Evaluating the possibility of a change from Mild Cognitive Impairment (MCI) to dementia of Alzheimer's Disease (AD) is a very difficult task. Single-Photon Emission Computerized Tomography (SPECT) scans of 2146 people using the Deep Learning (DL) algorithm was developed and checked to predict the development of AD dementia and MCI patients in a time-toevent assessment scenario. Furthermore, the danger of early development divided each participant into small groups with separate timeframes for developing AD dementia. The threat of development based on AD has been merged with the basic clinical measures, achieving improved prediction of the advancement of AD dementia. A hybrid SPECT - Improved Faster Recurrent Convolutional Neural Network (IFRCNN) provides a precise, affordable way to predict the disease and makes it easier to enroll patients in clinical studies who are expected to advance within a given time frame.