Brain age prediction methods, systems, media and electronic devices

By combining quantitative magnetic susceptibility imaging with a two-stage cascaded convolutional neural network based on deep learning algorithms, the problems of low accuracy and poor generalization in brain age prediction in existing technologies have been solved, achieving efficient brain age prediction and risk assessment for neurodegenerative diseases.

CN116313102BActive Publication Date: 2026-06-30SHANGHAI JIAOTONG UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHANGHAI JIAOTONG UNIV
Filing Date
2023-05-10
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing brain age prediction methods are mostly based on structural magnetic resonance imaging, which has limited morphological features and cannot obtain neural-related information. Furthermore, traditional machine learning algorithms are easily affected by human factors, resulting in poor model generalization and low accuracy.

Method used

A deep learning algorithm based on quantitative magnetic susceptibility imaging is used to automatically extract brain image features for brain age prediction by combining quantitative magnetic susceptibility images and gender information through a two-stage cascaded convolutional neural network. This includes steps such as phase unwinding, brain mask generation, and dipole inversion to construct a brain age prediction model.

Benefits of technology

It improves the accuracy and generalization ability of brain age prediction, can effectively detect pathological changes such as brain calcification and iron deposition, and provides a basis for risk assessment of neurodegenerative diseases, showing good clinical application prospects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116313102B_ABST
    Figure CN116313102B_ABST
Patent Text Reader

Abstract

This invention provides a method, system, medium, and electronic device for predicting brain age. The method includes the following steps: acquiring brain MRI data, gender, and true age of a healthy human body; reconstructing a quantitative magnetic susceptibility image based on the brain MRI data; constructing a brain age prediction model; training the brain age prediction model based on the quantitative magnetic susceptibility image, the gender, and the true age, and then using the trained brain age prediction model to achieve brain age prediction. The brain age prediction method, system, medium, and electronic device of this invention, based on quantitative magnetic susceptibility imaging, uses a deep learning algorithm to predict brain age, achieving excellent brain age prediction performance and possessing good generalization and clinical translational potential.
Need to check novelty before this filing date? Find Prior Art