Gliomas are a kind of tumor that starts in the neuroglia cells of the brain or the spine. Gliomas constitute about 80 percent of all malignant brain tumors and 30 percent of all brain and central nervous system tumors. An accurate segmentation of gliomas is crucial in order to diagnosis brain tumour as well as to perform the proper treatment planning and management. Gliomas segmentation involves determination of their locations and sizes in the brain. Brain glioma segmentation in Magnetic resonance imaging (MRI) images is important to assist diagnosis, treatment planning and surgical navigation. However, manual segmentation is time consuming and is prone to bias. Thus, robust software based automatic segmentation methods are required. Machine learning based techniques particularly deep learning methods have been proposed for brain glioma segmentation and they are successful. This paper presents a review on segmentation of brain glioma in MRI images using various recent and state-of-the-art deep learning methods. These methods are identified and their performances are compared.