A new study on the validity of fMRI results has found significant errors in the software that is standardly used in fMRI analysis. One of the bugs, which has since been fixed, had existed for 15 years, potentially making flawed much of the current thinking about how the brain works.
Anders Eklund from Linköping University in Sweden and his team gathered resting-state fMRI data from 499 healthy people sourced from databases around the world, split them up into groups of 20, and measured them against each other to get 3 million random comparisons. They tested the three most popular fMRI software packages for fMRI analysis and found false-positive rates of up to 70 percent (when 5% should be the expectation.)
The problem can exist because fMRI neuroscientists are not looking at the actual brain. Instead, they are looking at an image of the brain divided into tiny ‘voxels’ and then interpreted by a computer program. To the extents those programs have flaws, the patterns they reveal will have flaws.
A fuller discussion of the study and its implications can be found here and the article can be found here.
Comments are closed.