Nuisance effects and the limitations of nuisance regression in dynamic functional connectivity fMRI

Abstract

In resting-state fMRI, dynamic functional connectivity (DFC) measures are used to characterize temporal changes in the brain's intrinsic functional connectivity. A widely used approach for DFC estimation is the computation of the sliding window correlation between blood oxygenation level dependent (BOLD) signals from different brain regions. Although the source of temporal fluctuations in DFC estimates remains largely unknown, there is growing evidence that they may reflect dynamic shifts between functional brain networks. At the same time, recent findings suggest that DFC estimates might be prone to the influence of nuisance factors such as the physiological modulation of the BOLD signal. Therefore, nuisance regression is used in many DFC studies to regress out the effects of nuisance terms prior to the computation of DFC estimates. In this work we examined the relationship between seed-specific sliding window correlation-based DFC estimates and nuisance factors. We found that DFC estimates were significantly correlated with temporal fluctuations in the magnitude (norm) of various nuisance regressors. Strong correlations between the DFC estimates and nuisance regressor norms were found even when the underlying correlations between the nuisance and fMRI time courses were relatively small. We then show that nuisance regression does not necessarily eliminate the relationship between DFC estimates and nuisance norms, with significant correlations observed between the DFC estimates and nuisance norms even after nuisance regression. We present theoretical bounds on the difference between DFC estimates obtained before and after nuisance regression and relate these bounds to limitations in the efficacy of nuisance regression with regards to DFC estimates.

Nuisance effects in inter-scan functional connectivity estimates before and after nuisance regression

“…We provide a more extensive analysis of various nuisance effects on the FC estimates before and after NR. We generalize the theory developed in Nalci et al (2019) to static FC measures and confirm the validity of the theoretical limitation of nuisance regression for correlation-based static FC estimates. We introduce nuisance contamination maps which illustrate the spatial distribution and extent of correlations between nuisance norms and FC estimates across scans.…”

“…Our work significantly extends the preliminary results regarding static FC estimates presented in Nalci et al (2019). We provide a more extensive analysis of various nuisance effects on the FC estimates before and after NR.…”

“…More recently, in Nalci et al (2019), we demonstrated that dynamic FC (DFC) estimates were related (within a scan) to the norms of various nuisance regressors such as the HM, WM+CSF, GS, heart rate, and respiratory derived time courses. We found that NR was largely ineffective in removing nuisance effects from DFC estimates with significant relations between the nuisance norms and DFC estimates remaining even after NR.…”

“…Despite the fact that NR is adopted with the assumption that it removes nuisance confounds from the FC estimates, it has been previously shown that it can be quiet ineffective in reducing the effects of nuisance confounds (Bright and Murphy, 2015; Power et al, 2012; Nalci et al, 2019). For example, HM regression has been shown to be a largely ineffective approach for reducing HM confounds in FC estimates even after projecting out 12 motion regressors (Power et al, 2012; Satterthwaite et al, 2012; Van Dijk et al, 2012).…”

Transient states of network connectivity are atypical in autism: A dynamic functional connectivity study

“…However, dFC methods are relatively new, and gold standards have yet to be agreed upon with respect to sliding window parameters such as window length and overlap. Indeed, some have raised doubts about the reliability of sliding window methods, and the degree to which resulting states represent true FC configurations (Laumann et al, 2017;Nalci, Rao, & Liu, 2019;Shakil, Lee, & Keilholz, 2016). For example, in heterogeneous samples, differences in dFC measures may arise from nonneuronal sources of variability (Lehmann, White, Henson, Cam, & Geerligs, 2017).…”