In the last 25 years, functional MRIs have become an important part of neurological research, as physicians and other scientists use its techniques to look inside the brain. But what if long-held assumptions about what is standard and how to correct for outliers have been wrong?

According to Quartz, fMRI readers could have been misreading, and therefore misinterpreting, thousands of fMRI scans for more than two dozen years. That could mean the results of tens of thousands of neurological (and other) studies could be based on faulty information, which in turn could lead to ineffective or harmful treatment and false understandings of the workings of the human brain and body.

The scans rely on measuring blood flow in the brain. But if an fMRI of a dead salmon can show blood flow that obviously does not exist, what does that mean for how radiologists and others read the blood flow in other, not-dead fMRI subjects?

Check out Quartz to see the whole story of the dead fish in an MRI scanner and how it could have revealed questions with assumed false positive rates of the studies that depend on fMRI. Plus, find out why all these questions might not be problems at all.