The Development of an Algorithm to Detect Sleep Structure With a Wearable EEG Monitor in an Elderly Population

Not Applicable

Completed

• Conditions: Sleep

• Registration Number: NCT04755504
• Lead Sponsor: Universitaire Ziekenhuizen KU Leuven

Brief Summary

To evaluate whether it is able to perform sleep staging with EEG data recorded from 2 electrodes behind each ear.

Detailed Description

The Sensor Dot wearable device measures electroencephalography (EEG). It records from 2 electrodes behind each ear. The device was designed as a wearable for seizure detection in epilepsy patients. The purpose of this study is to test its ability to capture the information necessary for sleep monitoring in elderly patients. Trained electrophysiologists are unable to stage sleep on data from novel wearable devices, since AASM sleep scoring rules are only defined for standardized recording positions on the head. Therefore, we need an automated algorithm to perform sleep staging with data from the Sensor Dot device. We will train this algorithm using manual annotations made with the polysomnography simultaneously acquired with the wearable EEG.

Study Timeline

• Study Start: 2021-01-21
• Study First Submitted: 2021-02-08
• Study First Submitted QC: 2021-02-15
• Study First Posted: 2021-02-16
• Primary Completion: 2023-01-30
• Study Completion: 2023-01-30
• Status Verified: 2024-07
• Last Update Submitted: 2024-07-02
• Last Update Posted: 2024-07-03

Recruitment & Eligibility

• Status: COMPLETED
• Sex: All
• Target Recruitment: 100

Inclusion Criteria

• Subjects planned to undergo a diagnostic polysomnography
• > 60y old

Exclusion Criteria

• Patients unable to provide informed consent

Study & Design

• Study Type: INTERVENTIONAL
• Study Design: SINGLE_GROUP

Primary Outcome Measures

Name	Time	Method
Sleep algorithm	1 night	To develop an algorithm to characterize sleep architecture based on EEG measurement by 2 electrodes behind each ear.; To classify the sleep stages, a deep learning algorithm will be used. The algorithm will learn a complex function, transforming an input to an output, based on several examples. In this specific case, the input are 30s EEG epochs and the output are sleep stages. To classify the measured signal in the correct sleep stage, the deep learning algorithm will learn to extract useful features from the data.

Trial Locations

Locations (1)

UZ Leuven; 🇧🇪 Leuven, Belgium