Havforskningsinstituttets institusjonelle arkiv
Brage IMR -
Institutional repository of the Institute of Marine Research b r a g e i m rDette er forfatters siste versjon av den fagfellevurderte artikkelen, vanligvis omtalt som postprint. I Brage IMR er denne artikkelen ikke publisert med forlagets layout fordi forlaget ikke tillater dette. Du finner lenke til forlagets versjon i Brage-posten. Det anbefales at referanser til artikkelen hentes fra forlagets side.
Ved lenking til artikkelen skal det lenkes til post i Brage IMR, ikke direkte til pdf-fil.This is the author's last version of the article after peer review and is not the publisher's version, usually referred to as postprint. You will find a link to the publisher's version in Brage IMR. It is recommended that you obtain the references from the publisher's site.Linking to the article should be to the Brage-record, not directly to the pdf-file.
AbstractA bite-and-pull demand-feeding system was introduced to groups of cultured cod (Gadus morhua). For half of the groups trigger actuations were rewarded with food, while actuations were unrewarded in the other groups. Initially, cod responded with frequent triggering, irrespective of whether triggering was rewarded with food or not. The high initial curiositydriven triggering rate declined rapidly, and was almost perfectly described by an exponential decay model with a decay rate of 7% min -1 . After 3 hours, the triggering frequency of the rewarded fish diverged from that of unrewarded fish, and it remained higher throughout the 9 days of the experiment. The initial curiosity-driven triggering allowed the cod to establish the relationship between action and reward in a short time. It is inferred that the time trajectory of action frequency of rewarded fish is the result of several factors and that operant learning can only be verified by comparing action frequencies of rewarded and unrewarded fish, and not by the temporal development in action frequency of rewarded fish alone.