The act of screaming, often viewed as a primal expression of intense emotion, serves a cathartic purpose that can help individuals release negative emotions during times of nervousness. Psychologists have recognized the therapeutic benefits of emotional release through screaming and have incorporated it into various therapeutic approaches. Understanding the intricate workings of the human brain during the act of screaming can be achieved through innovative medical tools, with Magnetic Resonance Imaging (MRI) standing as a paramount choice. Even without dedicated fMRI analysis, extracting texture data from MRI images can contribute significantly to our understanding of the brain's functional processes. In this study, we involved a multidisciplinary team and the purpose was to investigate, for the fist time, if there are any brain oxygenation level changes that appear in the MRI scans while the subject is screaming and if these can be detected using deep learning models.MRI scans were carried out as part of evaluation of brain oxygenation levels during screaming for 4 subjects. Two types of recordings were acquired from each one of them: pre-screaming MRI and screaming MRI. The interpretation of the MRI images was carried out with deep-learning algorithms trained to automatically detect the differences in the images. We used an AlexNet classifier and the results show that brain oxygenation changes appear while a person is screaming and can be automatically detectable using MRI analysis. The classifier used attained an accuracy of 89.92%, underscoring discernible distinctions between MRI scans captured during both normal, non-screaming conditions and instances of screaming. The article summarizes significant open questions, lessons learned, and these preliminary results are very valuable, opening up numerous study directions for future research.