Functional MRI BOLD response to Tower of London performance of first-episode schizophrenia patients using cortical pattern matching
Introduction
Functional brain imaging techniques, such as functional magnetic resonance imaging (fMRI) and positron emission tomography (PET), provide detailed spatial information about brain activation. This information is increasingly used to generate maps of normal brain function as a reference for abnormal brain function in various neuropsychiatric conditions. While co-registration of changes in blood oxygenation levels or blood flow rates onto structural MRI scans of individual brains can be achieved with high levels of accuracy, grouping these data for quantitative comparisons (e.g., between diagnostic groups) poses a major methodological challenge due to the inter-individual variability of regional brain morphology (Brett et al., 2002). Some cortical areas, however, such as the primary motor, visual, and auditory cortex, are defined by their relationship to distinct anatomical landmarks such as central and calcarine sulci and Heschl's gyrus which are relatively invariant in position and configuration across individuals (Rademacher et al., 1993), while other areas, such as the prefrontal cortex, are highly variable across subjects thus making it difficult to choose meaningful anatomical labels (Rajkowska and Goldman-Rakic, 1995). For these regions of the neocortex, there is also little agreement on cytoarchitectonic boundaries and their relationship to sulcal and functional anatomy (Amunts et al., 2000).
At present, most analytical approaches are based on template or normalization techniques, which provide an approximation of spatial information for grouped data with reference to a coordinate system (e.g., the brain atlas of Talairach and Tournoux, 1988). However, small changes of cortical stereotactic coordinates can, in fact, represent a relatively large change in distance within the cortical surface given the three-dimensional folding pattern of cortex morphology (e.g., Fischl et al., 1999, Van Essen et al., 2000). The same limitations apply when registering stereotactic coordinates onto cytoarchitectonically-defined regions such as the Brodmann areas (Brodmann, 1909). The interpretation of such data can be further distorted by subtle brain pathology that is not sufficiently corrected by the spatial normalization procedures, potentially adding a systematic co-registration error to group comparisons of brain activation.
The present study addresses some of these limitations by employing three-dimensional continuum-mechanical image-warping techniques to derive well-resolved average representations of anatomy for the co-registration of functional brain imaging data. A key element of this strategy is the application of cortical pattern matching methodologies that permit aggregation of imaging data sets in the same anatomical reference locations across subjects, explicitly modeling and adjusting for individual differences in the cortical folding pattern as well as in overall brain size (Thompson et al., 1997, Thompson et al., 2000, Thompson et al., 2003). Many traditional spatial mapping techniques normalize data using only linear or global (e.g., Woods et al., 1993) or piece-wise linear transformation into stereotactic space, using the anterior and posterior commissures, and a small number of points on the cortical surface defined with reference to the AC–PC line, as fixed points (Talairach and Tournoux, 1988). Such a global transformation is highly insensitive to individual differences in cortical patterning.
By contrast, the cortical pattern matching method differs from traditional spatial normalization techniques by explicitly modeling major cortical landmarks, and through deforming the individual's cortical surface, creates the best fit to an average cortical pattern model. When comparing measures of gray matter (Thompson et al., 2003) and functional activation across subjects, statistical power is increased by explicitly adjusting for cortical patterning differences across individuals. Recent research shows that patients with schizophrenia show greater individual differences in the locations of major cortical surface than healthy controls (Narr et al., 2001). Thus, this method may be particularly relevant in avoiding systematic biases when comparing gray matter or functional measures in patients with neuropsychiatric conditions to healthy control subjects.
Cortical pattern matching has been applied in the current study to compare the spatial properties of brain activation of first-episode schizophrenia patients with that of matched healthy control subjects when performing the Tower of London (TOL) task. It was hypothesized that the gyral pattern-averaged model of the cortex will improve spatial co-registration of brain activation particularly in those regions of the cortex, which have a close association of cytoarchitecture and function (i.e., primary sensory projection areas).
The TOL is an adaptation of the Tower of Hanoi and consists of moving colored balls within a limited number of moves in order to achieve a given goal configuration. Previous studies reported poor TOL performance in patients with frontal brain lesions (e.g., Owen et al., 1990, Pantelis et al., 1997, Shallice, 1982) or frontal lobe dementia (Carlin et al., 2000), Parkinson's disease (e.g., Lewis et al., 2003, Morris et al., 1988, Owen et al., 1992, Robbins et al., 1994), depression (e.g., Purcell et al., 1997), and schizophrenia (e.g., Morris et al., 1995, Pantelis et al., 1997, Schall et al., 1998).
