You are here

Analysis of working memory from EEG signals under different emotional states

Analysis of working memory from EEG signals under different emotional states

Authors: 
Buket D. Barkana, Yusuf Ozkan, Joanna A. Badara
Year: 
2022
Journal: 
Biomedical Signal Processing and Control
Abstract: 

This study analyzes electroencephalography (EEG) measurements during short-term memory retention under different emotional states. A public-domain library with emotion-annotated images (IAPS) was used to stimulate neutral, negative, and positive emotions. The associated EEG data were acquired from twelve volunteers (between 20 and 26 years old; ten males and two females). Each participant was exposed to three sessions back-to-back on the same day. Each session corresponded to the induced emotional states (positive, negative and neutral) and consisted of relaxation, memorization of a list of ten words and ten numbers, watching a set of images to arouse emotion, and recalling the words and numbers memorized earlier. Statistical and spectral features of EEG data were analyzed for two instances: emotion recognition (neutral, negative, and positive) and recall events under the three emotional states. By designing two baseline machine-learning models, support vector machines (SVMs) and K-nearest neighbor (KNN), the significance of the EEG bands and the brain lobes were studied. Experimental results suggest that the short-term (working) memory recalls after exposure to neutral, negative, and positive images (to arouse neutral, negative, and positive emotions) differ from each other significantly (at alpha level 0.001). We have found that each EEG band carries unique information in both emotion and memory recall classification tasks and recommend that the entire EEG signal frequency range must be analyzed in future similar studies. On the other hand, we also have found that each brain region carries similar information as it relates to each task (i.e., memorization, recall), thus only one of the brain regions can be analyzed in future studies in order to avoid complexity and high computation time.

Read more >>>

© Copyright 1999 - 2022 ANT Neuro | www.ant-neuro.com | Terms of Use | Privacy Statement | Contact | USA Customers