Hemispheric Asymmetries of Individual Differences in Functional Connectivity

Abstract

Resting-state fMRI studies have revealed that individuals exhibit stable, functionally meaningful divergences in large-scale network organization. The locations with strongest deviations (called network “variants”) have a characteristic spatial distribution, with qualitative evidence from prior reports suggesting that this distribution differs across hemispheres. Hemispheric asymmetries can inform us on constraints guiding the development of these idiosyncratic regions. Here, we used data from the Human Connectome Project to systematically investigate hemispheric differences in network variants. Variants were significantly larger in the right hemisphere, particularly along the frontal operculum and medial frontal cortex. Variants in the left hemisphere appeared most commonly around the TPJ. We investigated how variant asymmetries vary by functional network and how they compare with typical network distributions. For some networks, variants seemingly increase group-average network asymmetries (e.g., the group-average language network is slightly bigger in the left hemisphere and variants also appeared more frequently in that hemisphere). For other networks, variants counter the group-average network asymmetries (e.g., the default mode network is slightly bigger in the left hemisphere, but variants were more frequent in the right hemisphere). Intriguingly, left- and right-handers differed in their network variant asymmetries for the cingulo-opercular and frontoparietal networks, suggesting that variant asymmetries are connected to lateralized traits. These findings demonstrate that idiosyncratic aspects of brain organization differ systematically across the hemispheres. We discuss how these asymmetries in brain organization may inform us on developmental constraints of network variants and how they may relate to functions differentially linked to the two hemispheres.

Evidence for a Compensatory Relationship between Left- and Right-Lateralized Brain Networks

“…Furthermore, we examined potential relationships between networks’ NSAR values and found evidence supporting a dependent hypothesis of lateralization. These findings shed new light on hemispheric specialization, which has implications for the understanding of brain organization and development (Toga & Thompson, 2003), individual differences (Perez et al, 2023), human-defining cognitive processes (Hartwigsen et al, 2021), and neurodevelopmental conditions (Eyler et al, 2012; Kong et al, 2022). Previously, functional lateralization has been assessed through a variety of approaches dependent on intrinsic connectivity, primarily at the group level.…”

“…This was the case for the Dorsal Attention-B network, which was not a significantly lateralized network across the three datasets. However, there is evidence for a left-lateralized dorsal attention network across both left-and right-handed individuals, stemming from a within-individual network variants approach (see Figure 7 Panel C of Perez et al, 2023). Additionally, it could be that a finer-grained parcellation deconstructs the dorsal attention network into one bilateral and one lateralized network, similar to previous within-individual work on the default network (Braga & Buckner, 2017; DiNicola et al, 2020).…”

“…These historical methods are complimented by neuroimaging metrics, many of which are functional connectivity-based. Of particular interest are the intrinsic laterality index (Liu et al, 2009), autonomy index (Wang et al, 2014), hemispheric contrast (Gotts et al, 2013), functional lateralization metric (Nielsen et al, 2013), classification metric (Friedrich et al, 2022), and network variants approach (Perez et al, 2023). Despite the unifying aim of estimating hemispheric specialization or lateralization, each of the listed methods varies in terms of how it approaches structural asymmetries, the addition of covariates such as handedness and gender, and short-and long-range connectivity.…”

“…Despite the unifying aim of estimating hemispheric specialization or lateralization, each of the listed methods varies in terms of how it approaches structural asymmetries, the addition of covariates such as handedness and gender, and short-and long-range connectivity. However, with the exception of the network variants approach (Perez et al, 2023), each method has been implemented on less than 12 minutes of resting-state fMRI data per participant, a tactic which is increasingly being exchanged for a within-individual “precision” approach.…”

A frontoparietal network for volitional control of gaze following

“…In our analysis, we were only able to detect activity in left prefrontal areas. Several studies have reported on asymmetries within the FP network in the sense that left and right FP networks have different patterns in timing and are related to distinct processes in cognitive control (Budisavljevic et al, 2017; Ogawa et al, 2022; Perez et al, 2022). There is a possibility that activation in the left FP network serves distinct functions such as rule and stimulus processing, target evaluation, and the decision making process itself rather than response evaluation and adjustment, which might be performed by the right FP network (Gold & Shadlen, 2007; Gratton et al, 2017).…”