Can structure predict function at individual level in the human connectome?

Several studies predicting Functional Connectivity (FC) from Structural Connectivity (SC) at individual level have been published in recent years, each promising increased performance and utility. We investigated three of these studies, analyzing whether the results truly represent a meaningful individual-level mapping from SC to FC. Using data from the Human Connectome Project shared accross the three studies, we constructed a predictor by averaging FC of training data and analyzed its performance in the same way. In each case, we found that group average FC is an equivalent or better predictor of individual FC than the predictive models in terms of raw prediction performance. Furthermore, we showed that additional analyses performed by the authors of the three studies, in which they attempt to show that their predicted FC has value beyond raw prediction performance, could also be reproduced using the group average FC predictor. This makes it unclear whether any of the three methods represent a meaningful individual-level predictive model. We conclude that either the methods are not appropriate for the data, that the sample size is too small, or that the data does not contain sufficient information to learn a mapping from SC to FC. We advise future individual-level studies to explicitly report results in comparison to the performance of the group average, and carefully demonstrate that their predictions contain meaningful individual-level information. Finally, we believe that investigating alternatives for the construction of SC and FC may improve the chances of developing a meaningful individual-level mapping from SC to FC.

Working memory performance in glioma patients is associated with functional connectivity between the right dorsolateral prefrontal cortex and default mode network.

In healthy subjects, activity in the default mode network (DMN) and the frontoparietal network (FPN) has consistently been associated with working memory (WM). In particular, the dorsolateral prefrontal cortex (DLPFC) is important for WM. The functional-anatomical basis of WM impairment in glioma patients is, however, still poorly understood. We investigated whether WM performance of glioma patients is reflected in resting-state functional connectivity (FC) between the DMN and FPN, additionally focusing on the DLPFC. Resting-state functional MRI data were acquired from 45 glioma patients prior to surgery. WM performance was derived from a pre-operative N-back task. Scans were parcellated into ROIs using both the Gordon and Yeo atlas. FC was calculated as the average Pearson correlation between functional time series. The FC between right DLPFC and DMN was inversely related to WM performance for both the Gordon and Yeo atlas (p = .010). No association was found for FC between left DLPFC and DMN, nor between the whole FPN and DMN. The results are robust and not dependent on atlas choice or tumor location, as they hold for both the Gordon and Yeo atlases, and independently of location variables. Our findings show that WM performance of glioma patients can be quantified in terms of interactions between regions and large-scale networks that can be measured with resting-state fMRI. These group-based results are a necessary step toward development of biomarkers for clinical management of glioma patients, and provide additional evidence that global disruption of the DMN relates to cognitive impairment in glioma patients.

Presurgical Executive Functioning in Low-Grade Glioma Patients Cannot Be Topographically Mapped.

Executive dysfunctions have a high prevalence in low-grade glioma patients and may be the result of structural disconnections of particular subcortical tracts and/or networks. However, little research has focused on preoperative low-grade glioma patients. The frontotemporoparietal network has been closely linked to executive functions and is substantiated by the superior longitudinal fasciculus. The aim of this study was to investigate their role in executive functions in low-grade glioma patients. Patients from two neurological centers were included with IDH-mutated low-grade gliomas. The sets of preoperative predictors were (i) distance between the tumor and superior longitudinal fasciculus, (ii) structural integrity of the superior longitudinal fasciculus, (iii) overlap between tumor and cortical networks, and (iv) white matter disconnection of the same networks. Linear regression and random forest analyses were performed. The group of 156 patients demonstrated significantly lower performance than normative samples and had a higher prevalence of executive impairments. However, both regression and random forest analyses did not demonstrate significant results, meaning that neither structural, cortical network overlap, nor network disconnection predictors explained executive performance. Overall, our null results indicate that there is no straightforward topographical explanation of executive performance in low-grade glioma patients. We extensively discuss possible explanations, including plasticity-induced network-level equipotentiality. Finally, we stress the need for the development of novel methods to unveil the complex and interacting mechanisms that cause executive deficits in low-grade glioma patients.