作者: Fabian Munoz , Greg Jensen , Benjamin C. Kennedy , Yelda Alkan , Herbert S. Terrace
DOI: 10.1101/689133
关键词: Ordered set 、 Virtual workspace 、 Mathematics 、 Artificial intelligence 、 Stimulus (physiology) 、 Pattern recognition 、 Neural substrate 、 Posterior parietal cortex 、 Population 、 Associative property 、 Premovement neuronal activity
摘要: Monkeys are able to learn the implied ordering of pairs images drawn from an ordered set, without ever seeing all simultaneously and explicit spatial or temporal cues. The learning order differs visual motor sequences. We recorded activity parietal neurons in rhesus macaques while they learned 7-item TI lists when only 2 items were presented on each trial. Behavior ensemble neuronal significantly influenced by ordinal relationship stimulus pairs, specifically symbolic distance (the difference rank) joint ranks sum ranks). Symbolic strongly predicted decision accuracy, was consistently faster as increased. An effect rank performance also found nested within effect. Across population neurons, there significant modulation firing correlated with relative two stimuli Neurons exhibited selectivity for during learning, but not before after. observed behavior is best explained a virtual workspace model, associative reward mechanisms. neural data support role posterior cortex representing several variables that contribute serial particularly information about given Thus, appears belong substrate abstract relationships workspace.