Structural connectivity of visuotopic intraparietal sulcus

Highlights

• This study combined functional mapping of visuotopic IPS regions with DTI.

• Posterior IPS regions are more likely connect to visual regions.

• Anterior IPS regions are more likely to connect to prefrontal regions.

• IPS1-4 show similar connectivity to insula, thalamus, and striatum.

Abstract

The intraparietal sulcus (IPS) contains topographically organized regions, similar to retinotopic maps in visual cortex. These regions, referred to as IPS1–4, show similar functional responses to the mapping tasks used to define them, yet differing responses to tests of other posterior parietal cortex (PPC) functions such as short-term memory, eye movements and object viewing, suggesting that they may have distinct patterns of structural connectivity to other parts of the brain. The present study combined functional magnetic resonance imaging (fMRI) mapping with diffusion tensor imaging (DTI) to describe white matter connections of visuotopic regions along the IPS, in 25 neurotypical young-adult participants. We found that posterior IPS more likely connects to retinotopically defined visual regions, and superior temporal gyrus, relative to anterior IPS. Anterior IPS regions 3 and 4 had higher connection probabilities to prefrontal regions, relative to posterior IPS. All four IPS regions showed inter-hemispheric connections to analogous regions in the opposite hemisphere, as well as consistent connections to the thalamus and regions of the striatum. Multivariate pattern classification at the group level reliably distinguished IPS regions from one another on the basis of connectivity patterns, especially for the most distal pairs of regions; occipital and prefrontal regions provided the most discriminating information. These findings advance our understanding of the structure of visuotopic IPS, with implications for functional differences between regions, and possible homologies between humans and macaques. Visuospatial functions dependent on the parietal cortex are frequently impaired in individuals with developmental disorders and those afflicted by cerebrovascular disease; the findings described here can be used as a basis for comparing connectivity differences in these populations.

Introduction

The intraparietal sulcus (IPS) contains several regions that show a topographic organization for visual–spatial information, similar to retinotopic maps in visual cortex (Silver and Kastner, 2009). While passive visual stimulation engages retinotopic responses in early visual areas (Sereno et al., 1995), responses in the IPS seem to be particularly engaged by spatial attention (Bressler and Silver, 2010, Saygin and Sereno, 2008) though see Swisher et al. (2007). Although these IPS regions respond in a similar manner to the tasks used to define them, such as visually guided saccades, or covertly attending to a rotating checkerboard wedge, the existence of multiple IPS regions suggests that they may have unique functions in other cognitive and sensory processes.

Several functional magnetic resonance imaging (fMRI) studies have probed the functionality of these regions using a subject-specific region of interest approach, wherein one task is used to map IPS sub-regions, and a follow-up task probes the involvement of these regions in classic parietal cortex functions. Such studies have demonstrated, for example, that more posterior IPS sub-regions have a greater increase in activation when visual short-term memory load is increased (Sheremata et al., 2010), and show greater object-specific responses, relative to anterior regions (Konen and Kastner, 2008b); for a detailed review see Silver and Kastner (2009).

To the extent that connectivity and function are inter-related, we would therefore expect that visuotopic IPS regions would show differing patterns of connectivity to the rest of the brain. A recent diffusion tensor-imaging (DTI) investigation suggested that posterior IPS sub-regions show greater connectivity to extra-striate visual regions than anterior sub-regions, and that this structural connectivity underlies the control of visual attention (Greenberg et al., 2012). This study did not, however, describe connections outside the visual cortex, delineate anterior IPS regions 3 or 4, or describe connections favored by these anterior IPS regions.

The posterior parietal cortex (PPC) is a functionally heterogeneous region, and several studies have used connectivity information to define sub-regions within the PPC. Rushworth et al. (2006) reported that, analogous to macaque MIP, AIP and LIP, the PPC can be divided into three sub-regions, a superior region showing strong connectivity to the superior colliculus, an anterior inferior parietal lobule (IPL) region showing preferential connections to ventral premotor cortex, and a posterior IPL region showing preferential connections to parahippocampal cortex. Cytoarchitectonic studies have identified subdivisions in the anterior IPS, labelled hIP1, 2, (Choi et al., 2006) and hIP3 (Scheperjans et al., 2008a, Scheperjans et al., 2008b); hIP3 is superior, hIP2 inferior-anterior and hIP1 inferior-posterior. A recent combined functional and structural connectivity study showed that hIP1 preferentially connects to the insula, hIP2 to prefrontal cortex (along the superior longitudinal fasciculus; SLF) and hIP3 to occipital cortex (Uddin et al., 2010). Several studies have used connectivity information to divide the inferior parietal lobule (Caspers et al., 2011, Wang et al., 2012), and relate sub-divisions to known structural layout in macaque (Caspers et al., 2011). Automated clustering of PPC voxels based on structural or functional connectivity has also been used to sub-divide the PPC into functional subdivisions in a data driven manner (Anderson et al., 2010, Mars et al., 2011, Nelson et al., 2010). For example, Mars et al. (2011) identified 10 functional sub-regions in superior and inferior parietal lobule and showed that anterior SPL clusters had stronger resting-state connectivity to dorsal pre-motor regions, while posterior IPS regions showed connectivity with extrastriate visual area V5.

Despite a growing body of literature on structural and functional connectivity of the PPC, the specific structural connections made by visuotopic IPS regions, and how these connections relate to functional differences between these regions remain largely unknown. In the present study, we combined functional mapping with DTI within subjects, to investigate differences in the patterns of connections made by IPS regions 1–4. Based on previous work, we hypothesized that posterior IPS regions would show preferential connectivity to occipital regions (Greenberg et al., 2012), and anterior IPS regions would show preferential connections to prefrontal regions (Mars et al., 2011, Uddin et al., 2010). We used multivariate pattern classification to uncover which structural connections most discriminated between IPS regions, at the group-level. A qualitative analysis was used to uncover the most consistent high-probability connections between IPS1 to 4 and other brain regions, across participants. Finally, as previous work has suggested hemispheric asymmetries in both function (Sack et al., 2005, Sheremata et al., 2010) and structural connectivity (Thiebaut de Schotten et al., 2011) of the PPC, we investigated hemispheric asymmetries in structural connectivity of IPS regions.