Brain age cascade refining prediction method and system based on structural magnetic resonance image

Abstract

The invention discloses a brain age cascade refining prediction method and system based on a structural magnetic resonance image. The brain age prediction method comprises the following steps: obtaining a training sample set comprising a brain structure magnetic resonance image, a gender label and a real age of a healthy person; constructing a two-stage cascade refining network, wherein the first-stage network and the second-stage network are multi-scale tight connection networks and the output of the first-stage network is connected with the input of the second-stage network; based on the training sample set, training the two-stage cascade refining network by adopting a back propagation algorithm and a gradient descent algorithm to obtain a brain age prediction model; verifying the brainage prediction model by adopting a cross-validation method to obtain a verified brain age prediction model; inputting the test set into the verified brain age prediction model to obtain a brain age prediction result, wherein the test set comprises a brain structure magnetic resonance image and a gender label of a to-be-tested person. The method can improve the prediction accuracy of the brain age.

Description

technical field [0001] The present invention relates to the technical field of brain age prediction, in particular to a brain age cascade refinement prediction method and system based on structural magnetic resonance images. Background technique [0002] As the global population ages, aging-related brain diseases are placing an increasing burden on society. The human brain undergoes some subtle structural changes with age, which can lead to a degeneration of the brain's normal function and show a significant correlation with brain diseases such as neurodegeneration. Reasons such as genes, environment, disease or injury may cause the aging rate of the brain to be significantly accelerated, and methods need to be provided to quantify this abnormal brain aging rate to assess the current aging stage of the brain. [0003] Artificial intelligence methods can use brain structural magnetic resonance images to establish a prediction model of brain aging, so as to predict the age of...

AI Technical Summary

Problems solved by technology

[0005] In addition, convolutional neural networks are being tried to be used in brain age prediction, and this method also has several shortcomings: (1) Structural magnetic resonance images are different from natural images, it is a three-dimensional image, and the training sample size is much smaller than Natural images, so this poses a high challenge to the structure of the convolutional neural network, requiring the convolutional neural network to fully extract and utilize the information contained in the magnetic resonance image

However, the convolutional neural network currently used in brain age prediction tasks is too traditional, only using the structure of convolution + pooling + full connection, and the extraction and utilization efficiency of features need to be improved

(2) In the process of using the backpropagation algorithm to optimize a large number of parameters in the convolutional neural network, the existing methods only use the mean absolute error (MAE) or mean square error ( MSE) as a loss function

Therefore, the brain age prediction method based on the traditional convolutional neural network has the problem of low prediction accuracy.

Images