Basic cognitive mechanisms, such as the abilities to briefly maintain, focus, and process information, decline with age. Related fields of cognitive aging research have been advancing rapidly, but mostly independently, at the biological, information processing, and behavioral levels. To facilitate integration, this article reviews research on cognitive aging at the different levels, and describes a recent integrative theory postulating that aging-related deficiencies in neurotransmission cause increased noise in information processing and less distinctive cortical representation, which in turn lead to cognitive deficits. Aging-related attenuation of catecholaminergic modulation can be modeled by lowering a neural network parameter to reduce the signal-to-noise ratio of information processing. The performance of such models is consistent with benchmark phenomena observed in humans, ranging from age differences in learning rate, asymptotic performance, and interference susceptibility to intra- and interindividual variability and ability dedifferentiation. Although the details of the conjectured sequence of effects linking neuromodulation to cognitive aging deficits await further empirical validation, cross-level theorizing of the kind illustrated here could foster the coevolution of related fields through cross-level data synthesis and hypothesis testing.