Over the past few years, challenges remain in producing an accurate brain structures segmentation due to the imag- ing nature of Magnetic Resonance images, that is known to exhibit similar intensity characteristics among subcortical structures such as the hippocampus, amygdala and caudate nucleus. Lack of a distinct image attributes that separate adjacent structures often hinders the accuracy of the segmentation. Therefore, researches have been directed to infer prior knowledge about the possible shape and spatial location to promote accurate segmentation. Realizing the importance of prior information, this focused review aims to introduce brain structures segmentation from the perspective of how the prior information has been utilized in the segmentation methods. A critical analysis on the methodology of the brain segmentation approaches, its’ advantages and issues pertaining to these methods has been discussed in detail. This review also provides an insight to the current happenings and future directions in brain structure segmentation.