Purchase this article with an account.
Elisabeth A. Karuza, Lauren L. Emberson, Matthew E. Roser, Michael S. Gazzaniga, Daniel Cole, Richard N. Aslin, Jozsef Fiser; Dynamic shifts in connectivity between frontal, occipital, hippocampal and striatal regions characterize statistical learning of spatial patterns. Journal of Vision 2014;14(10):955. doi: 10.1167/14.10.955.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Extensive behavioral evidence has revealed that humans automatically develop internal representations that are adapted to the temporal and spatial statistics of the environment. However, the neural systems underlying this statistical learning process are not fully understood. Recently, various neuroimaging methods have been employed to examine this topic, but these studies have focused exclusively on temporally ordered stimuli. Since spatial structure is a hallmark of object and scene perception in vision, the present functional magnetic resonance imaging (fMRI) study investigated the substrates and processes underlying complex spatial pattern learning. Neuroimaging data were obtained while 20 subjects passively viewed artificially created scenes with a pre-specified pair-based statistical structure. After three runs of exposure to 144 different 6-element scenes, subjects performed a yes/no task on base-pairs and cross-pairs. Using seed regions defined by relating magnitude of activation to this post-exposure behavioral learning performance, we examined changes in functional connectivity over the course of learning. In addition to a general increase in connectivity throughout exposure, we find a specific connectivity relationship between frontal, occipital, hippocampal and subcortical areas that was dynamically reconfigured as learning progressed. Specifically, we show that connectivity with frontal regions shifted from early visual areas to subcortical areas when comparing early and late phases of exposure. These results suggest that learning is not fully captured by a single, fixed "learning" network, but is reflected at least partially in dynamic shifts in connectivity across numerous cortical and subcortical areas.
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only