This is an intriguing article about the impact of head movement on fMRI data from the Kessler Foundation. Data from subjects who move their heads during imaging usually has to be discarded, as it is a source of random error. The article asserts that this often occurs with subjects who suffer from cognitively impairing disorders such as MS. Thus, there is a potential bias against subjects who are more impaired. Dr. Wylie, the associate director of Neuroscience in Neuropsychology & Neuroscience Research at the Kessler Foundation, said that as the difficulty of the task increased, there was an increase in movement that was larger among subjects with lower cognitive ability. A possible, albeit costly, solution is to recruit a larger number of subjects. This way, subjects of all different abilities are better represented.
I found this article to be very interesting, as it presents an important issue with fMRI that never occurred to me. Given that fMRI is becoming the standard in neuroimaging, it seems all that much more important that both researchers and consumers of research be aware of this issue. I am also curious if there are any other issues or biases that exist that are similar to this one. I can certainly think of several other diseases or disorders which would make it difficult for subjects to stay still during imaging.
As I was reading this article, I was reminded of another article by Alexis Madrigal, which highlights another important critique of fMRI. The article discusses the very real possibility of attaining false positives when using fMRI purely due to chance because of the large amount of data being collected by the machine. This assertion is supported by the fact that spots of activity appeared in the brain of a deceased salmon when it was put in the fMRI by neuroscientist Craig Bennett as a test subject. This surprising incident can be used to show the importance of sound statistical methods that take into account the possibility of false positives. Bennett highlights the importance of having the right amount of statistical power when conducting research.
Taken together, these two articles seem to suggest the importance of sound and careful attention to statistics when doing research in Cognitive Psychology. Since the discipline relies on neuroimaging as one of the methods of studying the brain, it seems particularly important to be aware of any pertinent issues that may exist with the imaging method that is being used in the given experiment. This could help to avoid unwanted complications in the experiment, such as confounds or low internal validity (which is very important in Cognitive Psychology). I would be interested to see if other methods of neuroimaging, such as CT scans and MRIs, have similar concerns associated with them. It seems as though it would be easier to avoid with other methods due to their more static nature. CT scan and MRIs both take images at a moment in time. On the other hand, an fMRI measures the amount of oxygen in the blood in a particular part of the brain.