Assignment of functional activations to probabilistic cytoarchitectonic areas revisited
Introduction
Studies in humans and non-human primates have shown that regionally specific differences in cortical function are underlined by concurrent differences in cyto-, myelo- and connectional architecture (Coq et al., 2004, Eickhoff et al., 2005b, Eickhoff et al., 2006a, Luppino et al., 1991, Matelli et al., 1991, Nelissen et al., 2005, Wilms et al., 2005, Wu and Kaas, 2003). Consequently, there is an emerging consensus that cortical areas, defined by their microstructure and/or connectivity, can be regarded as the functional modules of the cerebral cortex (Eickhoff et al., 2005b, Felleman and Van Essen, 1991, Passingham et al., 2002, Zilles et al., 2002). Integration of functional observations with a detailed knowledge of microstructural architecture is therefore crucial for understanding the principles underlying cortical organisation. This integration has a longstanding tradition in experimental research in non-human primates, where the electrode tracks and thus regionally specific functional properties can directly be related to cyto- or myeloarchitectonic preparations of the same individual (Luppino et al., 1991).
For humans, functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) can provide information about the functional organisation of the cerebral cortex with a spatial resolution of few millimetres. There is, however, currently no method to analyse function and microstructure in one and the same human brain, although recent studies using high-resolution MR imaging with a spatial resolution at almost microscopic level provide some future hope in that respect (Augustinack et al., 2005, Eickhoff et al., 2005a, Fatterpekar et al., 2002, Walters et al., 2007). Therefore, the contemporary approach for analysing the correspondence between structure and function in the human brain is to perform both analyses separately (in two groups of subjects) and integrate the obtained data by brain atlases. The development of probabilistic cytoarchitectonic maps (Amunts and Zilles, 2001, Eickhoff et al., 2005b, Zilles et al., 2002) as an anatomical reference for functional neuroimaging studies promoted such integration, as in contrast to classical anatomical brain atlases (e.g., Brodmann, 1909), these maps (Table 1) are based on observer-independent cytoarchitectonic analysis of 10 post-mortem brains (Schleicher et al., 2005) and provide stereotaxic information on the location and variability of cortical areas in the MNI reference space (Eickhoff et al., 2005b). More recently, add-ons to the analysis software packages SPM (http://www.fil.ion.ucl.ac.uk/spm), FSL (http://www.fmrib.ox.ac.uk/fsl) and AFNI (http://afni.nimh.nih.gov/afni/) have been introduced, which enable a routine, standardised application of these architectonic maps as an anatomical reference for functional activations (Eickhoff et al., 2005b). This incorporation into widely used environments for the analysis of functional imaging data assisted the use of probabilistic cytoarchitectonic maps in the context of neuroimaging studies and may establish a common anatomical reference framework for neuroimaging experiments. In this context, anatomical information may take two important roles, (i) to supply a-priori information which can be used to investigate hypothesised regionally specific effects (Eickhoff et al., 2006c), (ii) to enable objective inference on the cortical areas forming the structural substrate of a functional activation (Eickhoff et al., 2005b). This report focuses on the second aspect, illustrated by an fMRI activation in the anterior parietal lobe evoked by viewing videos of emotionally neutral hand gestures. Using this example data, we will first review the basic approaches for anatomical localisation of functional activations using probabilistic cytoarchitectonic maps and comment on some of their potential difficulties. We will then propose an additional quantification of the correspondence between functional imaging activations and probabilistic anatomical maps based on distribution analysis, and close by discussing the advantages and disadvantages of the various methods in different scenarios, focussing on the question which approach is most appropriate for a particular situation.
Section snippets
Example fMRI data
All methods are illustrated using fMRI data imaging the perception of dynamic hand movements and facial expressions (Grosbras and Paus, 2005). In particular, we examined the anatomical substrates of changes in fMRI signal, henceforth “activation”, that were observed on the left postcentral gyrus and anterior parietal cortex following the observation of emotionally neutral hand movements. Experimental details have been described elsewhere (Grosbras and Paus, 2005). In short, 20 healthy
Existing approaches
Volume-based labelling of functional activations is based on calculating the intersection between a cluster of activation and anatomically defined areas. When the intersecting volume is related to the size of the respective cluster, the resulting “cluster labelling” partitions the functional activation based on the underlying anatomical structures (e.g., 50% of this activation is allocated to area X, 30% to area Y and 20% to area Z). Conversely, the intersection volume can also be expressed as
Existing approaches
Whereas volume-based approaches treat all super-threshold voxels as equivalent, local maxima labelling focuses only on the most significant voxels, reflecting the location of the strongest (functional) effects. Furthermore, and in contrast to the total cluster volume, these centres of activation are not affected by the threshold applied for statistical inference.
The labelling of a local maximum can be performed by comparing its location to the MPM, testing whether this position is assigned to a
Discussion
Correlating the activation identified in functional imaging studies of the human brain with structural (e.g., cytoarchitectonic) information on the activated areas is a major methodological challenge for neuroscience research. Probabilistic cytoarchitectonic maps have provided a promising approach to these challenges in several studies of somatosensory (Eickhoff et al., 2006e, Eickhoff et al., 2006f, Grefkes et al., 2006, Naito et al., 2005, Young et al., 2004), motor (Binkofski et al., 2000,
Acknowledgments
This Human Brain Project/Neuroinformatics research was funded by the National Institute of Biomedical Imaging and Bioengineering, the National Institute of Neurological Disorders and Stroke and the National Institute of Mental Health. K.Z. acknowledges funding by the Deutsche Forschungsgemeinschaft (KFO-112) and the Volkswagenstiftung. We would like to thank Lars Hömke and Klaas Enno Stephan for valuable feedback on previous versions of the distribution-based labelling and all our colleagues at
References (58)
- et al.
Brodmann's areas 17 and 18 brought into stereotaxic space—Where and how variable?
NeuroImage
(2000) - et al.
Analysis of neural mechanisms underlying verbal fluency in cytoarchitectonically defined stereotaxic space—The roles of Brodmann areas 44 and 45
NeuroImage
(2004) - et al.
The human inferior parietal cortex: cytoarchitectonic parcellation and interindividual variability
NeuroImage
(2006) - et al.
A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data
NeuroImage
(2005) - et al.
Testing anatomically specified hypotheses in functional imaging using cytoarchitectonic maps
NeuroImage
(2006) - et al.
Segregation of visceral and somatosensory afferents. An fMRI and cytoarchitectonic mapping study
Neuroimage
(2006) - et al.
Anatomical mapping of functional activation in stereotactic coordinate space
NeuroImage
(1992) - et al.
Areas 3a, 3b, and 1 of human primary somatosensory cortex. Part 2. Spatial normalization to standard anatomical space
NeuroImage
(2000) - et al.
Human somatosensory area 2: observer-independent cytoarchitectonic mapping, interindividual variability, and population map
NeuroImage
(2001) - et al.
The role of the left Brodmann's areas 44 and 45 in reading words and pseudowords
Cogn. Brain Res.
(2005)