Computational modelling clarifies that the over-generalization of learned causes has to do with distress from delusions broadly, whereas under-generalization of learned absences may have more to do with the experience of reference delusions (as is commensurate with the latter's limited scope).
2/3
