Pareidolia

I love words. Radiology has reminded me of one of my favorite words, pareidolia. A type of apophenia, it refers to that common habit of seeing, hearing, or perceiving something random and interpreting it as significant. How many times have you looked at a weird cloud or listened to a song backwards and found meaning where it was not intended? Even the image shown above, taken by a NASA Viking mission to Mars in 1976 jumps out as a face, but it turns out to be a trick of the sunlight on a random rock formation.

Human psychology is wired to identify patterns, faces, sequences, and shapes even in random data. We are biased to find things, and it's more pronounced when we're looking for something in particular. We see what we expect to see. Thus, the best clinical studies are blinded. If you are conducting a study to see whether a drug does better than a placebo, if you know which patients are taking the drug, you are more likely to get a positive result. Subjects and interpreters should be blinded to what arm of the study patients are assigned to. This takes away bias.

We must keep this in mind in radiology. I am much more likely to call a chest X-ray normal if it's in a reading room rather than a lecture because I know and expect that images shown in lecture will be abnormal. I also think radiologists should know the clinical information regarding the studies they interpret. Sometimes clinicians will hold back on giving the history because they don't want to bias the radiologist. This is silly. Giving a radiologist more information only changes their pre-test probability. If they expect something to be wrong, they are more likely to identify it, especially if it is subtle. The good radiologist will then hit all the other boxes in the checklist to make sure nothing else is wrong. We see what we want to see. If radiologists know what the clinician expects, he changes his pre-test probability and is more likely to make the right diagnosis.

Image is in the public domain, from Wikipedia.