Connectome spatial smoothing (CSS): Concepts, methods, and evaluation

CoCoNest: A continuous structural connectivity-based nested family of parcellations of the human cerebral cortex

Despite the widespread exploration and availability of parcellations for the functional connectome, parcellations designed for the structural connectome are comparatively limited. Current research suggests that there may be no single "correct" parcellation and that the human brain is intrinsically a multiresolution entity. In this work, we propose the Continuous Structural Connectivitity-based, Nested (CoCoNest) family of parcellations-a fully data-driven, multiresolution family of parcellations derived from structural connectome data. The CoCoNest family is created using agglomerative (bottom-up) clustering and error-complexity pruning, which strikes a balance between the complexity of each parcellation and how well it preserves patterns in vertex-level, high-resolution connectivity data. We draw on a comprehensive battery of internal and external evaluation metrics to show that the CoCoNest family is competitive with or outperforms widely used parcellations in the literature. Additionally, we show how the CoCoNest family can serve as an exploratory tool for researchers to investigate the multiresolution organization of the structural connectome.

Spurious correlations in surface-based functional brain imaging

The study of functional MRI data is increasingly performed after mapping from volumetric voxels to surface vertices. Processing pipelines commonly used to achieve this mapping produce meshes with uneven vertex spacing, with closer neighbours in sulci compared to gyri. Consequently, correlations between the fMRI time series of neighbouring sulcal vertices are stronger than expected. However, the causes, extent, and impacts of this bias are not well understood or widely appreciated. We explain the origins of these biases, and using in-silico models of fMRI data, illustrate how they lead to spurious results. The bias leads to leakage of anatomical cortical folding information into fMRI time series. We show that many common analyses can be affected by this "gyral bias", including test-retest reliability, fingerprinting, functional parcellations, regional homogeneity, and brain-behaviour associations. Finally, we provide recommendations to avoid or remedy this spatial bias.

TractoSCR: a novel supervised contrastive regression framework for prediction of neurocognitive measures using multi-site harmonized diffusion MRI tractography

Neuroimaging-based prediction of neurocognitive measures is valuable for studying how the brain's structure relates to cognitive function. However, the accuracy of prediction using popular linear regression models is relatively low. We propose a novel deep regression method, namely TractoSCR, that allows full supervision for contrastive learning in regression tasks using diffusion MRI tractography. TractoSCR performs supervised contrastive learning by using the absolute difference between continuous regression labels (i.e., neurocognitive scores) to determine positive and negative pairs. We apply TractoSCR to analyze a large-scale dataset including multi-site harmonized diffusion MRI and neurocognitive data from 8,735 participants in the Adolescent Brain Cognitive Development (ABCD) Study. We extract white matter microstructural measures using a fine parcellation of white matter tractography into fiber clusters. Using these measures, we predict three scores related to domains of higher-order cognition (general cognitive ability, executive function, and learning/memory). To identify important fiber clusters for prediction of these neurocognitive scores, we propose a permutation feature importance method for high-dimensional data. We find that TractoSCR obtains significantly higher accuracy of neurocognitive score prediction compared to other state-of-the-art methods. We find that the most predictive fiber clusters are predominantly located within the superficial white matter and projection tracts, particularly the superficial frontal white matter and striato-frontal connections. Overall, our results demonstrate the utility of contrastive representation learning methods for regression, and in particular for improving neuroimaging-based prediction of higher-order cognitive abilities. Our code will be available at: https://github.com/SlicerDMRI/TractoSCR.

Blurred streamlines: A novel representation to reduce redundancy in tractography

Tractography is a powerful tool to study brain connectivity in vivo, but it is well known to suffer from an intrinsic trade-off between sensitivity and specificity. A critical - but usually underrated - parameter to choose that can heavily impact the quality of the estimates is the number of streamlines to be reconstructed for a given data set. In fact, sensitivity can be improved by generating more and more streamlines, as all real anatomical connections are likely reconstructed, but lots of false positives are inevitably introduced, too. Consequently, so-called tractography filtering techniques have become increasingly popular to get rid of these false positives and improve specificity. However, increasing number of streamlines introduces redundancy in tractography reconstructions, which may negatively impact the performance of filtering algorithms, especially those based on linear formulations. To address this problem, we introduce a novel streamlines representation, called "blurred streamlines", which drastically reduces the redundancy among streamlines by (i) clustering similar trajectories and (ii) spatially blurring the corresponding signal contributions. We tested the effectiveness of the blurred streamlines both on synthetic and in vivo data. Our results clearly show that this new representation is as accurate as state-of-the-art methods despite using only 5% of the input streamlines, thus significantly decreasing the computational complexity of filtering algorithms as well as storage requirements of the resulting reconstructions.

Statistical power in network neuroscience

Network neuroscience has emerged as a leading method to study brain connectivity. The success of these investigations is dependent not only on approaches to accurately map connectivity but also on the ability to detect real effects in the data - that is, statistical power. We review the state of statistical power in the field and discuss sample size, effect size, measurement error, and network topology as key factors that influence the power of brain connectivity investigations. We use the term 'differential power' to describe how power can vary between nodes, edges, and graph metrics, leaving traces in both positive and negative connectome findings. We conclude with strategies for working with, rather than around, power in connectivity studies.