# Evaluating Electroencephalogram-Based Predictive Model for Drowsiness Measurement to Reduce Accident Risk in Active Individuals: Protocol for a Preliminary Monocentric Study

## TL;DR

This study aims to develop and validate an EEG-based system to predict drowsiness in active individuals, helping reduce accident risks in high-stakes environments.

## Contribution

The study introduces a novel protocol for validating EEG-based predictive models of drowsiness using real-world sleep deprivation scenarios.

## Key findings

- The study will assess the effectiveness of the Objective Sleepiness Scale in measuring wakefulness.
- Multimodal EEG markers and cognitive performance will be evaluated under sleep deprivation conditions.
- Sociodemographic and clinical variables will be analyzed for their influence on drowsiness prediction.

## Abstract

Voluntary behaviors and socioeconomic factors, such as social jetlag and shift work, can lead to insufficient or disrupted sleep, resulting in drowsiness among active individuals. In occupational and driving contexts, drowsiness poses a serious safety risk by impairing alertness, slowing reaction times, and increasing the likelihood of accidents. Developing automatic and easy-to-implement tools for drowsiness detection or prediction is essential in the management of sleepy patients or in high-risk environments where sustained vigilance is critical.

This study aims to validate continuous or predictive methods for assessing drowsiness using automated analysis of a limited number of electroencephalogram (EEG) channels.

Designed as a single-center, nonrandomized, single-group study, this investigation will evaluate drowsiness and cognitive performance in 40 healthy volunteers exposed to 2 sleep deprivation conditions simulating real-world occupational scenarios. The primary outcome will be the Objective Sleepiness Scale (OSS) and its automated analysis, with a focus on its ability to measure objective wakefulness as assessed by the maintenance of wakefulness test (MWT). Secondary outcomes will include multimodal resting-state EEG markers, subjective and objective sleepiness measures, performance on a simulated driving task, attention, executive function, and vigilance assessments, as well as sleep quality, sleep quantity, and mind-wandering. The influence of sociodemographic and clinical variables on the measurement and prediction of drowsiness will also be systematically examined.

This study received funding from Physip and ANR (Agence Nationale de la Recherche, National Research Agency) in 2019, with ethical committee (Comité de Protection des Personnes, Committee for the Protection of Persons) approval in May 2022. Recruitment began in March 2023 and was completed in May 2025, with a database lock in June 2025. Data analysis started in June 2025 and is still ongoing.

By validating these novel EEG-based measures, this study aims to lay the groundwork for proactive strategies for drowsiness management in occupational, transportation, and clinical settings.