TY - GEN
T1 - Detecting PTSD Using Neural and Physiological Signals
T2 - 11th International Conference on Affective Computing and Intelligent Interaction, ACII 2023
AU - Kalanadhabhatta, Manasa
AU - Roy, Shaily
AU - Grant, Trevor
AU - Salekin, Asif
AU - Rahman, Tauhidur
AU - Bergen-Cico, Dessa
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - Post-traumatic stress disorder (PTSD) is a serious condition that is characterized by negative mood and affect, hyperarousal, irritability, and reactivity, as well as deterioration of cognitive processes such as attention and memory. Timely identification and treatment of PTSD symptoms can significantly improve symptom management and recovery. However, accurate prediction of PTSD outside clinical settings is often challenging. In this work, we investigate whether deficits in cognitive performance can be used to classify individuals with and without PTSD. We further examine whether neural and physiological signals such as prefrontal cortex activity, heart rate, respiration, and electrodermal activity recorded in conjunction with cognitive task performance can be leveraged to improve PTSD classification. Our results indicate that working memory tasks can achieve an F1 score of 0.80 at classifying individuals with PTSD, which can be further improved to 0.91 by combining multimodal information from neurophysiological signals. Based on our findings, we provide recommendations for in-the-wild PTSD classification.
AB - Post-traumatic stress disorder (PTSD) is a serious condition that is characterized by negative mood and affect, hyperarousal, irritability, and reactivity, as well as deterioration of cognitive processes such as attention and memory. Timely identification and treatment of PTSD symptoms can significantly improve symptom management and recovery. However, accurate prediction of PTSD outside clinical settings is often challenging. In this work, we investigate whether deficits in cognitive performance can be used to classify individuals with and without PTSD. We further examine whether neural and physiological signals such as prefrontal cortex activity, heart rate, respiration, and electrodermal activity recorded in conjunction with cognitive task performance can be leveraged to improve PTSD classification. Our results indicate that working memory tasks can achieve an F1 score of 0.80 at classifying individuals with PTSD, which can be further improved to 0.91 by combining multimodal information from neurophysiological signals. Based on our findings, we provide recommendations for in-the-wild PTSD classification.
KW - PTSD
KW - cognitive performance
KW - neural activity
KW - physiological signals
KW - post-traumatic stress disorder
KW - wearables
UR - http://www.scopus.com/inward/record.url?scp=85184659371&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85184659371&partnerID=8YFLogxK
U2 - 10.1109/ACII59096.2023.10388200
DO - 10.1109/ACII59096.2023.10388200
M3 - Conference contribution
AN - SCOPUS:85184659371
T3 - 2023 11th International Conference on Affective Computing and Intelligent Interaction, ACII 2023
BT - 2023 11th International Conference on Affective Computing and Intelligent Interaction, ACII 2023
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 10 September 2023 through 13 September 2023
ER -