Artificial intelligence predicts Alzheimer's disease. The accuracy rate of the disease exceeds 99%!

Introduction:

Currently, we have no effective treatment for Alzheimer's disease (AD), but early diagnosis and early intervention can improve the survival rate of patients

Alzheimer's disease (AD) is a progressive neurodegenerative disease with insidious onset

Researchers at Kaunas University in Lithuania have developed a method based on deep learning that can predict the possible incidence of Alzheimer's disease through brain images with an accuracy rate of over 99%

Published an article titled "Analysis of Alzheimer's Disease Features: Using Fine-tuned ResNet18 Network to Detect Early Functional Brain Changes from Magnetic Resonance Images" in "Diagnostics"

This method has performed MRI functional imaging analysis on 138 patients, and the accuracy, sensitivity and specificity of MRI imaging are better than previous methods

According to the World Health Organization, Alzheimer's disease is the most common cause of dementia, accounting for 70% of dementia cases

"Medical professionals all over the world are working hard to raise people's awareness of the early diagnosis of Alzheimer's disease, which provides patients with better opportunities for treatment benefits

Delegate image processing to the machine

Analyze the characteristics of Alzheimer's disease (AD) and create more effective and accurate tools based on the latest technology

One of the early symptoms of Alzheimer's disease may be mild cognitive impairment (MCI), which is a stage between normal aging and dementia

However, although theoretically feasible, manually analyzing fMRI images to try to identify changes related to Alzheimer's disease requires not only specific knowledge but also time
.
The application of artificial intelligence methods such as deep learning can greatly accelerate this speed
.
The discovery of MCI features does not necessarily mean that there is a disease, because it may also be a symptom of other related diseases, but it is more like an indicator and possible assistant that can guide medical professionals to evaluate
.

"Modern signal processing allows image processing to be delegated to a machine, which can complete image processing faster and more accurately
.
Of course, we dare not recommend that medical professionals rely 100% on any algorithm
.
Think of the machine as a robot
.
Complete the most tedious tasks , Sort the data, search for features
.
In this case, after the computer algorithm selects the cases that may be affected, the experts can study these cases more carefully, and ultimately, diagnose and treat faster, if it reaches the patient’s hands, every time Individuals will benefit,
said Maskeliūnas, head of the model research team
.

We need to make the most of data

This deep learning-based model was developed in a fruitful cooperation with researchers in the field of artificial intelligence in Lithuania
.
The model uses the famous ResNet 18 (residual neural network) to obtain functional MRI images of 138 subjects
.
The images are divided into six categories: from healthy to mild cognitive impairment (MCI) to Alzheimer's
.
A total of 78,753 images were studied
.
For evaluation, the data set is divided into training data set and validation data set
.
The segmentation rate was 70% (total 51443 images) and 30% (total 27310 images)
.

The data for this study comes from the ADNI (Alzheimer's Disease Neuroimaging Initiative) database (http://adni.
loni.
usc.
edu/)
.

The model can effectively discover the MCI features in a given data set
.
For early MCI and AD, late MCI and AD, MCI and early MCI, the best classification accuracy rates are 99.
99%, 99.
95%, and 99.
95%, respectively
.

Maskeliūnas said: "Although this is not the first attempt to diagnose early-onset Alzheimer's disease from similar data, our main breakthrough is the accuracy of the algorithm
.
Obviously, such a high number does not represent reality
.
But we are working with medical Agencies to collaborate to obtain more data
.
"

He said that the algorithm can be developed into software to analyze data collected by vulnerable groups (over 65 years of age, with a history of brain injury, high blood pressure, etc.
) and inform medical staff of abnormalities in data related to early-onset Alzheimer's disease
.

Maskeliūnas said: "We need to make full use of data
.
" "That's why our research group focuses on European open science principles so that anyone can use our knowledge and further develop them
.
I think this principle is important for social progress.
Meaning
.
Contribution
.
"

The main research area is the application of modern artificial intelligence methods in signal processing and multi-mode interfaces
.
He said that the above model can be integrated into a more complex system to analyze several different parameters, for example, it can also monitor eye movements
.
Track, "facial recognition", sound analysis
.
This technique can be used for self-checking and reminding
.
If there are any problems that cause your concerns, you can seek professional advice
.

Maskeliūnas said: "Technology can make medicines more accessible and cheaper
.
Although they will never (at least not soon) truly replace medical professionals, technology can encourage rapid diagnosis and help
.
"

(Source: Internet, reference only)