Early activation studies using single-photon emission computerized tomography (SPECT) reported that TOL performance is associated with increased activation of the frontal cortex in healthy subjects (Morris et al., 1993). Longer planning times and fewer moves to complete a problem were associated with significantly higher regional cerebral blood flow (rCBF) in the left prefrontal cortex whereas execution time was negatively correlated with both left and right prefrontal rCBF. Subsequent studies (Baker et al., 1996, Owen et al., 1996) reported additional task-related increase of rCBF in pre-motor areas, the visual cortex, thalamus, and caudate nucleus.
Dagher et al. (1999) compared rCBF dependent on task complexity as defined by the number of moves required to solve a TOL problem (i.e., correlational analysis) with rCBF contrasts of ‘on’ versus ‘off’ task performance (i.e., categorical analysis). Activated brain areas in which increases of rCBF did not correlate with task complexity were regions belonging to the dorsal stream of visual input processing (i.e., visual cortical and posterior parietal cortical areas) and regions involved in the execution and sequencing of arm movements (e.g., cerebellum, primary motor cortex, and supplementary motor area). Brain regions where levels of rCBF correlated with task complexity included the dorsolateral prefrontal cortex, the lateral pre-motor cortex, the rostral anterior cingulate cortex, and the dorsal caudate nucleus.
A similar regression model–with number of moves required to solve a problem as the independent variable and blood oxygenation level-dependent (BOLD) signal changes as the dependent variable–served as the functional MR measure in the present study (e.g., Schall et al., 2003, Van den Heuvel et al., 2003). So far, however, only a limited number of reports have been published with TOL as an activation task in schizophrenia. For instance, Andreasen et al. (1992) reported decreased rCBF (when using [133Xe]) in the left mesiofrontal cortex and cingulate in drug-naïve and neuroleptic withdrawn schizophrenia patients that was associated with the severity of frontal brain dysfunction as expressed by negative symptoms.
While neuropsychological data suggest impaired TOL performance already at an early stage of the disorder (Thienel et al., 2000), corresponding functional brain imaging studies are lacking. Such data are of particular interest since subtle structural decline of gray matter in those cortical areas–which are also subserving TOL processing–have been reported at very early stages of the disease (Pantelis et al., 2003). In early onset cases, the chronologically earliest deficits were found in parietal brain regions concerned with visuo-spatial and associative thinking, where adult deficits are known to be mediated by environmental (non-genetic) factors (Cannon et al., 2002). Over 5 years, these deficits progressed anteriorly into temporal lobes, engulfing sensorimotor cortices, executive and cognitive function associated dorsolateral prefrontal cortices, and into the frontal eye fields. Faster loss rates of gray matter in the frontal cortex were strongly correlated with the severity of negative symptoms. Statistical maps on the cortical surface further revealed regionally specific linkages between gray matter losses and several IQ sub-tests on information processing, comprehension, and vocabulary. These findings suggest an overall deterioration of global functioning and cognitive decline, consistent with progressive structural deterioration. These emerging dynamic patterns were controlled for medication and IQ effects, replicated in independent groups of males and females, and charted in individuals and groups (Thompson et al., 2001).
The present study directly assesses function–structure relationships by mapping task difficulty-dependent BOLD changes as a function of cortical gray matter thickness. We hypothesized that the extent of BOLD response–and, by extension, the performance on the task–is partially dependent on the integrity of regions of gray matter that are subserving TOL problem solving performance (i.e., prefrontal, frontal, and parietal cortex).
Section snippets
Materials and methods
The current study was conducted as part of the Brain Atlasing Initiative of the Neuroscience Institute of Schizophrenia and Allied Disorders (NISAD). Ethics approval was granted by the human research ethics committees of the contributing institutions: the University of Newcastle Human Research Ethics Committee, the Hunter Area Health Research Ethics Committee, the University of New South Wales Ethics Research Committee, the South Western Sydney Area Health Service Human Research Ethics
Results
Functional activation maps of healthy subjects corresponded with previous findings (reviewed by Schall et al., 2003) and confirmed an increasing BOLD response with TOL task difficulty (i.e., number of moves) in the left and right superior frontal gyrus (BA 10 and 6), the right middle frontal gyrus (BA 6/9/44/46), the left and right opercularis region (BA 1/2/40/43), the left and right superior parietal lobule (BA 5 and 7), and the left inferior occipital lobe (BA 17/18/19/37; Figs. 1B and D;
Discussion
The current study pursued three objectives. First, to compare BOLD responses co-registered on an intensity-averaged model (IA) with BOLD responses deformed onto the gyral pattern-averaged (GPA) model of the cortical surface; second, to compare the cortical BOLD response pattern of first-episode male schizophrenia patients with age-matched male healthy control subjects; and third, to analyze differences of cortical gray matter thickness between the two groups and its association with BOLD
Acknowledgments
We are grateful for the skilled radiographic assistance of Steve Hudson, Gary O'Connor, Jo Donovan, and Mary Dwyer. The Neuroscience Institute of Schizophrenia and Allied Disorders (NISAD) and the Hunter Medical Research Institute (HMRI) are supported by infrastructure funding from NSW Health. R.T. and S.B were supported by the IFORES Program, Faculty of Medicine, the University of Duisburg-Essen. Algorithm development was supported in part by National Institute for Biomedical Imaging and
References (45)
- et al.
