Premium
Hippocampal lesion and transitive inference: Dissociation of inference‐based and reinforcement‐based strategies in pigeons
Author(s) -
Lazareva Olga F.,
Kandray Kaitlyn,
Acerbo Martin J.
Publication year - 2015
Publication title -
hippocampus
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.767
H-Index - 155
eISSN - 1098-1063
pISSN - 1050-9631
DOI - 10.1002/hipo.22366
Subject(s) - associative property , reinforcement , stimulus (psychology) , psychology , hippocampal formation , lesion , transitive relation , neuroscience , recall , dissociation (chemistry) , cognitive psychology , inference , artificial intelligence , computer science , social psychology , chemistry , mathematics , combinatorics , psychiatry , pure mathematics
A typical nonverbal transitive inference task (TI) consists of several overlapping discriminations (A+ B−, B+ C−, C+ D−, D+ E−, where letters indicate stimuli and pluses and minuses denote reinforcement and nonreinforcement). A choice of stimulus B in a novel pair BD is interpreted as indicative of a TI (if B > C and C > D, then B > D). Although hippocampus has been implicated in nonverbal TI, it is not clear whether it simply maintains memory of associative values or stores an ordered representation of stimuli. We investigated the effect of hippocampal lesion on TI in pigeons while controlling reinforcement history so that reliance on associative values would lead to a choice of a stimulus D in the pair BD instead of a choice of a stimulus B expected by inferential mechanisms. Prior to the lesion, some of the pigeons (relational group; n = 4) have selected B over D indicating TI, while other birds (associative group; n = 6) chose D over B suggesting reliance on associative values. Hippocampal lesion had no effect on postlesion performance in associative group. In contrast, the relational group that preferred stimulus B in a pair BD before lesion showed a near‐chance performance after the lesion. Our results demonstrate that hippocampus may be involved in creating a representation of an ordered series of the stimuli instead of maintaining reinforcement history of each stimulus. In addition, we provide a behavioral procedure suitable for dissociating different behavioral strategies used to solving TI task. Finally, we show for the first time the involvement of avian hippocampus in the task that is not explicitly spatial in nature. These results further confirm the notion that avian hippocampus is functionally analogous to mammalian hippocampus despite the significant differences in their anatomy. © 2014 Wiley Periodicals, Inc.