Abstract:Standard fMRI connectivity analyses depend on aggregating the time series of individual voxels within regions of interest (ROIs). In certain cases, this spatial aggregation implies a loss of valuable functional and anatomical information about smaller subsets of voxels that drive the ROI level connectivity. We use two recently published graphical search methods to identify subsets of voxels that are highly responsible for the connectivity between larger ROIs. To illustrate the procedure, we apply both methods to longitudinal high-resolution resting state fMRI data from regions in the medial temporal lobe from a single individual. Both methods recovered similar subsets of voxels within larger ROIs of entorhinal cortex and hippocampus subfields that also show spatial consistency across different scanning sessions and across hemispheres. In contrast to standard functional connectivity methods, both algorithms applied here are robust against false positive connections produced by common causes and indirect paths (in contrast to Pearson's correlation) and common effect conditioning (in contrast to partial correlation based approaches). These algorithms allow for identification of subregions of voxels driving the connectivity between regions of interest, recovering valuable anatomical and functional information that is lost when ROIs are aggregated. Both methods are specially suited for voxelwise connectivity research, given their running times and scalability to big data problems.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to recover the functional and anatomical information lost in the standard functional connectivity methods by identifying sub - regions of effective connectivity in functional magnetic resonance imaging (fMRI) analysis. Specifically, the authors used two recently published graph - search methods (FASK and Two - Step algorithm) to identify sub - collective voxels that are highly responsible for the connections between these regions from larger regions of interest (ROIs). This helps to more precisely understand the communication mechanisms between different regions of the brain, especially in resting - state fMRI data. ### Main problems 1. **Limitations of standard fMRI connectivity analysis**: - Standard fMRI connectivity analysis relies on aggregating the time series of individual voxels within the regions of interest (ROIs). This spatial aggregation can lead to the loss of valuable functional and anatomical information, especially information about small voxel subsets that drive ROI - level connections. 2. **Identifying connected sub - regions**: - The authors used two graph - search methods (FASK and Two - Step algorithm) to identify voxel subsets that drive the connections between larger ROIs. These methods can identify voxel subsets that are spatially consistent across different scanning sessions and different hemispheres. 3. **Improving the robustness of connectivity analysis**: - Compared with the traditional Pearson correlation and partial correlation methods, these two algorithms are more robust to false positive connections generated by common causes and indirect paths. They can also avoid false connections due to common - effect conditioning. ### Methods 1. **FASK algorithm**: - The FASK algorithm first estimates which time - series voxels are adjacent in the graph representation, but does not specify the direction of influence. It uses the first stage of the PC - stable algorithm (called FAS - stable) to avoid false - positive connections through a series of iterative conditional independence evaluations. 2. **Two - Step algorithm**: - The Two - Step algorithm estimates the connection matrix in the linear causal structure by the independent component analysis (ICA) method. It is divided into two steps: the first step infers the undirected adjacency relationships between the observed variables; the second step optimizes the non - zero entries to make the estimated components as independent or non - Gaussian as possible and imposes a sparsity constraint. ### Data - Longitudinal resting - state fMRI data of a single individual was used, from the studies of Poldrack et al. (2015) and Laumann et al. (2015). The data contains multiple scanning sessions, with 4,610 data points per session, covering four ROIs: the entorhinal cortex (ENT), the hippocampal sub - region (CA32DG), hippocampal area 1 (CA1) and the subiculum (SUB). ### Results - The two algorithms showed good consistency in different data sets, especially in terms of the Jaccard similarity index of the undirected graph. For specific ROI pairs, such as (ENT, CA32DG) and (ENT, CA1), the consistency of the connected sub - regions was lower, which may reflect the sparsity of the functional connections or the actual changes in the connections of these regions in the resting state. ### Conclusions - By identifying sub - regions of effective connectivity, this study provides a more refined method for understanding the communication mechanisms between different regions of the brain and helps to recover important information lost in standard functional connectivity analysis.

Identification of Effective Connectivity Subregions

A Deconvolution-Based Approach to Identifying Large-Scale Effective Connectivity.

Concurrent brain parcellation and connectivity estimation via co‐clustering of resting state fMRI data: A novel approach

Identifying effective connectivity parameters in simulated fMRI: a direct comparison of switching linear dynamic system, stochastic dynamic causal, and multivariate autoregressive models

Assessing the Influence of Different ROI Selection Strategies on Functional Connectivity Analyses of fMRI Data Acquired During Steady-State Conditions

A Weighted-Rv Method to Detect Fine-Scale Functional Connectivity During Resting State.

A Mixed Model Approach for Estimating Regional Functional Connectivity from Voxel-level BOLD Signals

Large-scale correlation screening under dependence for brain functional connectivity network inference

Consistency of Regions of Interest as nodes of functional brain networks measured by fMRI

Connectivity analyses for task-based fMRI

A personalized cortical atlas for functional regions of interest

Decomposition of individual-specific and individual-shared components from resting-state functional connectivity using a multi-task machine learning method

Group-representative functional network estimation from multi-subject fMRI data via MRF-based image segmentation

Estimating effective connectivity in linear brain network models

Performing Group-Level Functional Image Analyses Based on Homologous Functional Regions Mapped in Individuals.

Reliability and reproducibility of individual differences in functional connectivity acquired during task and resting state

Hierarchical Extraction of Functional Connectivity Components in Human Brain Using Resting-State fMRI

Investigating the sub-regions of the superior parietal cortex using functional magnetic resonance imaging connectivity

Subspace-based Identification Algorithm for characterizing causal networks in resting brain

Exploring Functional Connectivity Networks in fMRI Data Using Clustering Analysis.

A Network Analysis Approach to fMRI Condition-Specific Functional Connectivity