P003 Can the electroencephalogram (EEG) alpha power predict mean sleep latency?

Abstract Introduction A Multiple Sleep Latency Test (MSLT) is resource intensive. An alternative, simpler method would be valuable. This study evaluates whether EEG total alpha power (TAP) from the 20-second eyes-closed calibration exercise conducted at the start of nap 1 of the MSLT can predict mean sleep latency (MSL). Method We retrospectively analyzed 400 consecutive polysomnography (PSG) and Multiple Sleep Latency Test (MSLT) reports and raw data from studies conducted at an Australian tertiary hospital adult sleep laboratory from January 2022 to February 2025. We excluded 156 reports due to an Apnea-Hypopnea Index (AHI) greater than 15, a Total Sleep Time of less than 6 hours, and Sleep Efficiency of less than 75%. Additionally, 37 MSLT raw data were excluded due to poor EEG quality. We collected data on the Arousal Index and medications for further analysis. We compared TAP values from 4 EEG signals (O1, O2, F3, F4) during the 20-second eyes-closed maneuver in nap 1 with the MSL using a one-way ANOVA. We averaged the TAP values from the same sites (occipital and frontal) and grouped them into three categories: low power (less than 20 μV2), medium power (20 to 40 μV2), and high power (greater than 40 μV2) before comparing them with MSL using a one-way ANOVA and post hoc analysis. Results We found a correlation between MSL and F3-TAP (p=.015). There was no significant difference in the average MSL among the three groups for F-TAP (p=.26) and O-TAP (p=.62), but we saw an increasing mean average MSL in the three groups of F-TAP and O-TAP. Low F-TAP: 8.35 (SD 4.62), medium F-TAP: 8.62 (SD 5.09) and high F-TAP: 9.76 (SD 5.13). Low O-TAP: 8.21 (SD 4.64), medium O-TAP: 8.94 (SD 5.03) and high O-TAP: 8.99 (SD 5.00). Conclusion We observed some evidence that higher alpha power from the frontal lobe may be associated with increased sleep latency, though more data and further examination of the impact of various medications and arousal index is required.