Machine Learning-Based Approach for Early Detection of Neurodegenerative Diseases Using Multimodal Neuroimaging Data

Advanced clinical diagnoses for neurodegenerative illnesses (NIs) are almost ready to be implemented due to the development of strong analytical techniques. Alzheimer's disease (AD) is one of the NIs for which there is no known cure. Early detection aids AD sufferers in maintaining a regular lifestyle. Identifying and categorizing AD phases appropriately presents a major challenge for both scientists and physicians. Numerous investigations have shown that multidimensional neuroimaging data (MND) can offer important new information about the morphological and cognitive alterations in the brain linked to AD. Employing strong computing techniques, machine learning (ML) systems can recognize trends and connections in MND to reliably classify AD stages. Using ML methods and MND from magnetic resonance imaging (MRI) methods can help diagnose AD more quickly and possibly even anticipate the way the condition will develop. Using ML models and AD screening statistics, it was additionally feasible to predict the dementia of specific elderly persons. The algorithm's efficiency can be improved by using the patient's MRI data and pre-existing diseases in predicting the AD patient condition. This study introduced a methodology depending on supervised learning algorithms for classifying dementia subjects depending on long-term brain MRI data as either AD or non-AD. Outcomes showed that the gradient boosting technique beats other algorithms with 97.6% accuracy, outperforming six distinct supervised algorithms used for the categorization of AD individuals.

Ҳали таржима қилинмаган