Investigation of EEG signals for emotional state analysis
This research investigates the electroencephalography (EEG) signal and how features within the signal correlates with the human emotional states. EEG signals were collected from 4 male and 4 female test subjects while exposed to audio-visual stimuli. The stimuli were selected to evoke 4 groups of em...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Published: |
2011
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This research investigates the electroencephalography (EEG) signal and how features within the signal correlates with the human emotional states. EEG signals were collected from 4 male and 4 female test subjects while exposed to audio-visual stimuli. The stimuli were selected to evoke 4 groups of emotions i.e., sad, fear, happiness and disgust. The signals were then processed to remove artifacts using an adaptive empirical mode decomposition (EMD). This novel method was designed to remove artifact for a limited number of EEG channel. Subsequently, the spectral features namely alpha peak frequency, alpha band power, beta peak frequency, beta band power and
alpha to beta band power ratio were extracted from the artifact-free EEG of each electrode
using modified Welch periodogram. An investigation was conducted on finding the optimum parameters for Welch in terms of the number of segments and the epoch length. Three different Welch segments
which are, 2, 4 and 8 segments were applied and analyzed on three types of EEG epoch lengths that are, 2, 5 and 10 seconds. The hypothesis derived from the first experiment was subsequently tested on an additional 7 male subjects. It was observed that the alpha peak frequency consistently had the highest magnitudes for happiness-evoked emotion for male subjects. This observation was not reflected for all other type of emotions. Hence signifies the correlation of alpha peak frequency and the human emotion of happiness. The results of the investigation also showed that the optimum
number of segments for Welch is 4 segments while the optimum epoch length of a
signal is 5 seconds. |
|---|