Alzheimer’s disease is the most common form of dementia, accounting for up to 70 percent of all dementia cases, according to the World Health Organization. Around 24 million people are impacted worldwide, and this number is anticipated to double every 20 years. Because of societal ageing, the disease will become an expensive public health burden in the coming years.
Mild cognitive impairment (MCI), the stage between the expected cognitive decline of normal ageing and dementia, is one of the possible early indicators of Alzheimer’s disease.
According to the study authors of new study, functional magnetic resonance imaging (fMRI) can be utilised to identify brain regions related to the start of Alzheimer’s disease.
The first stages of MCI frequently have absolutely no obvious symptoms, but neuroimaging can detect them in a small number of individuals.
However, while theoretically conceivable, manually analyzing fMRI scans in order to identify the alterations associated with Alzheimer’s disease is not only time-consuming but also requires specific knowledge; the use of Deep learning and other AI approaches can significantly speed up this process.
Finding MCI traits does not definitely indicate the presence of sickness, since it can also be a sign of other associated disorders, but it is more of an indicator and possible guide to a medical professional’s evaluation.
To overcome all these challenges, the deep learning-based model was created as a result of a fruitful collaboration between leading Lithuanian researchers in the Artificial Intelligence sector, who used a modified version of the well-known fine-tuned ResNet 18 (residual neural network) to classify functional MRI images from 138 subjects. The photos were divided into six categories, ranging from healthy to mild cognitive impairment (MCI) to Alzheimer’s disease. The Alzheimer’s Disease Neuroimaging Initiative fMRI dataset was used to train and validate 51,443 and 27,310 images, respectively.
The model successfully identified MCI features in the supplied dataset, with classification accuracy of 99.99 percent, 99.95 percent, and 99.95 percent for early MCI vs. AD, late MCI vs. AD, and MCI vs. early MCI, respectively.
According to senior researchers, the above-described model can be integrated into a more complex system, analyzing several different parameters, such as monitoring eye-movement tracking, face reading, voice analysis, and so on.
Such technology might subsequently be used for self-checking and alerting people to seek expert help if anything is causing them concern.
Image Credit: iStock
