Am J Psych: Chinese scientists use primate magnetic resonance imaging fusion machine learning technology to successfully analyze the loop mechanism of human mental illness

, June 18, 2020 /
PRNewswire/ -- A recent publication in the International Journal ofAmerican Journal of theentitled "Diagnostic for Human Autis m and Creed-Compulsive Disorder On Machine Learning From from a Primate In model's study, "
Diagnostics
for Autism and Obsessive Compulsive Disorders based on primate genetically modified models", scientists from institutions such as the Center for Excellence in Brain Science and Intelligent Technology excellence in the Chinese Academy of Sciences successfully analyzed the loop mechanism of human mental illness using primate magnetic resonance imaging fusion machine learning techniquesphoto source: Yafeng Zhan, et aldoi: 10.1176/
app
i.ajp.2020.19101091
The CDC's 2018 statistics show that the incidence of autism among children is as high as 1/59, with an average of one in every 100 children in the countryAutism (also known as autism) is a kind of widespread developmental disorder
hereditarymental illness, which is characterized by social disorders, repetitive stereotypes and narrow interestsMoreover, nearly 75 percent of children with autism are often associated with other mental disorders, such as generalized anxiety disorder, attention deficit hyperactivity disorder, and obsessive compulsive disorderThe research on the pathogenesis of autism, as well as the diagnosis andand intervention treatment of clinicalare facing great challenges due to its high heterogeneity, high co-incidence rate and clinical epialatical overlapthe brain structure and function of non-human primates are similar in evolution to humans, and have become the first choice for studying advanced cognitive functions of the brain, and it is also the best animal model to explore the pathogenesis of complex brain diseases, develop diagnostic techniques and transform them into clinical practiceA genetically modified crab monkey model with multiple copies of MECP2, created by a team of neuro-researchers and Sun Qiang, showed symptoms such as repetitive stereotypes and anxiety that were highly similar to those in humans with autism (Nature, 2016)The team's earlier use of magnetic resonance imaging technology has found that genetically modified macaques are highly similar to the brain function connectivity map in a small number of autistic patients (J Neurosci, 2020)As a result, the researchers continued to explore the mapping relationship between monkeys and humans on the neural loop of the brain, building machine learning models, and for the first time in the world realized a cross-species migration-like classification prediction model for magnetic resonance imagingdiagnosticfor human mental illnessthe experimental and challenging problem of how to effectively refine the brain network mapping relationship between macaque model and human patients for ultra-high dimensional functional magnetic resonance imaging dataWang Zheng, a researcher at the Brain Science and Intelligence Technology Center of the Chinese Academy of Sciences, worked closely with a team of researchers to boldly carry out the original innovation, and for the first time proposed a "monkey-human" cross-species migration-like classification prediction model for imagingdiagnosticof human mental illness, with a brief flow shown in Figure 1Complex brain network maps can be abstracted into the composition of nodes (brain regions) and edges (functional connections)An edge indicates that there is some "connection" between the two nodes to which it is connectedEvolutionaryly, primates and human brains have higher homologous ness and relatively conservative functions in the brain regionTherefore, the researchers chose to construct a sparse learning model with brain region as the core, extract the core brain region associated with genetically modified effect in the brain map of the functional magnetic resonance image of macaques, and then map these core brain regions to the brain map of the human functional magnetic resonance image, extract the functional connection characteristics associated with the core brain region, and train the sparse learning model to classify and predict the patients with mental illness builds a predictive model for cross-species machine learning classification using functional magnetic resonance brain map features, Extracting magnetic resonance imaging features assists diagnosis mental illness in humans photo source: Yafeng Zhan, et al doi:10.1176/
app
i.ajp.2020.19101091
researchers used sparse learning methods to migrate nine core brain regions extracted from macaque image data to 336 human trials screened from the human autism magnetic resonance imaging public database, ABIDE-I This cross-species classification prediction model distinguished between autistic and normal people with an accuracy of 82.14 percent, and was validated in 169 separate human trials (from ABIDE-II), also achieving 75.17 percent accuracy (Figure 2B) Neuroimaging evidence suggests that autism has similar functional network abnormalities to mental illnesses such as attention deficit hyperactivity disorder and obsessive-compulsive disorder As a result, the researchers tried to apply this model to diagnosis people with ADHD and obsessive compulsive disorder in humans The results showed that the prediction accuracy rate of human obsessive-compulsive patients reached 78.36 percent, and the classification performance was significantly better than that of the prediction model based on human autistic patients, but the identification of people with adhd attention deficit was not significantly improved Further analysis of the correlation between the connectivity characteristics of the brain map and clinical symptoms revealed that the right outer abdominal prefrontal cortex functional connection plays a dual role in autism and obsessive compulsive disorder, corresponding to different dimensions of symptom phenotype - compulsive ness in autism and obsessive compulsive disorder (A) learned nine core brain regions on genetically modified macaques; (B) used classifiers derived from the core brain regionof the genetically modified macaque and the core brain region seismographic in humans with autism to classify the magnetic resonance imaging data of people with autism, obsessive compulsive disorder and hyperactivity to predict the ROC curve of related performance, and (C) single-gene mutant primate models to help resolve the neuro-loop characteristics of autism and obsessive-compulsive disorder sharing in humans photo source: Yafeng Zhan, et al doi:10.1176/
app
i.ajp.2020.19101091
this study establishes the intraloop phetype of autism and obsessive compulsive disorder in the core prefrontal cortex, prefrontal cortex and other core brain regions of the right abdomen, providing an objective basis for imaging diagnosis for mental illness More importantly, the results of this study open up a new road map for exploring the loop mechanism of complex brain diseasein in humans with the help of non-human primate models, and provide new reference clues for the development of new interventional treatment techniques (BioValleyBioon.com) original origins: Yafeng Zhan, M.S., Gianze Wei, M.S., Jian Liang, Ph.D., et al.
Diagnostic Classification for Human Autism and Compulsive Disorder On Machine Learning from a Primate Genetic Model , American Journal of (2020) doi: 10.1176/
app
i.ajp.2020.19101091