Abstract:
Negative feelings, e.g., hopelessness, helplessness, sadness, etc., often result in a loss of purpose and meaning in life, which can get associated even with lethal outcomes such as suicides. The negative feelings are generally analyzed through interviews that become extremely difficult in the contexts of marginalized communities such as refugees. The difficulties arise due to diversified barriers covering language and cultural barriers, lack of literacy and technological skills, lack of trust to reveal sensitive information to a stranger, etc. To overcome the barriers, we propose using non-verbal biomarkers such as non-invasive electroencephalogram (EEG) brainwave signals and head movement data for the purpose of revealing and analyzing negative feelings. To do so, in this study, we collect EEG and head movement data along with conducting interviews on negative feelings over Rohingya refugees (n = 135). Then, we analyze associations among different negative feelings based on the collected interview data through applying graph theoretic approaches to develop various models of associations. Besides, we use various statistical measures to identify potential neurobiological markers. We also leverage machine-learning algorithms for classifying the negative feelings. Our study demonstrates novel outcomes on the associations over different negative feelings. Besides, our study also presents substantial (up to 95%) accuracy in classifying negative feelings based on EEG signals and head movement data in isolation and in combination.
