Cross-species machine learning can improve the accuracy of MRI diagnosis of mental diseases

[ Instrument R&D of Instrument Network ] On June 17, the American Journal of Psychiatry published a research paper titled Diagnostic Classification for Human Autism and Obsessive-Compulsive Disorder based on Machine Learning from a Primate Genetic Model online.
The research was conducted in collaboration with the research team of Wang Zheng, a researcher of the Center for Excellence in Brain Science and Intelligent Technology of the Chinese Academy of Sciences (Institute of Neuroscience), Shanghai Brain Science and Brain-like Research Center, State Key Laboratory of Neuroscience, and He Ran's research group of the Institute of Automation, Chinese Academy of Sciences Complete, integrate primate animal models and functional MRI data of patients with clinical psychiatric diseases, design the monkey-human cross-species machine learning analysis process for the first time in the world, and use the features learned from the transgenic macaque model to construct clinical psychiatric patients The classifier model, in-depth analysis of the neural circuit mechanism of human autism and obsessive-compulsive disorder, provides new evidence for the accurate diagnosis of mental diseases, and discovers the use of non-human primate models to serve the clinical application needs of mental diseases. way.
Autism (ASD) is a developmental disorder of nervous system disorders with high heterogeneity. Patients are often accompanied by complications such as obsessive-compulsive disorder (OCD) and attention deficit hyperactivity disorder (ADHD), which provide clinical diagnosis and pathological mechanisms. Research brings great challenges. Non-human primate model animals and humans are relatively similar in brain structure and function. Researchers have previously found that transgenic primate models can exhibit similar symptom phenotypes as human clinical patients, such as MECP2 overexpressing cynomolgus monkeys. Stereotyped behaviors, social behavior disorders and other autistic symptoms (Nature, 2016), and abnormalities in the brain circuit are also similar to some autistic patients (J Neurosci, 2020).
Based on the previous work (IEEE TMI, 2015), the research team explored possible evolutionary conservative characteristics among primate species, assuming that based on the conserved brain function, a psychiatric classification and prediction model that can migrate across species (Figure A). In this study, the Group LASSO algorithm was used to screen brain function map data from 5 transgenic rhesus monkeys and 11 wild-type rhesus monkeys, and 9 core brain regions were identified (Figure B); 9 brain regions were mapped to The human brain is used to form feature sets using the functional connections of the brain regions, and constructs a sparse logistic regression classifier, which is used for the diagnosis and classification of patients with autism, obsessive-compulsive disorder, and attention deficit hyperactivity disorder, respectively. The patient data came from 4 clinical image databases including ABIDE-I (1112 persons), ABIDE-II (1114 persons), OCD (186 persons) and ADHD-200 (776 persons). After cross-validation, the study found that the classification model based on the characteristics of transgenic macaques can distinguish between autistic patients and normal persons in the ABIDE-I data set with an accuracy rate of 82.14%, and 75.17% of the human subjects in the ABIDE-II database. The accuracy rate is significantly higher than the performance of building a classifier based on the characteristics of the patients with autism and obsessive-compulsive disorder (Figure C). When the same 9 brain regions were extended to obsessive-compulsive disorder image data, the study found that the macaque monkey feature classification model can still achieve 78.36% accuracy, which is significantly higher than the performance of the classifier based on the features of the autistic patients. However, these features learned based on the macaque model failed to significantly improve the classification accuracy of ADHD patients. Further analysis of the relationship between the functional connections in these superior classifiers and the clinical symptoms of psychiatric disorders, the study found that the right ventral lateral prefrontal cortex plays a dual role in autism and obsessive-compulsive disorder, corresponding to their respective Specific dimension symptom phenotype (Figure D).
The study was jointly completed by Wang Zheng and He Ran under the supervision of doctoral students Zhan Yafeng and Wei Jianze, and was supported by Cambridge University, Pediatric Hospital Affiliated to Fudan University, Kunming Institute of Zoology, Chinese Academy of Sciences and the Ministry of Science and Technology, National Natural Science Foundation of China, Chinese Academy of Sciences, Shanghai, Guangdong and other subsidies.

Kaiten Conveyor Belt Accessories

Established in 2008, Guangzhou Yuyang Food Machinery Co., Ltd. has been committed to becoming a domestic technology enterprise specializing in automated food delivery, providing professional solutions for large-scale catering brand enterprises with intelligent and automated food delivery.

For the mid to high-end customer market, Guangzhou Yuyang Machinery has also targeted the development of intelligent automatic rail delivery vehicles, which can deliver meals directly from the kitchen to customers through different styles of molds.

Kaiten Conveyor Belt Accessories,Food Grade Rotary Sushi Belt,Rotary Sushi Belt,Conveyor Belt Accessories

Guangzhou Yuyang Food Machinery Co., Ltd , https://www.yysushibelt.com

Posted on