Aerobic glycolysis, the efficiency tradeoff hypothesis, and the biological basis of neuroimaging: A solution to a metabolic mystery at the heart of neuroscience.
Aerobic glycolysis is a form of glucose-inefficient metabolism that occurs when cells metabolize glucose without oxygen, despite oxygen being abundant; the result is less energy per glucose molecule and increased glucose consumption. Aerobic glycolysis in the brain is a metabolic paradox: this inefficient metabolic process is a hallmark of neural activity, yet brains supposedly evolved to be energy-efficient. We discuss this paradox and introduce a possible solution, formalized as the efficiency tradeoff hypothesis: aerobic glycolysis, despite using glucose inefficiently, allows for energy-efficient communication. It allows axon diameter to be minimized (decreasing energy costs of communication) while allowing energy production to closely adhere to unpredictable, rapid-on/rapid-off energy demands. We expand on this hypothesis—linking observations across levels of analysis, from cognitive function to its biological implementation in the brain—culminating in a novel interpretation of the blood-oxygen level-dependent (BOLD) signal, which is closely related to localized metabolic changes caused by aerobic glycolysis. We hypothesize that the BOLD signal indexes bottom-up sensory encoding, or more specifically, prediction error in predictive processing models. This implies that much of a brain’s function, which is implemented with predictive signaling, is not indexed by BOLD fMRI. We then elaborate on the implications of our account for (a) how the evolution of human cytoarchitecture may relate to metabolism and brain function, (b) how social behavior may depend on metabolic cost functions, and (c) how metabolism may play a fundamental role in mental illness. We conclude that aerobic glycolysis and the efficiency tradeoff hypothesis offer a generative foundation for future neuroscientific research.