Predictive analysis of brain PET images through an algorithm based on Machine Learning detects early cases of Alzheimer’s with 81% effectiveness.
- Where did I put it? – Did I take him? – I don’t remember your name. Who are you? The reality of a person with Alzheimer’s makes his day to day a forgotten memory that he has to recover. Habits, actions, decisions seem to leave and not return, although we always talk about those small moments of lucidity, where memories seem to return. Early and effective detection of Alzheimer’s could slow its progress and even stop it. Once again, technology, and in particular Artificial Intelligence, becomes the best ally for the most accurate possible prediction of this disorder. In fact, IT’s not just IT that helps fight or cope with this disorder. From the BlogThinkBig.com we reveal three other technological initiatives that are carried out to make life easier for people who suffer from this disease.
Alzheimer’s represents 70% of dementia cases that arise in modern society and its projection in the future is less than encouraging. According to the 2016 global Alzheimer’s report, the number of people affected by dementia is expected to be 131.5 million by 2050, three times the 2015 figure of 46.8 million. Therefore, scientists and doctors advance every day in their research, which combines medicine and technology to be able to “forget forever” Alzheimer’s. In this scenario, publications such as Hipertextual have also told in depth how it affects another of the degenerative diseases that exist today, the reality of Parkinson.
Apparently, a group of researchers from the Public Health Center in South Korea they carried out a study that replaced the brain image analysis work of a doctor with a machine that incorporates machine learning technology. Predictive analysis of brain PET imaging what the machine does through a new algorithm, according to the study, discovers the invisible signals to the human eye that would help identify early the chances of developing the disorder with 81% accuracy. However, in advanced cases, research is still ongoing to curb its effects, which is why we emphasize the importance of early detection of Alzheimer’s, as it becomes the best method to combat the disease. The South Korean team assures that thanks to this machine they will be able to predict with a large percentage of accuracy the chances that a person has of suffering Alzheimer’s during the next three years.
Machine learning that you don’t forget
We all notice the passage of time, we age, and it is normal to suffer a cognitive decline within our brain. Most doctors determine it as a mild impairment, being inevitable for all people. But when an unusual decline occurs, it can lead to disorders such as Alzheimer’s. In fact, almost without realizing it you can move from one side to another, that is, many people who have a mild deterioration can end up suffering from Alzheimer’s. Among most of its first symptoms we find: lack of vocabulary or a misuse of it, not recognizing certain relatives, as well as memory loss in everyday habits or disorientation.
The best way that exists today to identify the second group of patients prone to Alzheimer’s is mainly based on the analysis of brain PET images. This type of imaging based on positron emission technology is able to reflect changes in the body at the cellular level. Therefore, it becomes the best method for the early detection of a disease. But, once the images are taken, the decision falls to the human eye of a doctor, a fairly complex task that in some cases becomes erroneous. This scenario is when intelligent technology appears, the machine learning built into a machine capable of “guessing the brain future” of a patient.
And how does this machine work? Its operation is very simple, the philosophy of Big Data is again the protagonist. The new algorithm, integrated into the appliance, collects a database with more than 300 brain images between healthy people and Alzheimer’s patients. During the first test comparing a healthy brain and another with Alzheimer’s, the accuracy of its diagnosis was close to 90%. In the second test they decided to change the sample to 181 patients who suffer a mild deterioration, of which 79 were already detected with Alzheimer’s within three years. The aim of this second experiment was to train the machine to discover these 79 cases. The results were very positive. The algorithm installed in the device had an efficiency of 81%, a high figure in contrast to specialist doctors who use the human eye to analyze brain images.
The use of this technique, in addition to its high probability of success, will speed up the process of early detection of this type of disease. Researchers explain that once again it has been discovered how machine learning becomes the best tool for analytical medicine. It is therefore demonstrated that Artificial Intelligence has come to stay in all areas of our lives.