Real-Time Resting-State Functional Magnetic Resonance Imaging Using Averaged Sliding Windows with Partial Correlations and Regression of Confounding Signals

Background/Introduction: There is considerable interest in using real-time functional magnetic resonance imaging (fMRI) for monitoring functional connectivity dynamics. To date, the majority of real-time resting-state fMRI studies have examined a limited number of brain regions. This is, in part, due to the computational demands of traditional seed- and independent component analysis-based methods, in particular when using increasingly available high-speed fMRI methods. Methods: This study describes a computationally efficient, real-time, seed-based, resting-state fMRI analysis pipeline using moving averaged sliding-windows (ASW) with partial correlations and regression of motion parameters and signals from white matter and cerebrospinal fluid. Results: Analytical and numerical analyses of ASW correlation and sliding-window regression as a function of window width show selectable bandpass filter characteristics and effective suppression of artifactual correlations resulting from signal drifts and transients. The analysis pipeline is compatible with multislab echo-volumar imaging and simultaneous multislice echo-planar imaging with repetition times as short as 136 msec. High-speed, resting-state fMRI data in healthy controls demonstrate the effectiveness of this approach for minimizing artifactual correlations in white and gray matter, which was comparable to conventional regression across the entire scan. Integrating sliding-window averaging (width: W₁) within a second-level sliding-window (width: W₂) enabled monitoring of intra- and internetwork correlation dynamics of up to 12 resting-state networks with bandpass filter characteristics determined by the first-level sliding-window and temporal resolution W₁ + W₂. Conclusions: The computational performance and confound tolerance make this seed-based, resting-state fMRI approach suitable for real-time monitoring of data quality and resting-state connectivity dynamics in neuroscience and clinical research studies. Impact statement Using averaged sliding-windows for seed-based correlation and regression of confounding signals provides a powerful model-free approach to increase tolerance to artifactual signal transients in resting-state analysis. The algorithmic efficiency of this sliding-window approach enables real-time, seed-based, resting-state functional magnetic resonance imaging (fMRI) of multiple networks with computation of connectivity matrices and online monitoring of data quality. Integration of a second-level sliding-window enables mapping of resting-state connectivity dynamics. Sensitivity and tolerance to confounding signals compare favorably with conventional correlation and confound regression across the entire scan. This methodological advance has the potential to enhance the clinical utility of resting-state fMRI and facilitate neuroscience applications.

Functional Connectivity Alterations between Networks and Associations with Infant Immune Health within Networks in HIV Infected Children on Early Treatment: A Study at 7 Years

Although HIV has been shown to impact brain connectivity in adults and youth, it is not yet known to what extent long-term early antiretroviral therapy (ART) may alter these effects, especially during rapid brain development in early childhood. Using both independent component analysis (ICA) and seed-based correlation analysis (SCA), we examine the effects of HIV infection in conjunction with early ART on resting state functional connectivity (FC) in 7 year old children. HIV infected (HIV+) children were from the Children with HIV Early Antiretroviral Therapy (CHER) trial and all initiated ART before 18 months; uninfected children were recruited from an interlinking vaccine trial. To better understand the effects of current and early immune health on the developing brain, we also investigated among HIV+ children the association of FC at 7 years with CD4 count and CD4%, both in infancy (6–8 weeks) and at scan. Although we found no differences within any ICA-generated resting state networks (RSNs) between HIV+ and uninfected children (27 HIV+, 18 uninfected), whole brain connectivity to seeds located at RSN connectivity peaks revealed several loci of FC differences, predominantly from seeds in midline regions (posterior cingulate cortex, paracentral lobule, cuneus, and anterior cingulate). Reduced long-range connectivity and increased short-range connectivity suggest developmental delay. Within the HIV+ children, clinical measures at age 7 years were not associated with FC values in any of the RSNs; however, poor immune health during infancy was associated with localized FC increases in the somatosensory, salience and basal ganglia networks. Together these findings suggest that HIV may affect brain development from its earliest stages and persist into childhood, despite early ART.

Functional connectivity analyses using emulated and conventional resting-state data: parts versus the whole story

Continuous resting-state (RS) functional magnetic resonance imaging (fMRI) has become particularly useful to identify changes in functional connectivity (FC) in CNS disorders. Fair et al. proposed a method of volume extraction to emulate RS fMRI from block-design experiments. Whether the validity of this approach holds true in multiple sclerosis (MS) patients has not been tested formally so far. Twelve MS patients and 18 controls underwent conventional RS fMRI and a cognitive block-design fMRI. The total amount of volumes as well as the truncated set of volumes of both functional datasets was separately analyzed using a seed-based approach. Overall, seed-based analyses of FC from the anterior cingulated cortex allowed identification of the same key-network constituents using different analytical approaches, whereas higher-level within-group analyses of emulated RS versus continuous RS also revealed significant distinct differences in FC networks. Using the emulated RS approach, a general identification of connectivity networks similar to those obtained using conventional RS data also appears feasible in diseased brains. Higher-level contrasts, however, yielded different results attesting to a significant impact of employed methodology.