In this paper, a three dimensional convolutional neural network based solution is proposed for classification of brain tissues as lesion or healthy in terms of ischemic stroke disease. Three dimensional data used in this work are obtained by magnetic resonance imaging technique. Proposed method is compared with traditional methods that are in the same category, via K-fold cross validation technique in terms of sensitivity, specificity and accuracy measures. In conclusion, it is obtained nearly 89% accuracy using our proposed method. Comparing this method with others, our proposed method is the best method.