TY - GEN
T1 - Prediction of Cannabis Addictive Patients with Graph Neural Networks
AU - Wen, Shulin
AU - Yang, Shihao
AU - Ju, Xinglong
AU - Liao, Ting
AU - Liu, Feng
N1 - Publisher Copyright:
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2023
Y1 - 2023
N2 - Neurological research is closely intertwined with public health issues, and artificial intelligence (AI) holds substantial potential in this domain. This study aims to investigate the enhancement of brain imaging classification performance in diverse populations using Graph Neural Networks (GNN) and its variants. Brain activity data are sourced from public neuroimaging databases, including functional Magnetic Resonance Imaging (fMRI) data of cannabis addicts and a healthy control group. Our results show that, compared to the healthy control group, cannabis addicts exhibit significant alterations in functional connectivity in certain brain regions. With the application of AI tools, we can distinguish the two groups based on brain imaging. We observed a significant improvement in brain imaging classification performance, and this model has achieved an accuracy rate of approximately 80%. These AI tools’ robust generalizability and vast developmental potential were also highlighted. These findings not only provided a novel perspective on the role of AI in brain imaging studies but also suggested potential new strategies for addressing public health issues.
AB - Neurological research is closely intertwined with public health issues, and artificial intelligence (AI) holds substantial potential in this domain. This study aims to investigate the enhancement of brain imaging classification performance in diverse populations using Graph Neural Networks (GNN) and its variants. Brain activity data are sourced from public neuroimaging databases, including functional Magnetic Resonance Imaging (fMRI) data of cannabis addicts and a healthy control group. Our results show that, compared to the healthy control group, cannabis addicts exhibit significant alterations in functional connectivity in certain brain regions. With the application of AI tools, we can distinguish the two groups based on brain imaging. We observed a significant improvement in brain imaging classification performance, and this model has achieved an accuracy rate of approximately 80%. These AI tools’ robust generalizability and vast developmental potential were also highlighted. These findings not only provided a novel perspective on the role of AI in brain imaging studies but also suggested potential new strategies for addressing public health issues.
KW - Deep Learning
KW - Graph Neural Network
KW - functional Magnetic Resonance Imaging (fMRI)
UR - http://www.scopus.com/inward/record.url?scp=85172422027&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85172422027&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-43075-6_26
DO - 10.1007/978-3-031-43075-6_26
M3 - Conference contribution
AN - SCOPUS:85172422027
SN - 9783031430749
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 297
EP - 307
BT - Brain Informatics - 16th International Conference, BI 2023, Proceedings
A2 - Liu, Feng
A2 - Wang, Hongjun
A2 - Zhang, Yu
A2 - Kuai, Hongzhi
A2 - Stephen, Emily P.
T2 - 16th International Conference on Brain Informatics, BI 2023
Y2 - 1 August 2023 through 3 August 2023
ER -