I'm bad at math, so when I see articles that seek to explain why some people are good at math, and those articles don't have any math obviously present in them, I'm interested. This article, via EurekAlert, describes an investigation into dyscalculia, "the numerical equivalent of dyslexia." So, it's possible this doesn't apply to me specifically -- I don't think I'm dyscalculic, I'm just not very good at math. The study, which compared fMRI scans in which subjects were determining whether two groups of objects matched in number, and confirming the accuracy of equations presented to them.
Consistent with previous studies, the researchers found that the basic number-matching task activated the right parietal cortex, while the addition and subtraction tasks produced additional activity in the left parietal cortex. But they also found something new: During the arithmetic tasks, communication between the left and right hemispheres increased significantly compared with the number-matching task. Moreover, people who exhibited the strongest connection between hemispheres were the fastest at solving the subtraction problems.
Math is generally characterized as overwhelmingly centered in one part of the brain -- the left hemisphere. It's a fairly substantial recontexualization, in my mind, to think of math as requiring the whole brain, the way music does. That said, I'm not a neuroscientist -- I have very nearly no idea what I'm talking about when it comes to brain science.