This perspective emphasizes that the brain-machine interface (BMI) research has the potential to clarify major mysteries of the brain and that such clarification of the mysteries by neuroscience is needed to develop BMIs. I enumerate five principal mysteries. The first is how is information encoded in the brain? This is the fundamental question for understanding what our minds are and is related to the verification of Hebb’s cell assembly theory. The second is how is information distributed in the brain? This is also a reconsideration of the functional localization of the brain. The third is what is the function of the ongoing activity of the brain? This is the problem of how the brain is active during no-task periods and what meaning such spontaneous activity has. The fourth is how does the bodily behavior affect the brain function? This is the problem of brain-body interaction, and obtaining a new body by a BMI leads to a possibility of changes in the owner’s brain. The last is to what extent can the brain induce plasticity? Most BMIs require changes in the brain’s neuronal activity to realize higher performance, and the neuronal operant conditioning inherent in the BMIs further enhances changes in the activity.