High-precision eye-tracking in nonvisual settings

ongoing

In the rapidly evolving landscape of cognitive neuroscience, a significant surge in research has highlighted eye movements as a critical yet previously underestimated window into neuro-cognitive processes. Our project is at the forefront of this paradigm shift, proposing an innovative machine learning-based approach to decode eye movements with high precision using electrooculography (EOG) channels alone. This method is set to revolutionize eye-tracking technologies, particularly in non-visual contexts such as closed eyes or during sleep, where traditional methods are ineffective. The initial phase of our research involves the extensive collection of simultaneous eye-tracking and EOG data under various conditions, including simulated sleep patterns. Using state-of-the-art deep neural networks to map high-precision eye-tracking data onto the simultaneously collected EOG signal, we aim to achieve a level of EOG electrode precision that has previously been unattainable. A key focus of our deep-learning model is its ability to generalize, thereby minimizing the need for extensive individual calibration. In practical terms, our project holds transformative potential across multiple domains. In the scientific field, it enables the exploration of eye movements with closed eyes, which is necessary for mind-reading research and mental states like meditation, providing new insights into the human mind. In medical monitoring, it opens avenues for non-invasive sleep analysis and the diagnosis of sleep-related disorders, previously hindered by the low precision of current gaze-tracking technology. The most groundbreaking application lies in developing a wearable device that enables widespread, effortless tracking and analysis of eye movements during sleep, thus democratizing the use of advanced cognitive monitoring tools to offer insights into sleep patterns and build a healthy sleep routine.