Bayesian Model Selection for Group Studies

Wellcome Trust Centre for Neuroimaging, Institute of Neurology, University College London, London, UK.
NeuroImage (Impact Factor: 6.36). 04/2009; 46(4):1004-17. DOI: 10.1016/j.neuroimage.2009.03.025
Source: PubMed


Bayesian model selection (BMS) is a powerful method for determining the most likely among a set of competing hypotheses about the mechanisms that generated observed data. BMS has recently found widespread application in neuroimaging, particularly in the context of dynamic causal modelling (DCM). However, so far, combining BMS results from several subjects has relied on simple (fixed effects) metrics, e.g. the group Bayes factor (GBF), that do not account for group heterogeneity or outliers. In this paper, we compare the GBF with two random effects methods for BMS at the between-subject or group level. These methods provide inference on model-space using a classical and Bayesian perspective respectively. First, a classical (frequentist) approach uses the log model evidence as a subject-specific summary statistic. This enables one to use analysis of variance to test for differences in log-evidences over models, relative to inter-subject differences. We then consider the same problem in Bayesian terms and describe a novel hierarchical model, which is optimised to furnish a probability density on the models themselves. This new variational Bayes method rests on treating the model as a random variable and estimating the parameters of a Dirichlet distribution which describes the probabilities for all models considered. These probabilities then define a multinomial distribution over model space, allowing one to compute how likely it is that a specific model generated the data of a randomly chosen subject as well as the exceedance probability of one model being more likely than any other model. Using empirical and synthetic data, we show that optimising a conditional density of the model probabilities, given the log-evidences for each model over subjects, is more informative and appropriate than both the GBF and frequentist tests of the log-evidences. In particular, we found that the hierarchical Bayesian approach is considerably more robust than either of the other approaches in the presence of outliers. We expect that this new random effects method will prove useful for a wide range of group studies, not only in the context of DCM, but also for other modelling endeavours, e.g. comparing different source reconstruction methods for EEG/MEG or selecting among competing computational models of learning and decision-making.

Download full-text


Available from: Karl J Friston,
    • "not through intrinsic hierarchical connections) provided the most accurate and parsimonious account of the at-risk functional MRI data. Taking the ratio of the posterior probability of models in both the control and at-risk groups yields an odds ratio of non-linear versus bilinear models of 2.6, corresponding to a robust effect (Stephan et al., 2009a). "
    [Show abstract] [Hide abstract]
    ABSTRACT: The emotional and cognitive vulnerabilities that precede the development of bipolar disorder are poorly understood. The inferior frontal gyrus-a key cortical hub for the integration of cognitive and emotional processes-exhibits both structural and functional changes in bipolar disorder, and is also functionally impaired in unaffected first-degree relatives, showing diminished engagement during inhibition of threat-related emotional stimuli. We hypothesized that this functional impairment of the inferior frontal gyrus in those at genetic risk of bipolar disorder reflects the dysfunction of broader network dynamics underlying the coordination of emotion perception and cognitive control. To test this, we studied effective connectivity in functional magnetic resonance imaging data acquired from 41 first-degree relatives of patients with bipolar disorder, 45 matched healthy controls and 55 participants with established bipolar disorder. Dynamic causal modelling was used to model the neuronal interaction between key regions associated with fear perception (the anterior cingulate), inhibition (the left dorsolateral prefrontal cortex) and the region upon which these influences converge, namely the inferior frontal gyrus. Network models that embodied non-linear, hierarchical relationships were the most strongly supported by data from our healthy control and bipolar participants. We observed a marked difference in the hierarchical influence of the anterior cingulate on the effective connectivity from the dorsolateral prefrontal cortex to the inferior frontal gyrus that is unique to the at-risk cohort. Non-specific, non-hierarchical mechanisms appear to compensate for this network disturbance. We thus establish a specific network disturbance suggesting dysfunction in the processes that support hierarchical relationships between emotion and cognitive control in those at high genetic risk for bipolar disorder.
    Brain 11/2015; 138(11):3427-3439. DOI:10.1093/brain/awv261 · 9.20 Impact Factor
  • Source
    • "The estimated model evidence reflects the plausibility of the model taking into account both goodness of fit and model complexity. We used random-effects Bayesian model comparison to evaluate the plausibility of every model of the model space across the population (Stephan et al. 2009) and report the results in terms of protected exceedance probabilities (Rigoux et al. 2014) throughout the study. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Interactions between motivational, cognitive, and motor regions of the striatum are crucial for implementing behavioral control. Work with experimental animals indicates that such interactions are sensitive to modulation by dopamine. Using systematic pharmacological manipulation of dopamine D2-receptors and resting-state functional imaging, we defined the functional architecture of the human striatum and quantified the effects of dopaminergic drugs on intrinsic effective connectivity between striatal subregions. We found that dopamine modulates interactions between motivational and cognitive regions, as well cognitive and motor regions of the striatum. Stimulation and blockade of the dopamine D2-receptor had opposite (increasing and decreasing) effects on the efficacy of those interactions. Furthermore, trait impulsivity was specifically associated with dopaminergic modulation of ventral-to-dorsal striatal connectivity. Individuals with high trait impulsivity exhibited greater drug-induced increases (after stimulation) and decreases (after blockade) of ventral-to-dorsal striatal connectivity than those with low trait impulsivity. These observations establish a key link between dopamine, intrinsic effective connectivity between striatal subregions, and trait impulsivity.
    Cerebral Cortex 10/2015; DOI:10.1093/cercor/bhv243 · 8.67 Impact Factor
    • "The 'regulate'-condition was assumed to modulate connectivity in either bottom-up, top-down or both directions. This results in a total of nine models which were compared with Bayesian model selection (BMS) (Stephan et al., 2009, 2010). The first eigenvariate of the single-subject time courses was extracted from volumes located in the right amygdala and the vmPFC (spheres with radius = 6 mm) and adjusted for the effects of interest. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Down-regulation of the amygdala with real-time fMRI neurofeedback (rtfMRI NF) potentially allows targeting brain circuits of emotion processing and may involve prefrontal-limbic networks underlying effective emotion regulation. Little research has been dedicated to the effect of rtfMRI NF on the functional connectivity of the amygdala and connectivity patterns in amygdala down-regulation with neurofeedback have not been addressed yet. Using psychophysiological interaction analysis of fMRI data, we present evidence that voluntary amygdala down-regulation by rtfMRI NF while viewing aversive pictures was associated with increased connectivity of the right amygdala with the ventromedial prefrontal cortex (vmPFC) in healthy subjects (N=16). In contrast, a control group (N=16) receiving sham feedback did not alter amygdala connectivity (Group x Condition t-contrast: p<.05 at cluster-level). Task-dependent increases in amygdala-vmPFC connectivity were predicted by picture arousal (β=.59, p<.05). A dynamic causal modeling analysis with Bayesian model selection aimed at further characterizing the underlying causal structure and favored a bottom-up model assuming predominant information flow from the amygdala to the vmPFC (xp=.90). The results were complemented by the observation of task-dependent alterations in functional connectivity of the vmPFC with the visual cortex and the ventrolateral PFC in the experimental group (Condition t-contrast: p<.05 at cluster-level). Taken together, the results underscore the potential of amygdala fMRI neurofeedback to influence functional connectivity in key networks of emotion processing and regulation. This may be beneficial for patients suffering from severe emotion dysregulation by improving neural self-regulation.
    NeuroImage 10/2015; 125. DOI:10.1016/j.neuroimage.2015.10.027 · 6.36 Impact Factor
Show more