Resting-state functional connectivity of ventral parietal regions associated with attention reorienting and episodic recollection

S. M. Daselaar, W. Huijbers, K. Eklund, M. Moscovitch, R. Cabeza

In functional neuroimaging studies, ventral parietal cortex (VPC) is recruited by very different cognitive tasks. Explaining the contributions of VPC to these tasks has become a topic of intense study and lively debate. Perception studies frequently find VPC activations during tasks involving attention-reorienting, and memory studies frequently find them during tasks involving episodic recollection. According to the Attention to Memory (AtoM) model, both phenomena can be explained by the same VPC function: bottom-up attention. Yet, a recent functional MRI (fMRI) meta-analysis suggested that attention-reorienting activations are more frequent in anterior VPC, whereas recollection activations are more frequent in posterior VPC. Also, there is evidence that anterior and posterior VPC regions have different functional connectivity patterns. To investigate these issues, we conducted a resting-state functional connectivity analysis using as seeds the center-of-mass of attention-reorienting and recollection activations in the meta-analysis, which were located in the supramarginal gyrus (SMG, around the temporo-parietal junction-TPJ) and in the angular gyrus (AG), respectively. The SMG seed showed stronger connectivity with ventrolateral prefrontal cortex (VLPFC) and occipito-temporal cortex, whereas the AG seed showed stronger connectivity with the hippocampus and default network regions. To investigate whether these connectivity differences were graded or sharp, VLPFC and hippocampal connectivity was measured in VPC regions traversing through the SMG and AG seeds. The results showed a graded pattern: VLPFC connectivity gradually decreases from SMG to AG, whereas hippocampal connectivity gradually increases from SMG to AG. Importantly, both gradients showed an abrupt break when extended beyond VPC borders. This finding suggests that functional differences between SMG and AG are more subtle than previously thought. These connectivity differences can be explained by differences in the input and output to anterior and posterior VPC regions, without the need of postulating markedly different functions. These results are as consistent with integrative accounts of VPC function, such as the AtoM model, as they are with models that ascribe completely different functions to VPC regions.