A simulation and comparison of dynamic functional connectivity methods

Abstract

WHT and PF designed the simulation analysis. WHT and CGR designed and tested the jackknife correlation method. WHT and PPS designed the statistical models. WHT performed the analysis. WHT and PF wrote the paper. CGR and PPS critically revised the paper. AbstractThere is a current interest in quantifying brain dynamic functional connectivity (DFC) based on neuroimaging data such as fMRI. Many methods have been proposed, and are being applied, revealing new insight into the brain's dynamics. However, given that the ground truth for DFC in the brain is unknown, many concerns remain regarding the accuracy of proposed estimates. Since there exists many DFC methods it is difficult to assess differences in dynamic brain connectivity between studies. Here, we evaluate five different methods that together represent a wide spectrum of current approaches to estimating DFC (sliding window, tapered sliding window, temporal derivative, spatial distance and jackknife correlation). In particular, we were interested in each methods' ability to track changes in covariance over time, which is a key property in DFC analysis. We found that all tested methods correlated positively with each other, but there were large differences in the strength of the correlations between methods. To facilitate comparisons with future DFC methods, we propose that the described simulations can act as benchmark tests for evaluation of methods. In this paper, we present dfcbenchmarker, which is a Python package where researchers can easily submit and compare their own DFC methods to evaluate its performance.

Brain network segregation and integration during an epoch-related working memory fMRI experiment

“…changes of functional connectivity we used the jackknife correlation (JC) method. The JC method was first introduced in neuroscience for the purpose of measuring single trial coherence and Granger causality on ECoG data [52] and it has recently been applied to estimate dFC in fMRI data [22]. Since we are interested in estimating BOLD signal covariance at individual data time-points, the jackknife correlation between two time-series…”

“…The individual values of connectivity yielded by the JC method have little meaning in themselves, but, importantly, they become meaningful in relation to the rest of the timeseries of dFC values. Our choice of using the JC method was based on a recent simulation study that compared five different methods to derive point-based dFC estimates [22]. In that study we showed that the JC method had superior performance in terms of tracking changes in fluctuations in signal co-variance over time [22].…”

“…Our choice of using the JC method was based on a recent simulation study that compared five different methods to derive point-based dFC estimates [22]. In that study we showed that the JC method had superior performance in terms of tracking changes in fluctuations in signal co-variance over time [22].…”

“…However, sliding-window based methods have limited temporal resolution due to the necessity for analyzing large portions of consecutive data time-points (typically 30-60 seconds). Moreover, we have in a recent study shown that window-based methods fail to accurately track fluctuations in signal co-variance over time [22].…”

Characterizing functional and structural brain alterations driven by chronic alcohol drinking: a resting-state fMRI connectivity and voxel-based morphometry analysis

“…Several approaches have focused on studying the dynamic nature of whole brain activity in the resting state as measured by fMRI [185,186,[194][195][196]. Evaluation of these methods is typically done either using simulations or by correlating dynamic functional connectivity (dFC) metrics against non-brain measures like disease status [185,186,[194][195][196][197][198][199]. Sliding windows approaches have been popular to investigate time-varying patterns or FC [200,201].…”