Estimation of Single Trial ERPs and EEG Phase Synchronization with Application to Mental Fatigue
- When?
- Thursday 21 July 2011, 10:30 to 11:30
- Where?
- 39BB02
- Open to:
- Staff, Students
- Speaker:
- Miss Delaram Jarchi
Monitoring mental fatigue is a crucial and important step for prevention of fatal accidents. This may be achieved by understanding and analysis of brain electrical potentials. Electroencephalography (EEG) is the record of electrical activity of the brain and gives the possibility of studying brain functionality with a high temporal resolution. EEG has been used as an important tool by researchers for detection of fatigue state. However, their proposed methods have been limited to classical statistical solutions and the results given by different researchers are somehow conflicting. Therefore, there is a need for modification of the existing methods for reliable analysis of mental fatigue and detection of fatigue state.
In addition to the raw EEG, event related potentials (ERPs), which are direct measures of brain responses to the specified stimuli, have been used in mental fatigue analysis since the attention related ERPs have shown to be effective for detection of fatigue state.
In this study we aim to extend and further develop the existing signal processing methods for EEG- and ERP-based mental fatigue analysis.First, a new approach is proposed for measuring synchronization of EEG oscillations in different frequency bands across brain regions. The approach is used to find the relevant and effective features for detection of the fatigue state. It uses adaptive methods such as empirical mode decomposition (EMD) and adaptive line enhancer (ALE) for extracting and de-noising the EEG oscillations. Then, Hilbert transform (HT) is used for computing the linear and nonlinear synchronization measures. A new method based on particle filtering (PF) is proposed for direct estimation of instantaneous phase of an oscillation. This method can be developed more in future studies for phase synchronization analysis of the EEG oscillations before and during the fatigue state.
ERP subcomponents are estimated using PF. Based on the proposed method, ERP subcomponents are separated in temporal domain across different trials and their inter-trial variability is tracked using a coupled PF. The method is applied to mental fatigue data to show the potential use of the method in ERP subcomponent estimation for detection of fatigue state. Then, a new spatiotemporal filtering is designed for estimation of the correlated ERP subcomponents. The method is robust against both temporal and spatial correlations of the ERP subcomponents. It is compared to the existing methods in different scenarios and its superiority is confirmed by using the simulated signals. It is also applied to real data to show its potential use in ERP subcomponent estimation.
Finally, an auditory based paradigm is implemented to evaluate the effectiveness of the designed mental fatigue detection system. By applying the proposed methods for estimation of single trial P300 subcomponents and EEG phase synchronization, it is demonstrated that the proposed auditory paradigm can be effectively used in a mental fatigue detection system.
