This document describes a novel approach to automated classification of brain tumors using probabilistic neural networks (PNN). It discusses how principal component analysis (PCA) can be used to reduce the dimensionality of magnetic resonance (MR) brain images, and then a PNN can classify the tumors. The proposed method involves using PCA for feature extraction and a PNN for classification. This is intended to provide faster and more accurate classification of brain tumors in MR images than conventional human-based methods.