The brain is thought to implement canonical neural computations, i.e. the same mathematical operations, in a variety of contexts. The method of population receptive fields (pRFs) allows quantitative modeling of responses to external stimuli, and has been used extensively to investigate neural computations in health and disease. Here, we build a pRF model based on divisive normalization. Divisive normalization pRF models provide a biologically-inspired, unified modeling framework for seemingly different properties of responses to spatial visual stimuli observed across the visual hierarchy. Moreover, a fundamental question is how canonical computations link to the underlying biological substrate. We shed light on this question by investigating the computational roles of neurotransmitter receptors in divisive normalization. We find significant alignment between neurotransmitter receptor densities and the divisive normalization parameters throughout the human visual hierarchy. These results provide an algorithmic link between the brain's computational capabilities and their biological implementation. We propose that the brain employs canonical computations, and that neurotransmitter systems provide response flexibility within the framework of these canonical computations.