Study On Neural Network Of Alzheimer Disease Based On Clinical Scale And Diffusion Tensor Imaging

With increasing aged population,research on high incidence of senile diseases is particularly important.Alzheimer Disease(AD)is one of the disease because of its early characteristics are not obvious and it is almost incurable in the latter stage.Nowadays,the clinically recognized stage of AD can be divided into the stage of amnestic mild cognitive impairment(aMCI),and the stage of severe dementia of AD.Meanwhile related studies have shown that the cognitive function of AD inpatients with first injury is episodic memory.At present,the description of the disease is not comprehensive and objective and the interdiscipline between different disciplines is not enough.Hence this paper propose a novel means combining imaging with clinical genomics and provide high-precision and low-cost auxiliary diagnosis and treatment methods forAD.Magnetic Resonance Image(MRI)can observe the internal structure of brain without intervention,hence this method has become an important means of clinical diagnosis and treatment of AD.In this paper,we utilize Diffusion Tensor Image(DTI)to track the trend of nerve fibers,the individual brains are partitioned by related templates.Then we generate the corresponding neural fiber network and utilize the graph theory.At the same time,we utilize multi-dimensional scale test in the field of neuropsychological to analyze clinical genomics.We select theAuditory Verbal Learning Test(AVLT)to test the context memory function.In order to analyze AD comprehensively and improve the diagnosis efficiency of the AD,this paper will separately study the imaging aspect and clinical aspect and establish the connection between the two aspects.We extract the imaging parameters of the brain regions of AD,aMCI and NC episodes.Then we further analyze the differences between groups and extract the features related to the disease.Meanwhile,we analyze the correlation between characteristic parameters and the AVLT scale score.It provides a theoretical basis for the subsequent construction of the auxiliary diagnosis and treatment model.In order to significantly improve the diagnosis efficiency of AD and reduce the resource waste of disease diagnosis,this paper utilize machine learning to construct an AD auxiliary diagnosis and treatment model that takes into account the accuracy and ease of use.Meanwhile considering the accuracy of the prediction model,propose a novel regression algorithm model based on the stacking framework is proposed,which can effectively use the image graph theory parameters to predict the corresponding AVLT scale scores.It improves the efficiency of AD diagnosis and can be used as a more accurate model of AD auxiliary diagnosis and treatment in the future.In summary,we utilize graph theory to quantify the neural fiber network of related brain regions.Then we extract some effective features and further analyze the correlation.Basedon the above analysis,we propose a high-precision and low-cost AD disease auxiliary diagnosis and treatment model,which improves the diagnostic efficiency of AD disease and is essential for further exploration of the diagnosis and treatment ofAD disease.