A study on Multimodal approach for early detection of Dementia using Deep Learning

Alzheimer impacts a patient's memory, leading to cognitive decline and the priority lies in risk reduction, providing early intervention and precise symptom diagnosis, rather than finding a cure for it. In the brain the Alzheimer's Disease is caused by the neurodegenerative changes. Approximately one in nine individuals aged 65 and older has Alzheimer's. According to the report, the Alzheimer's disease was noted as the sixth-leading cause for the loss of life in the United States. In the Alzheimer's early stages, It is not easy to predict. A medication given at an initial stage of the disease is more curable, and it causes some minor damage than a medication done at higher stage. By using various algorithms for identifying best parameters to predict Alzheimer's Disease. Utilizing a Multimodal approach with Deep Learning for early detection significantly contributes to reducing annual mortality rates associated with Alzheimer's Disease. This multimodal approach includes different types of data modalities such as biomarkers, demographics, medical history, neuro psychological test results and voice data. In view of enhancing efficiency, accuracy, precision, recall and meanF1 scores of the trained model by integrating multi- modality data to yield superior results compared to single modality model and reduce error rate.