Linear Mixed-Effects Modeling Approach to FMRI Group Analysis.

Scientific and Statistical Computing Core, NIMH/NIH/HHS, USA. Electronic address: .
NeuroImage (Impact Factor: 6.13). 01/2013; DOI: 10.1016/j.neuroimage.2013.01.047
Source: PubMed

ABSTRACT Conventional group analysis is usually performed with Student-type t-test, regression, or standard AN(C)OVA in which the variance-covariance matrix is presumed to have a simple structure. Some correction approaches are adopted when assumptions about the covariance structure is violated. However, as experiments are designed with different degrees of sophistication, these traditional methods can become cumbersome, or even be unable to handle the situation at hand. For example, most current FMRI software packages have difficulty analyzing the following scenarios at group level: (1) taking within-subject variability into account when there are effect estimates from multiple runs or sessions; (2) continuous explanatory variables (covariates) modeling in the presence of a within-subject (repeated measures) factor, multiple subject-grouping (between-subjects) factors, or the mixture of both; (3) subject-specific adjustments in covariate modeling; (4) group analysis with estimation of hemodynamic response (HDR) function by multiple basis functions; (5) various cases of missing data in longitudinal studies; and (6) group studies involving family members or twins. Here we present a linear mixed-effects modeling (LME) methodology that extends the conventional group analysis approach to analyze many complicated cases, including the six prototypes delineated above, whose analysis would be otherwise either difficult or unfeasible under traditional frameworks such as AN(C)OVA and general linear model (GLM). In addition, the strength of the LME framework lies in its flexibility to model and estimate the variance-covariance structures for both random effects and residuals. The intraclass correlation (ICC) values can be easily obtained with an LME model with crossed random effects, even at the presence of confounding fixed effects. The simulations of one prototypical scenario indicate that the LME modeling keeps a balance between the control for false positives and the sensitivity for activation detection. The importance of hypothesis formulation is also illustrated in the simulations. Comparisons with alternative group analysis approaches and the limitations of LME are discussed in details.

  • [Show abstract] [Hide abstract]
    ABSTRACT: There are imminent needs for longitudinal analysis to make physiological inferences on NIH MRI study of normal brain development. But up to date, two critical aspects for longitudinal analysis, namely the selections of mean and covariance structures have not been addressed by the neuroimaging community. For the mean structure, we employed a linear free-knot B-spline regression in combination with quasi-least square estimating equations to approximate a nonlinear growth trajectory with piecewise linear segments for a friendly physiological interpretation. For covariance structure selection, we have proposed a novel time varying correlation structure considering not only the time separation between the repeated measures but also when these acquisitions occurred. We have demonstrated that the proposed covariance structure has a lower Akaike information criterion value than the commonly used Markov correlation structure.
    05/2014; 2014:1206-1209.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Uncertainty is a ubiquitous feature of our daily lives. Although previous studies have identified a number of neural and peripheral physiological changes associated with uncertainty, there are limited data on the causal mechanisms underlying these responses in humans. In this study, we address this empirical gap through a novel application of fMRI in neurosurgical patients with focal, bilateral ventromedial prefrontal cortex (vmPFC) damage. The fMRI task involved cued anticipation of aversive and neutral picture stimuli; “certain” cues unambiguously indicated the upcoming picture valence, whereas “ambiguous” cues could precede either picture type. Healthy subjects exhibited robust bilateral insula responses to ambiguous cues, and this cue-related insula activity significantly correlated with heart rate variability during the task. By contrast, the vmPFC lesion patients exhibited altered cue-related insula activity and reduced heart rate variability. These findings suggest a role for vmPFC in coordinating neural and physiological responses during anticipation.
    Journal of Neuroscience 07/2014; 34(31):10394-10401. · 6.75 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Degree, which is the number of connections incident upon a node, measures the relative importance of the node within a network. By computing degree metrics in voxel-wise functional brain networks, many studies performed high-resolution mapping of brain network hubs using resting-state functional magnetic resonance imaging. Despite its extensive applications, defining nodes as voxels without considering the different sizes of brain regions may result in a network where the degree cannot accurately represent the importance of nodes. In this study, we designed a data-driven method to reduce this impact of the region size in degree metrics by (1) disregarding all self-connections among voxels within the same region and (2) regulating connections from voxels of other regions by the sizes of those regions. The modified method that we proposed allowed direct evaluation of the impact of the region size, showing that traditional degree metrics overestimated the degree of previous identified hubs in humans, including the visual cortex, precuneus/posterior cingulate cortex, and posterior parietal cortex, and underestimated the degree of regions including the insular cortex, anterior cingulate cortex, parahippocampus, sensory and motor cortex, and supplementary motor area. However, the locations of prominent hubs were stable even after correcting the impact. These findings were robust under different connectivity thresholds, degree metrics, data-preprocessing procedures, and datasets. In addition, our modified method improved test-retest reliability of degree metrics as well as the sensitivity in group-statistic comparisons. As a promising new tool, our method may reveal network properties that better represent true brain architecture without compromising its data-driven advantage.
    Frontiers in Neurology 10/2014; 5:199.