4 ms·

The article is about someone whose brain implant failed, and concludes that the problem was caused by electromagnetic interference. But this is pure guesswork;

The article is about someone whose brain implant failed, and concludes that the problem was caused by electromagnetic interference. But this is pure guesswork; no measurements were taken, no tests were performed and no electrical engineers were consulted.

zellyn 9y ago
> 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.
- mi100hael 9y ago
  The plural of anecdote is not data.
  - saghm 9y ago
    50 out of 50 people is not insignificant
  - zellyn 9y ago
    That's a cute (albeit tired) quote. But it's not 50 anecdotes. It's an informal survey of 50 people.
throwanem 9y ago
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.
jerf 9y ago
"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.
- sin7 9y ago
  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.
lowbloodsugar 9y ago
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.

zellyn 9y ago

> 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.

mi100hael 9y ago
The plural of anecdote is not data.
- saghm 9y ago
  50 out of 50 people is not insignificant
- zellyn 9y ago
  That's a cute (albeit tired) quote. But it's not 50 anecdotes. It's an informal survey of 50 people.

throwanem 9y ago

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.

jerf 9y ago

"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.

sin7 9y ago
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.

lowbloodsugar 9y ago

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.