The transmission of commands and signals from the brain to an external device is done using Brain-computer interfaces (BCIs) bypassing the conventional output channels of peripheral muscles. First BCIs fetch signals and then translate brain activity signals coming from conscious and subconscious levels. In a recent report by NIH (National Institute on Health), it is stated that more than 57 million people are suffering from dementia which causes Alzheimer’s disease. This problem can be dealt with in the BCI system by learning emitted patterns by the brain and then assisting the patient in daily activity. Dementia typically attacks the brain’s functions like communication ability to make the judgments on the situation. Machine learning applications are capable of detecting cases of dementia early by health healthcare professionals, hence enabling the implementation of timely therapeutic interventions to mitigate the progression of cognitive decline. In this research paper, a BCI framework to classify EEG signals is proposed that applies SVM, XG Boost, and Neural Network classification algorithms. For the feature selection, principal component analysis (PCA) is utilized to select the most suitable features. XGBoost algorithm performs best in classification tasks to give 80% accuracy in detecting the mental disorder.