Friston K.J. Unified segmentation

Wellcome Department of Imaging Neuroscience, 12 Queen Square, London, WC1N 3BG, UK.
NeuroImage (Impact Factor: 6.36). 08/2005; 26(3):839-51. DOI: 10.1016/j.neuroimage.2005.02.018
Source: PubMed


A probabilistic framework is presented that enables image registration, tissue classification, and bias correction to be combined within the same generative model. A derivation of a log-likelihood objective function for the unified model is provided. The model is based on a mixture of Gaussians and is extended to incorporate a smooth intensity variation and nonlinear registration with tissue probability maps. A strategy for optimising the model parameters is described, along with the requisite partial derivatives of the objective function.

55 Reads
  • Source
    • "Preprocessing steps included bias-field correction and segmentation into gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF). Segmented images were registered to standard Montreal Neurological Institute (MNI) space using the high-dimensional Dartel approach (Ashburner and Friston, 2005 "
    [Show abstract] [Hide abstract]
    ABSTRACT: There is increasing research interest in the structural and functional brain correlates underlying creative potential. Recent investigations found that interindividual differences in creative potential relate to volumetric differences in brain regions belonging to the default mode network, such as the precuneus. Yet, the complex interplay between creative potential, intelligence, and personality traits and their respective neural bases are still under debate. We investigated regional gray matter volume (rGMV) differences that can be associated with creative potential in a heterogeneous sample of N = 135 individuals using voxel-based morphometry (VBM). By means of latent variable modeling and consideration of recent psychometric advancements in creativity research, we sought to disentangle the effects of ideational originality and fluency as two independent indicators of creative potential. Intelligence and openness to experience were considered as common covariates of creative potential. The results confirmed and extended previous research: rGMV in the precuneus was associated with ideational originality, but not with ideational fluency. In addition, we found ideational originality to be correlated with rGMV in the caudate nucleus. The results indicate that the ability to produce original ideas is tied to default-mode as well as dopaminergic structures. These structural brain correlates of ideational originality were apparent throughout the whole range of intellectual ability and thus not moderated by intelligence. In contrast, structural correlates of ideational flueny, a quantitative marker of creative potential, were observed only in lower intelligent individuals in the cuneus / lingual gyrus.
    NeuroImage 12/2015; 7. DOI:10.1016/j.neuroimage.2015.02.002 · 6.36 Impact Factor
    • "The VOIs were designed and then converted into the NIFTI format by means of the software MRIcron ( mricron) to ensure compatibility with the DPARSF-A software (Yan & Zang, 2010). Next, the fMRI scans were realigned, the volumetric T1 image was co-registered to the functional images and then subjected to the unified segmentation algorithm (Ashburner & Friston, 2005) to obtain maps of the grey matter, white matter and cerebrospinal fluid (CSF). At this stage, signal decomposition was applied. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Three cognitive components may play a crucial role in both memory awareness and in anosognosia for memory deficit (AMD): (1) a personal data base (PDB), i.e., a memory store that contains "semantic" representations about the self, (2) monitoring processes (MPs) and (3) an explicit evaluation system (EES), or comparator, that assesses and binds the representations stored in the PDB with information obtained from the environment. We compared both the behavior and the functional connectivity (as assessed by resting-state fMRI) of AMD patients with aware patients and healthy controls. We found that AMD is associated with an impoverished PDB, while MPs are necessary to successfully update the PDB. AMD was associated with reduced functional connectivity within both the default-mode network and in a network that includes the left lateral temporal cortex, the hippocampus and the insula. The reduced connectivity between the hippocampus and the insular cortex was correlated with AMD severity.
    Consciousness and Cognition 09/2015; 37:160-177. DOI:10.1016/j.concog.2015.08.010 · 2.31 Impact Factor
  • Source
    • "Finally, grey matter segments were smoothed using a 6x6x6 mm 3 full-width-at-half maximal Gaussian kernel to increase the signal-to-noise ratio and ensure a Gaussian distribution allowing data analysis in the general linear model (Ashburner and Friston 2005). Analyses were performed on a voxel-by-voxel basis employing the framework of the General Linear Model within SPM8. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Genetic, behavioural and functional neuroimaging studies have revealed that different vulnerabilities characterise children with conduct problems and high levels of callous-unemotional traits (CP/HCU) compared with children with conduct problems and low callous-unemotional traits (CP/LCU). We used voxel-based morphometry to study grey matter volume (GMV) in 89 male participants (aged 10-16), 60 of whom exhibited CP. The CP group was subdivided into CP/HCU (n = 29) and CP/LCU (n = 31). Whole-brain and regional GMV were compared across groups (CP vs. typically developing (TD) controls (n = 29); and CP/HCU vs. CP/LCU vs. TD). Whole-brain analyses showed reduced GMV in left middle frontal gyrus in the CP/HCU group compared with TD controls. Region-of-interest analyses showed reduced volume in bilateral orbitofrontal cortex (OFC) in the CP group as a whole compared with TD controls. Reduced volume in left OFC was found to be driven by the CP/HCU group only, with significant reductions relative to both TD controls and the CP/LCU group, and no difference between these latter two groups. Within the CP group left OFC volume was significantly predicted by CU traits, but not conduct disorder symptoms. Reduced right anterior cingulate cortex volume was also found in CP/HCU compared with TD controls. Our results support previous findings indicating that GMV differences in brain regions central to decision-making and empathy are implicated in CP. However, they extend these data to suggest that some of these differences might specifically characterise the subgroup with CP/HCU, with GMV reduction in left OFC differentiating children with CP/HCU from those with CP/LCU.
    Journal of Abnormal Child Psychology 09/2015; DOI:10.1007/s10802-015-0073-0 · 3.09 Impact Factor
Show more