29-30 January, 2020 - Szeged, Hungary


Abstract details



Balázs Knakker1, Anna Padányi1, Viktória Pál1, Kata Kóczán1, Evelin Kiefer1, Balázs Lendvai1, István Hernádi1

1 Translational Neuroscience Research Group, Grastyán Endre Translational Research Centre, Szentágothai Research Centre, University of Pécs

Keeping known objects in visual short-term memory (VSTM) is thought to entail activation of long-term memory representations, while VSTM for novel objects might rely on sensory-attentional memory processes to a greater degree. The CANTAB Paired Associates Learning (PAL) paradigm is a VSTM task where the locations of sequentially presented abstract visual objects have to be held in memory during a delay period, after which all objects are probed in random order. Twelve male adult rhesus macaques have been intensively trained in this task on every working day for 22 months. We hypothesized that the high level of performance that the animals achieved could be permitted by deep knowledge of the 67 visual objects used throughout the training period. To test this hypothesis, we introduced completely new objects that the animals had never seen before on 5 sessions during a 4-week period. We show that while initially the proportion of correct choices dropped by an average of 25%, after 5 days the animals could memorize the location of novel objects as efficiently as old objects. In new-set sessions, however, trials with old objects had 5-10% better performance, which could be due to lower inter-trial interference or heightened motivation for old objects. Memory performance also increased when new and old objects were mixed within trials, which could demonstrate the facilitatory effect of new/old status as a salient memory cue. We conclude that visual object knowledge is beneficial, but is not a prerequisite for macaques to efficiently perform the PAL object-location VSTM task.