Thalamic stereo-EEG detects sleep stages N2 and N3 with discrimination approaching scalp EEG

Thalamic stereotactic EEG (stereo-EEG) holds promise for accurately identifying sleep stages N2 and N3, often challenging to differentiate using traditional scalp EEG. Research involving 477 hours of stereo-EEG data collected from six adults, coupled with simultaneous scalp EEG monitoring, revealed that thalamic bipolar channels outperformed other intracranial sites in detecting these critical sleep stages.

Utilizing two classifiers, a random forest with hidden Markov decoding (RF + HMM) and a pretrained algorithm called SleepSEEG, the study evaluated the performance of thalamic channels.

The results demonstrated that thalamic channels significantly improved the discrimination of N2 and N3 stages from other sleep stages when classified by a sleep expert. Among the various regions sampled, thalamic channels achieved the highest area under the curve (AUC) scores, reaching up to 0.99 with the RF + HMM classifier. Even thalamic electrodes individually demonstrated impressive capabilities, with AUCs ranging from 0.79 to 0.99, with 10 out of 11 electrodes achieving at least 0.89 AUC with the RF + HMM classifier.

These findings, while derived from offline analysis, highlight the potential of thalamic steros-EEG in identifying sleep stages N2 and N3. This has significant implications for sleep-targeted neuromodulation, suggesting that real-time evaluation of thalamic channels may prove valuable in future research and clinical applications.