Measures of selective attention processing like latent inhibition (LI) and previous termconditioned blockingnext term (CB) are disturbed in some patients with previous termschizophrenia.next term [LI is the delay in learning about the associations of previous termanext term stimulus that has been associated with no event (versus de novo learning); CB is the delay in learning the associations of previous termanext term stimulus-component when the other component has already started to acquire these associations.] We proposed: (1) to replicate the reported decreases of CB in patients without paranoid-hallucinatory previous termsymptoms;next term (2) to see if CB depends on the previous termage of illness-onset and its duration,next term as reported for LI. We studied 101 young and old, acute and chronically ill patients with previous termschizophrenia,next term of whom 62 learned previous termanext term modified ‘mouse-in-house’ CB task, and compared them with 62 healthy controls matched for previous termage,next term education and socio-economic background. CB was more evident in patients with previous termanext term diagnosis of paranoid previous termschizophrenianext term than other subtypes. An unusual persistence of high CB scores through testing was associated with productive previous termsymptomsnext term (including positive thought disorder). Reduced CB related to the expression of (previous terma)next term Schneider's first rank previous termsymptomsnext term of ideas-of-reference and (b) to negative previous termsymptomsnext term like poor rapport and poor attention. CB was less evident in the older patients and those with an earlier previous termillness-onset.next term In contrast to the similar LI test of selective attention, CB is found in patients with paranoid previous termschizophrenianext term and its expression is not related closely to previous termillness duration.next term This implies that the two tests reflect the activity of different underlying processes. We suggest that reduced CB on initial test-trials in nonparanoid previous termschizophrenianext term reflects the unusual persistence of controlled information processing strategies that would normally become automatic during conditioning. In contrast, continued CB during testing reflects an unusual persistence of automatic processing strategies.