Predictive processing is one of the most influential ideas from computational neuroscience for the experimental neurosciences. However, definitions of predictive processing vary broadly, to the extent that “predictive coding” is used sometimes in a very narrow sense (there are specific cell types for negative or positive expectation errors) or in a very broad sense (anything related to error signals or expectation mismatch is predictive processing).

Jerome Lecoq now started the great initiative to write a review about error signals for predictive processing, but in a very collaborative manner. He invited anybody interested to join the writing of the review.

I think that this way of writing a collaborative and open review is a great idea, even though it might be difficult to reconcile all the different opinions! This link will lead you to the Google document with the main text and with the instructions on how to contribute. And if you don’t feel like you want to contribute, it is at least a useful opportunity to learn about the current opinions in the field and how people agree or disagree about the interpretation of predictive coding, error signals and the literature that covers both. Take a look!