Sleep Stage Analysis using EEG Data

This hypnogram displays a typical, healthy sleep cycle. The subject progresses through sleep stages in regular cycles, with deep sleep (S3) appearing prominently in the early part of the night and REM sleep becoming more frequent in the later hours. Brief awakenings are normal and limited. MCAP events (in red) are evenly distributed and sparse, suggesting stable and undisturbed sleep architecture.

This subject shows clear signs of disrupted sleep. Long periods of wakefulness are seen, especially at the beginning and middle of the night. Transitions between stages are less frequent, and REM sleep is notably absent or very limited. MCAP events are more clustered, hinting at micro-arousals or instability in sleep continuity—hallmarks of insomnia.

The hypnogram indicates frequent and abrupt transitions between wake, REM, and light sleep stages. REM sleep appears unusually early and recurs throughout the night, which is a classic marker of narcolepsy. Sleep cycles are fragmented, with reduced time spent in deep sleep (S3). High MCAP density aligns with disturbed and unstable sleep.

This shows that a normal individual exhibits the highest normalized Delta power (around 4-5 units) during wakefulness (W), with moderate Theta power (around 2 units) across various stages, and negligible Alpha and Beta activity. This pattern reflects typical brain activity, with dominant Delta waves during wakefulness and a smooth transition to lighter sleep stages, consistent with healthy sleep architecture.

This shows that individuals with insomnia exhibit elevated Theta power during the S1 sleep stage (around 3-4 units), with noticeable Delta and Alpha activity during wakefulness (W) and lighter sleep stages (S2, S3). Beta power remains low across all stages. This suggests difficulty transitioning to deeper sleep, with persistent wake-like brain activity (higher Alpha and Theta) during intended sleep periods, consistent with insomnia patterns where deep sleep (Delta dominance) is reduced.

This shows that individuals with narcolepsy exhibit a notable Delta power peak (around 4 units) during wakefulness (W), with minimal activity in other bands. Theta and Alpha bands show moderate power during S2 and R stages, while Beta power remains negligible across all stages. This pattern suggests abnormal Delta wave presence during wakefulness, a hallmark of narcolepsy, reflecting disrupted sleep-wake transitions and potential rapid entry into sleep or REM-like states.

Normal sleepers show a well-balanced distribution across all stages. The majority of their sleep is spent in Stage 2 (S2) at ~43%, which is typical. Deep sleep (S3) is also well-represented at ~29%, indicating healthy restorative sleep. REM accounts for ~14%, supporting cognitive processing and dreaming, while wakefulness (W) is minimal at ~7.7%, reflecting efficient sleep cycles.
Insomniacs exhibit disrupted sleep patterns with a 27.5% spent awake (W), nearly four times higher than normals. Deep sleep (S3) is severely reduced to 15.3%, and REM also drops to 11.6%, both signs of poor sleep quality. Although S2 still dominates (~39.7%), the loss in restorative and REM stages highlights the impact of insomnia on sleep architecture.
Narcoleptic subjects spend a significant amount of time in REM sleep (17.4%), higher than both normals and insomniacs, consistent with rapid REM onset often seen in narcolepsy. They maintain decent S3 (23%) and high S2 (47%) proportions, suggesting relatively preserved non-REM structure. Wake time is low (~7.5%), but the altered REM balance points to their unique sleep profile.

The power spectral density (F4-C4) plot for Subject n1, a normal person, shows the strongest brain activity (around 0 dB/µV²/Hz) at low frequencies below 5 Hz, which is the Delta range. This gradually drops to about -30 dB/µV²/Hz by 50 Hz as it moves through Theta, Alpha, and Beta bands. This pattern reflects typical brain waves, with the most activity at lower frequencies during a restful state, matching a healthy EEG.

The power spectral density (F4-C4) plot for Subject ins2, an insomniac shows a peak power around 0 dB/µV²/Hz at low frequencies below 5 Hz, likely in the Delta range, followed by a decline to about -30 dB/µV²/Hz by 50 Hz. A noticeable rise in power around 10-15 Hz (Alpha range) suggests increased activity, which, combined with the "ins2" label, may indicate insomnia-like patterns where wakeful brain activity persists. This pattern is atypical for a restful state, as it shows more activity in the Alpha range than expected during sleep, indicating disrupted sleep architecture.

The power spectral density (F4-C4) plot for Subject narco1, a narcoleptic, shows a peak power around 0 dB/µV²/Hz at low frequencies below 5 Hz (Delta range), with sustained activity into higher frequencies, declining to -30 dB/µV²/Hz by 50 Hz. The persistent low-frequency power during wakefulness suggests abnormal Delta wave intrusion, a key indicator of narcolepsy, where sleep-like brain patterns disrupt normal wakeful states

The Time Domain Analysis (TDA) and Channel Activity Heatmaps provide a detailed visualization of EEG activity across different brain regions and time intervals for each subject type:

• Normal Subject: The heatmap shows balanced and consistent activity across all channels, with dominant Delta waves during deep sleep stages. This reflects stable and restorative sleep patterns typical of healthy individuals.
• Insomniac Subject: The heatmap reveals fragmented and uneven activity, with reduced Delta power and increased Alpha or Beta activity in certain regions. This indicates disrupted sleep architecture and difficulty maintaining deep sleep.
• Narcoleptic Subject: The heatmap highlights irregular activity, with Delta waves appearing inappropriately during wakefulness or light sleep. This pattern is characteristic of narcolepsy, where sleep-wake transitions are abnormal and unstable.

These visualizations help identify key differences in sleep patterns and brain activity across normal, insomniac, and narcoleptic subjects, providing insights into their unique sleep characteristics.