 Hz (bipolar derivation C(z)-P(z), 1440 single sweeps; 2048 data points each), calculated the time course of delta (1-3.5 Hz), theta (3.5-7.5 Hz), alpha (7.5-15 Hz), and beta (15-35 Hz) activity over the night, and determined the correlation coefficients of these different EEG rhythms separately for rapid eye movement (REM) and non-rapid eye movement (NREM) sleep. RESULTS: For both groups we detected a clear difference between REM and NREM sleep cycles at certain frequency bands. The most impressive changes occurred for the delta/beta and theta/beta correlations, which change their signs between NREM (negatively correlated) and REM (positively correlated) sleep cycles. Following an analysis of variance model with repeated measurement design, a statistically significant group effect (P=0.024) between depressives and controls was observable during NREM sleep for the delta/beta (P=0.010) and theta/beta (P=0.018) interactions. CONCLUSION: We detected alterations of certain sleep EEG oscillations during the NREM sleep cycle, where the delta/beta as well as the theta/beta activities were higher (negatively) compared to healthy controls. Together with previous investigations on the influence of antidepressants on the microstructure of sleep EEGs, this is another hint that the NREM sleep cycle plays a major role in depression.     

Effect of sleep deprivation on sleep and EEG power spectra in the rat

EEG power spectra of the rat were computed for consecutive 4-s epochs of the daily light period and matched with the scores of the vigilance states. Sleep was characterized by a progressive decline of low frequency spectral values (i.e. slow wave activity) in non-rapid eye movement (non-REM) sleep, and a progressive increase in the amount of REM sleep. During recovery from 24-h total sleep deprivation (TSD) the following changes were observed: an increase of slow wave activity in non REM sleep with a persisting declining trend; an enhancement of theta activity (7.25-10.0 Hz) both in REM sleep and waking; a decrease of non-REM sleep and an increase of REM sleep. In addition, a slow wave EEG pattern prevailed in the awake and behaving animal during the initial recovery period. In selective sleep deprivation paradigms, either REM sleep or slow wave activity in non-REM sleep was prevented during a 2-h period following upon 24-h TSD. During both procedures, non-REM sleep spectra in the lowest frequency band showed no increase. There was no evidence for a further enhancement of slow wave activity after its selective deprivation. The results indicate that: (1) slow wave activity in non-REM sleep and theta activity in REM sleep may reflect sleep intensity; and (2) REM sleep and active waking, the two states with dominant theta activity, may be functionally related.     

28例路易体痴呆患者的视频-多导睡眠图分析

目的研究路易体痴呆（DLB）患者的视频～多导睡眠图,探讨路易体痴呆患者的睡眠结构。方法对临床诊断DLB28例及正常对照28人进行研究。所有入组者均进行简易精神状态检查量表（MMSE）和视频～多导睡眠图（Video-PSG）监测。结果（1）睡眠结构分析：与对照组相比,DLB组总睡眠间期时间（SPT）减少,差异具有统计学意义（P〈0．05）;总睡眠时间（TST）减少、睡眠效率（SE）下降、总醒觉时间（TWT）、入睡后清醒时间（WASO）增多、1期睡眠时间（TS1）、2期睡眠时间（TS2）、NREM睡眠时间（TNREMS）和REM睡眠时间（TREMS）均明显减少,差异具有统计学意义（P〈0．01）;（2）睡眠呼吸事件分析：DLB组与对照组相比,各项指标差异均无统计学意义（P〉0．05）;（3）其他睡眠事件分析：与对照组相比,DLB组睡眠期周期性肢体运动次数（PLMS）、快动眼睡眠行为异常（RBD）明显增多,差异具有统计学意义（P〈0．01）。结论DLB患者存在睡眠结构紊乱,睡眠异常行为亦很常见。视频-多导睡眠图对于研究DLB患者的睡眠障碍很有帮助。     

Sleep and treatment response in depression: new findings using power spectral analysis

This study examined quantitative measures of sleep electroencephalogram (EEG) and phasic rapid eye movements (REM) as correlates of remission and recovery in depressed patients. To address correlates of remission, pre-treatment EEG sleep studies were examined in 130 women outpatients with major depressive disorder treated with interpersonal psychotherapy (IPT). To address correlates of recovery, baseline and post-treatment EEG sleep studies were examined in 23 women who recovered with IPT alone and 23 women who recovered with IPT+fluoxetine. Outcomes included EEG power spectra during non-rapid eye movement (NREM) sleep and REM sleep and quantitative REMs. IPT non-remitters had increased phasic REM compared with remitters, but no significant differences in EEG power spectra. IPT+fluoxetine recoverers, but not IPT recoverers, showed increases in phasic REM and REM percentage from baseline to recovery. In NREM sleep, the IPT+fluoxetine group showed a decrease in alpha power from baseline to recovery, while the IPT group showed a slight increase. The number of REMs was a more robust correlate of remission and recovery than modeled quantitative EEG spectra during NREM or REM sleep. Quantitative REMs may provide a more direct measure of brainstem function and dysfunction during REM sleep than quantitative sleep EEG measures.