Linear Mixed-Effects Modeling Approach to FMRI Group Analysis.

Scientific and Statistical Computing Core, NIMH/NIH/HHS, USA. Electronic address: .
NeuroImage (Impact Factor: 6.36). 01/2013; 73. 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.

  • Source
    • "doc/manual/3dDeconvolve). For the group-level analyses, a linear model effects (3dLME) analysis was conducted (Chen et al. 2013). A Monte Carlo simulation was conducted to establish the appropriate cluster-level threshold to correct for multiple comparisons. "
    [Show abstract] [Hide abstract]
    ABSTRACT: The present study utilized functional near infrared spectroscopy (fNIRS) to detect neural activation differences in the orbitofrontal brain region between individuals with multiple sclerosis (MS) and healthy controls (HCs) during a working memory (WM) task. Thirteen individuals with MS and 12 HCs underwent fNIRS recording while performing the n-back WM task with four levels of difficulty (0-, 1-, 2-, and 3-back). Subjects were fitted with the fNIRS cap consisting of 30 'optodes' positioned over the forehead. The results revealed different patterns of brain activation in MS and HCs. The MS group showed an increase in brain activation, as measured by the concentration of oxygenated hemoglobin (oxyHb), in the left superior frontal gyrus (LSFG) at lower task difficulty levels (i.e. 1-back), followed by a decrease at higher task difficulty (2- and 3-back) as compared with the HC group. HC group achieved higher accuracy than the MS group on the lower task loads (i.e. 0- and 1-back), however there were no performance differences between the groups at the higher task loads (i.e. 2- and 3-back). Taken together, the results suggest that individuals with MS experience a task with the lower cognitive load as more difficult than the HC group, and the brain activation patterns observed during the task confirm some of the previous findings from functional magnetic resonance imaging (fMRI) studies. This study is the first to investigate brain activation by utilizing the method of fNIRS in MS during the performance of a cognitive task.
    Brain Imaging and Behavior 06/2014; 9(2). DOI:10.1007/s11682-014-9307-y · 4.60 Impact Factor
  • Source
    Statistical Parametric Mapping, 01/2007: pages 166-177; , ISBN: 9780123725608
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVE Poor threat-safety discrimination reflects prefrontal cortex dysfunction in adult anxiety disorders. While adolescent anxiety disorders are impairing and predict high risk for adult anxiety disorders, the neural correlates of threat-safety discrimination have not been investigated in this population. The authors compared prefrontal cortex function in anxious and healthy adolescents and adults following conditioning and extinction, processes requiring threat-safety learning. METHOD Anxious and healthy adolescents and adults (N=114) completed fear conditioning and extinction in the clinic. The conditioned stimuli (CS+) were neutral faces, paired with an aversive scream. Physiological and subjective data were acquired. Three weeks later, 82 participants viewed the CS+ and morphed images resembling the CS+ in an MRI scanner. During scanning, participants made difficult threat-safety discriminations while appraising threat and explicit memory of the CS+. RESULTS During conditioning and extinction, the anxious groups reported more fear than the healthy groups, but the anxious adolescent and adult groups did not differ on physiological measures. During imaging, both anxious adolescents and adults exhibited lower activation in the subgenual anterior cingulate cortex than their healthy counterparts, specifically when appraising threat. Compared with their age-matched counterpart groups, anxious adults exhibited reduced activation in the ventromedial prefrontal cortex when appraising threat, whereas anxious adolescents exhibited a U-shaped pattern of activation, with greater activation in response to the most extreme CS+ and CS-. CONCLUSIONS Two regions of the prefrontal cortex are involved in anxiety disorders. Reduced subgenual anterior cingulate cortex engagement is a shared feature in adult and adolescent anxiety disorders, but ventromedial prefrontal cortex dysfunction is age-specific. The unique U-shaped pattern of activation in the ventromedial prefrontal cortex in many anxious adolescents may reflect heightened sensitivity to threat and safety conditions. How variations in the pattern relate to later risk for adult illness remains to be determined.
    American Journal of Psychiatry 08/2013; 170(10). DOI:10.1176/appi.ajp.2013.12050651 · 13.56 Impact Factor
Show more