Functional neuroimaging of small brainstem structures in humans is gaining increasing interest due to their potential importance in aging and many clinical conditions. Researchers have used different methods to measure activity in the locus coeruleus (LC), the main noradrenergic nucleus in the brain. However, the reliability of the different methods for identifying this small structure is unclear. In the present article, we compared four different approaches to estimate localization of the LC in a large sample (N = 98): 1) a probabilistic map from a previous study, 2) masks segmented from neuromelanin-sensitive scans, 3) components from a masked-independent components analysis of the functional data, and 4) a mask from pupil regression of the functional data. The four methods have been used in the community and find some support as reliable ways of assessing the localization of LC in vivo in humans by using functional imaging. We report several measures of similarity between the LC masks obtained from the different methods. In addition, we compare the similarity between functional connectivity maps obtained from the different masks. We conclude that sample-specific masks appear more suitable than masks from a different sample, that masks based on structural versus functional methods may capture different portions of LC, and that, at the group level, the creation of a "consensus" mask using more than one approach may give a better estimate of LC localization. KeywordsLocus coeruleus, brainstem, MRI, neuromelanin, functional connectivity Figure 1. Illustration of approaches for LC localization. Group approaches emphasize voxels of highest probability of containing LC at the group level. Individual approaches focus on identifying LC-containing voxels in the individual brainstem.A recent systematic review of imaging studies of LC (Liu et al., 2017) highlights the large variety in the methods used to identify the region that more closely corresponds to the LC. A majority of the reviewed studies used coordinates or masks