"Olhoeft is a retired professor of geophysics in Colorado who taught courses on electromagnetism, so with those two details, it wasn’t hard for him to figure out what had probably gone wrong. His implant, he says, operated at the same electromagnetic frequency as Best Buy’s theft detection system, and the two signals interfered with each other."

"To test such situations, he uses a detector that tells him the frequency of things like security systems to make sure they’re operating on a different frequency than his device. If the frequency is a match, he asks to go around."

Also, the statistical power of these results are quite strong. (Sometimes I think people screw up in teaching the "scientific method" as the traditional presentation tends to assume that you are dealing with some effect where you don't have a lot of statistical power, so people get this idea that you must have massive samples and do complicated statistics. This is not true. There are time you are justified in drawing strong conclusions from very small sets of data, because the statistical power is sufficient to justify it. You know this, because you do it all the time.)

And finally, this absolutely conforms with all current understandings of mainstream science and engineering. Massive amounts of engineering effort in the radio world are spent preventing, dealing with, and in some military type applications, causing radio interference. We're not talking about claims that UFOs are giving this guy Parkinson's, we're talking about perfectly sensible straightforward theories that were by far the dominant theory that an informed person would think of even before the overwhelming evidence that it is EM based.

Sample size is too small and correlation does not equal causation are the worst. Teachers did a great job teaching those things, but sometimes correlation does equal causation or at the very least is interesting enough to warrant a closer look rather than to be dismissed so easily. At the same time, a sample size of one is good enough to prove something in a scientific way. If I claim to know the lottery numbers, I can prove it by winning the next lottery. In doing so I beat 1 to 5 million odds, which are greater than the 1 in twenty that most studies require.

I'd be extremely hesitant to perform detailed testing, too, if part of the system under test happened to be installed inside my actual skull. It's not wrong to say that further investigation is warranted, but to discount the present result entirely because it has yet to be closely characterized strikes me as cavalier in the extreme.

FTA: In its manuals, Medtronic, the maker of the device, clearly advises patients that things like hairdryers, cell phones, power tools, and yes, in-store security systems may impact devices.

> I go to a support group for people with deep brain stimulation implants and I gave a talk about interference. I asked how many people had an experience like mine at Best Buy and all 50 people put their hands up.

The plural of anecdote is not data.

That's a cute (albeit tired) quote. But it's not 50 anecdotes. It's an informal survey of 50 people.

50 out of 50 people is not insignificant