Prediction of Cannabis Addictive Patients with Graph Neural Networks

Shulin Wen, Shihao Yang, Xinglong Ju, Ting Liao, Feng Liu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

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.

Original languageEnglish
Title of host publicationBrain Informatics - 16th International Conference, BI 2023, Proceedings
EditorsFeng Liu, Hongjun Wang, Yu Zhang, Hongzhi Kuai, Emily P. Stephen
Pages297-307
Number of pages11
DOIs
StatePublished - 2023
Event16th International Conference on Brain Informatics, BI 2023 - Hoboken, United States
Duration: 1 Aug 20233 Aug 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13974 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Conference on Brain Informatics, BI 2023
Country/TerritoryUnited States
CityHoboken
Period1/08/233/08/23

Keywords

  • Deep Learning
  • Graph Neural Network
  • functional Magnetic Resonance Imaging (fMRI)

Fingerprint

Dive into the research topics of 'Prediction of Cannabis Addictive Patients with Graph Neural Networks'. Together they form a unique fingerprint.

Cite this