<p>In the modern world, brain disorders have become a major global healthcare challenge of contemporary life due to their high prevalence and associated morbidity. The frequency of brain disorders that include a spectrum of illnesses is demanding rapid treatment, which has prompted a global effort to develop robust remedies. Magnetic Resonance Imaging (MRI) is the most preferred digital imaging modality for brain abnormality detection. It provides a detailed insight into the anatomical and functional structure of the brain. Artificial Intelligence (AI) offers promising tools, particularly through classification methods, to analyze these MRI images and identify brain anomalies. This article aims to provide a comprehensive overview of various classification methods employed within AI-based systems that emphasize the pivotal roles played by Machine Learning (ML) and Deep Learning (DL) in the efficient detection of brain abnormalities in MRI images. Though traditional ML classification models like Support Vector Machines (SVM) and Random Forests (RF) support extracted features from MRI images to distinguish normal from abnormal structures, the surge of DL-based classification methods has revolutionized this field. DL classification models such as Convolutional Neural Networks (CNNs) excel at automatic feature extraction, capturing intricate patterns in MRI data. Also, advancements in DL classification methods with Generative Adversarial Networks (GANs) and Autoencoders (AE) enhance data augmentation and anomaly reconstruction, leading to higher accuracy. Further hybrid classifiers using ML classifiers coupled with DL demonstrate notable efficacy in enhancing diagnostic accuracy. This article summarizes the potential, challenges, and future prospects of various AI classification methods, showcasing promising advancement towards improving accuracy and efficiency for brain abnormality detection in MRI images, ultimately aiming to improve early diagnosis and patient care.</p>