Home SECURITY Artificial intelligence has learned to measure chaos in the brain and find signs of Alzheimer’s disease

Artificial intelligence has learned to measure chaos in the brain and find signs of Alzheimer’s disease

0
Artificial intelligence has learned to measure chaos in the brain and find signs of Alzheimer’s disease

[ad_1]

Artificial intelligence has learned to measure chaos in the brain and find signs of Alzheimer’s disease

Scientists have developed a new algorithm that evaluates the degree of disorder in the data and classifies electroencephalograms.

Scientists from Russia and other countries submitted a new algorithm that uses a neural network to estimate the degree of disorder in data and helps classify data using machine learning. As an example, the scientists applied their algorithm to electroencephalograms of healthy people and patients with Alzheimer’s disease and were able to distinguish between them with an accuracy of more than 70%. The work, supported by the Russian Science Foundation, was published in the journal Algorithms.

To analyze data from different fields of science and practice, the concept of entropy is often used, which characterizes the degree of chaos, disorder and uncertainty in data. The higher the entropy, the more difficult it is to predict the behavior of a system or event. There are many ways to calculate entropy, but they are not always accurate and resistant to noise. Therefore, scientists have developed a new method that uses a neural network to determine a special type of entropy – NNetEn entropy (Neural Network Entropy – entropy on a neural network).

A neural network is trained on time series – sequences of numbers or random variables that change over time. As training data, the scientists used the MNIST database, consisting of handwritten numbers from 0 to 9. The neural network converted the numbers into time series and calculated their NNetEn entropy. The scientists then applied their algorithm to other types of data, such as electroencephalograms.

An electroencephalogram (EEG) is a recording of the electrical activity of the brain that can serve as a diagnostic tool for various diseases. One of these diseases is Alzheimer’s disease, which leads to a gradual loss of memory and cognitive functions. The scientists took a ready-made database of 65 patients, 29 of whom were healthy and 36 with Alzheimer’s disease. The task of the neural network was to separate these two groups by the value of the entropy NNetEn.

It turned out that the NNetEn entropy alone was not enough for accurate classification, so the scientists used a combination of different types of entropies that react differently to the randomness of the data. So, when adding another entropy to the NNetEn entropy, for example, approximate, approximate, permutation or fuzzy entropy, the classification accuracy increased to 73%. This means that the new NNetEn entropy works synergistically with other entropies and may be useful in detecting early signs of Alzheimer’s disease.

Scientists note that their method can be applied not only to EEG, but also to other types of data that have a chaotic structure. For example, to audio signals, seismic vibrations, cardiograms and charts of currency pairs. The neural network for calculating the entropy NNetEn is in the public domain and can be used by other researchers for their own tasks.

[ad_2]

Source link

www.securitylab.ru

LEAVE A REPLY

Please enter your comment!
Please enter your name here