Actigraphy-based sleep estimation in adolescents and adults: a comparison with polysomnography using two scoring...

Area of Science:

• Biomedical Engineering
• Sleep Medicine
• Wearable Technology

Background:

• Actigraphy is a common tool for estimating sleep-wake patterns.
• Limited data exists on the comparability of different actigraphy devices and algorithms.
• Accurate sleep-wake estimation is crucial for clinical and research applications.

Purpose of the Study:

• To compare sleep-wake time estimates from two wrist actigraphs (GT3X+ and Actiwatch Spectrum) against in-home polysomnography (PSG).
• To evaluate the influence of two scoring algorithms (Sadeh and Cole-Kripke) on GT3X+ actigraphy data.
• To assess the agreement between devices and algorithms in healthy volunteers.

Main Methods:

• 35 healthy participants (children and adults) were enrolled.
• Simultaneous recording of GT3X+ and Actiwatch Spectrum for at least five days in 22 adults.
• Concurrent actigraphy and PSG measurements over one night in 25 participants (12 adults, 13 children).
• Statistical analysis included intraclass correlation coefficients (ICCs), epoch-by-epoch comparisons, paired t-tests, and Bland-Altman plots.

Main Results:

• Both actigraphs demonstrated comparable accuracy (ICC: 0.81-0.86) to PSG for overall sleep-wake estimation.
• The Cole-Kripke algorithm exhibited higher sensitivity for sleep detection but lower specificity for wake detection compared to the Sadeh algorithm on the GT3X+.
• Agreement for wake after sleep onset between actigraphy and PSG was poor (ICC: 0.00-0.10).
• The GT3X+ with the Cole-Kripke algorithm showed comparable total sleep time and wake after sleep onset estimates to the Actiwatch Spectrum in adults.

Conclusions:

• Wrist actigraphy devices provide comparable and reasonably accurate sleep-wake estimates versus PSG.
• Actigraphy algorithms significantly influence sleep-wake estimation, particularly for wake episodes.
• While both devices showed limitations in identifying wakefulness, the GT3X+ using the Cole-Kripke algorithm yielded results similar to the Actiwatch Spectrum.