Brodmann's areas 17 and 18 brought into stereotaxic space—Where and how variable?
NeuroImage
(2000) - et al.
Neural systems engaged by planning: a PET study of the Tower of London task
Neuropsychologia
(1996) - et al.
Planning impairments in frontal lobe dementia and frontal lobe lesion patients
Neuropsychologia
(2000) - et al.
Using executive heterogeneity to explore the nature of working memory deficits in Parkinson's disease
Neuropsychologia
(2003) - et al.
Frontal and parietal participation in problem solving in the Tower of London: fMRI and computational modeling of planning and high-level perception
Neuropsychologia
(2003) - et al.
Neural correlates of planning ability: frontal lobe activation during the Tower of London test
Neuropsychologia
(1993) - et al.
Problem solving in schizophrenia: a specific deficit in planning ability
Schizophr. Res.
(1995) - et al.
Planning and spatial working memory following frontal lobe lesions in man
Neuropsychologia
(1990) - et al.
Neuroanatomical abnormalities before and after onset of psychosis: a cross-sectional and longitudinal MRI comparison
Lancet
(2003) - et al.
Functional brain maps of Tower of London performance: a PET and fMRI study
NeuroImage
(2003)
Frontostriatal system in planning complexity: a parametric functional magnetic resonance version of Tower of London task
NeuroImage
A voxel-based method for the statistical analysis of gray and white matter density applied to schizophrenia
NeuroImage
Hypofrontality in neuroleptic-naive patients and in patients with chronic schizophrenia. Assessment with xenon 133 single-photon emission computed tomography and the Tower of London
Arch. Gen. Psychiatry
Effects of olanzapine versus clozapine on executive functions in schizophrenia
Schizophr. Res.
The problem of functional localization in the human brain
Nat. Rev., Neurosci.
Vergleichende Lokalisationslehre der Grosshimrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues
Cortex mapping reveals regionally specific patterns of genetic and disease-specific gray-matter deficits in twins discordant for schizophrenia
Proc. Nat. Acad. Sci. U. S. A.
Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space
J. Comput. Assist. Tomogr.
Mapping the network for planning: a correlational PET activation study with the Tower of London task
Brain
The pharmacology of human working memory
Int. J. Neuropsychopharmacol.
High-resolution intersubject averaging and a coordinate system for the cortical surface
Hum. Brain Mapp.
Spatial registration and normalization of images
Hum. Brain Mapp.
Cited by (90)
Anti-Hebbian plasticity in the motor cortex promotes defensive freezing
2023, Current BiologyEffects of aripiprazole versus risperidone on brain activation during planning and social-emotional evaluation in schizophrenia: A single-blind randomized exploratory study
2017, Progress in Neuro-Psychopharmacology and Biological PsychiatryCitation Excerpt :Previous studies have also shown altered activation in posterior visual and parietal brain regions after treatment with both aripiprazole and strong dopamine receptor antagonists (Ikuta et al., 2014; Handley et al., 2013). Patients also have a lower activation in occipital regions compared to healthy subjects during ToL performance (Rasser et al., 2005). It has been suggested that from onset of psychosis, the earliest deficits may be found in parietal brain regions involved in visuo-spatial and associative thinking (Cannon et al., 2002).
Multimodal Fusion of Brain Imaging Data: A Key to Finding the Missing Link(s) in Complex Mental Illness
2016, Biological Psychiatry: Cognitive Neuroscience and NeuroimagingAcute Exercise and Event-Related Potential: Current Status and Future Prospects
2016, Exercise-Cognition Interaction: Neuroscience Perspectives