Yesterday I read a great article about neuroscientist Craig Bennet who purchased an Atlantic salmon and put it under an fMRI machine to scan its (presumably non-working) brain. Oh, yes, by the way, it was a *dead* salmon. With the fish in the scanner, it was shown a series of pictures, and the salmon was "asked to determine what emotion the individual in the photo must have been experiencing."
While this already sounds like a fishy story, wait - it gets better: during the scanning, it appeared that "voxels" (video images indicating activity) showed up in the fish's brain. The simple conclusion? Dead fish can think!
Of course, that is ridiculous, and the purpose of the scanning was not to show that dead fish have brain activity when asked questions, but to look at how on occaision, random noise that creeps into these scans appears to show false positive information, with the objective to suggest more rigorous validation of statistical methods when attempting to filter out random noise so as to prevent drawing conclusions from what is potentially flawed data.
While I would call this a less orthodox process for establishing the value of data quality, it certainly provides a general lesson regarding quality of information on output from a reporting activitiy.
Posted September 24, 2009 6:09 AM
Permalink | No Comments |