I have only read the text diagonally while looking for some keywords, but my 2 cents: in many cases, when a neuroscience study claims to be able to "predict" so

I have only read the text diagonally while looking for some keywords, but my 2 cents: in many cases, when a neuroscience study claims to be able to "predict" something, they mean "var X and var Y show a strong correlation". Actual predictive power is not checked unless there is a real test set that is left out of the training phase and all that (cross-validation, etc). When this article claims: > When data from GAD patients and healthy controls were combined, relatively low CHO predicted both relatively higher IQ and worry scores. ... it is actually saying that CHO is inversely correlated with higher IQ and worry scores. (Unless, again, as noted above, they are cross-validating in a way that was not obvious to me when I checked for it.)