Sleepiness detection on a single channel ear-eeg

Loading...
Thumbnail Image
Date
2019-06
Authors
Choong, Chia How
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Sleep deprivation, also known as not enough sleep is one of the common issues creating a lot of consequences for mental and physical health. Through sleepiness monitoring, the need for sleep can be detected. Sleepiness can be monitored using EEG monitoring. However, current clinical practice in sleepiness EEG monitoring is Scalp-electrode monitoring. This technique needs to place large number of wet electrodes tightly on the scalp with wire attached on the electrodes, causing the movement of patient to be restricted. This project was focused on sleepiness detection using a wireless single channel Ear-EEG, objectives of this project were to design Ear-EEG sensor for signal acquisition, instrumentation circuit for signal conditioning, wireless system for wireless signal transmission and apply the designed system for detecting sleepiness. Ear-EEG sensor with several signal channels was built in this project, and channel with better performance was selected as a single channel Ear-EEG sensor. The electrode technique used in this project was dry contact electrode technique, this technique eliminates the unstable connection due to time of use, and uncomfortable wearing experience. Besides, the instrumentation circuit which included pre-amplifier, difference amplifier, notch filter, low pass filter, and post-amplifier were designed. This series of instrumentation circuit was successfully filtered and amplified EEG signal to reach amplitude of 140mV, creating 28 quantization levels for ADC in Arduino Nano. Arduino Nano and nRF24L01 are used to design wireless transmitter to transmit the EEG signals. The receiver side of wireless EEG system was designed using Arduino Uno, this wireless system had 18m of effective transmission distance. Arduino Uno transferred the EEG signal to LabVIEW software for signal processing such as bandpass filter and FFT. A visual comparison and statistical comparison were done for the Ear-EEG and Scalp-EEG, it gave maximum error of 1Hz for the highest peak detected in the power spectrum. A sleepiness data model was recorded from a volunteer, the result was discussed, and the sleep spindle was detected at 12 minutes. A sleepiness detection system was developed in LabVIEW to detect the brain state, the detection gives 83.33% of accuracy with one subject of test. In conclusion, sleepiness detection on a single channel Ear-EEG was successful.
Description
Keywords
Citation