Several hypotheses have been proposed to explain the limitation of brain size in vertebrates. Here, we test three hypotheses of brain size evolution using marine teleost fishes: the direct metabolic constraints hypothesis (DMCH), the expensive tissue hypothesis and the temperature-dependent hypothesis. Our analyses indicate that there is a robust positive correlation between encephalization and basal metabolic rate (BMR) that spans the full range of depths occupied by teleosts from the epipelagic (< 200 m), mesopelagic (200-1000 m) and bathypelagic (> 4000 m). Our results disentangle the effects of temperature and metabolic rate on teleost brain size evolution, supporting the DMCH. Our results agree with previous findings that teleost brain size decreases with depth; however, we also recover a negative correlation between trophic level and encephalization within the mesopelagic zone, a result that runs counter to the expectations of the expensive tissue hypothesis. We hypothesize that mesopelagic fishes at lower trophic levels may be investing more in neural tissue related to the detection of small prey items in a low-light environment. We recommend that comparative encephalization studies control for BMR in addition to controlling for body size and phylogeny.