Problem Statement/Objectives


Recently, the correlation between the level of alpha waves and insomnia has been studied. Research papers such as “Alpha-waves frequency characteristics in health and insomnia during sleep” published in the Journal of Sleep Research in January 2016, and “Covert Waking Brain Activity Reveals Instantaneous Sleep Depth,” published by Harvard University in March 2011 suggest the level of alpha waves carries information about sleep stability. Also, quantitative research about “EEG spectral analysis of NREM sleep in a large sample of patients with insomnia and good sleepers: effects of age, sex and part of the night” published in September 2016 evaluates differences between patients with insomnia (N=803) and those who are not affected (N=811). Often the number of participants is limited to at most 15, but this specific research with quantitative data could obtain statistically significant correlation.


Several research papers have attempted to detect insomnia-related changes by means of the EEG, and some of the authors adopted power spectral analysis as a method to detect the signals (Freedman, 1986; Merica et al., 1998; Spiegelhalder et al., 2012). The increased alpha, beta, sigma and gamma power found during sleep in such patients are consistent with the theory of chronic, pathological hyper-arousal (Feige et al., 2013; Riemann et al., 2010, 2015).


The frequency of alpha waves measured by the EEG is a strikingly stable trait in healthy humans with a population standard deviation of only about 1 Hz (Grandy et al., 2013; Niedermeyer et al., 1997). Recently, a correlation between the power spectrum level of alpha waves and certain pathological processes has been revealed. Especially, larger deviations from the normal, physiological alpha range have been associated with clinical symptoms such as deteriorating working memory performance, schizophrenia, Alzheimer’s disease, and addiction. As quantitative and qualitative research papers about the correlation between the level of alpha waves and insomnia have been published recently, we will implement a system that can accurately self-diagnose insomnia given a person’s dataset. We are looking forward to expanding our research to implement similar models for diagnosing other pathological processes mentioned above after finishing our senior design project.


Aims or objectives of the project


Broad objectives or goals of the project

Our ultimate goal of the project is to develop a method to detect insomnia, given a dataset of a patient’s alpha waves obtained by an EEG apparatus. We tested digital signal processing (DSP) principles we learned in classes, including filtering principles and Windowing. After applying DSP techniques, we diagnose insomnia with the extracted alpha frequencies and their standard deviations from EEG dataset form PhysioNet.


Specific objectives or aims of the project

Our specific objective for the project is to increase the accuracy of our self-diagnosis program with additional information such as time taken for a participant before sleep-onset, once we finish building the model. If time allows, we will try to implement power spectral analysis to assure whether our method is capable of detecting insomnia-related changes in the EEG as the increased alpha, beta, sigma, and gamma power were previously found to be consistent with the theory of chronic, pathological hyper-arousal.