Modulating slow wave sleep at home with a mobile EEG system

Since ancient times sleep has fascinated philosophers and researchers alike. But it was not until the beginning of the 20th century that, with the introduction of electroencephalography (EEG), researchers and scientists were able to study sleep through the non-invasive measurements of brain activity. The introduction of EEG has enabled a better understanding of sleep and its relation to health and diseases. Considering the importance of sleep for health and well-being several methods to improve sleep, and more specific slow wave sleep (SWS), through external stimulation have been studied. Auditory stimulation delivered in-phase with slow waves during non-rapid eye movement sleep has shown to be an effective technique to enhance SWS with improvements in memory consolidation and cognition. So far, research using this approach was limited to complex and costly laboratory set-ups that required controlled human interaction. For this reason, lab-based sleep studies have so far included a low number of participants, and have been restricted to very few recording nights, therefore limiting their impact. Thus, the study of sleep in a more natural unobtrusive environment is desired. With the proliferation of mobile EEG technologies, brain activity recordings outside clinical or laboratory settings have become possible allowing for realistic long-term monitoring at larger scale and lower costs. However, available consumer devices for sleep monitoring and interventions, such as auditory stimulation, do not provide enough flexibility and data transparency to be used for research. To address these challenges, we developed a configurable and mobile 8-channel EEG system that enables sleep biosignal recording and autonomous auditory SWS stimulation. The system provides data transparency of the recorded biosignal and implemented algorithms' output for post-processing analysis and flexible electrode selection and position. In addition, it features a general-purpose connection that enables connectivity to other sensors. SWS enhancement has been observed only when the acoustic tones were played during the SW up-phase as compared to nights when the tones were delivered during SW down-phase or no stimulation was presented. Therefore, we propose two novel methodologies to improve the accuracy of state-of-the-art real-time slow wave phase detection, especially in the presence of low-amplitude slow waves. This represents an important contribution, especially for populations whose sleep is hallmarked by low amplitude slow waves. Therefore, elderly or Parkinson patients could benefit most from such non-invasive interventions. Additionally, we propose a novel method to automatically select a filter design for EEG signal preprocessing based on minimum filter-based distortions in the temporal, frequency, or both domains. The results of this thesis contribute to the transition of sleep research to home-based settings and constitute a step toward more robust and accurate in-phase auditory stimulation with resource-constrained, battery-powered devices while maintaining reliable performance