- Email: [email protected]
IC-P2-078: Altered brain activation during verbal working memory in mild cognitive impaired subjects
Alzheimer’s Imaging Consortium IC-P2: Poster Presentations of remaining skull and dura. All measurements were conducted by one rater blind to group status. Scans were then segmented into separate tissue classes using an adaptive fuzzy C-means algorithm. Volumes of gray matter and white matter tissue as well as total brain volume were calculated. Results: At both Time 1 and Time 2, the high-risk sample showed significantly smaller white matter volumes (Time 1 p ⫽ 0.05, Time 2 p ⫽ 0.04). Change across the time points, however, did not significantly differ between groups (control ⫽ -50ccm, high-risk ⫽ -53.28ccm). Conclusions: While the similar rate of white matter loss across groups may reflect aging, this would not provide an explanation for the initial significantly smaller white matter volume seen in the high-risk group. Further study will be needed to more fully understand the significance of reduced white matter volume as it relates to AD. IC-P2-077
AMYLOID DEPOSITION RELATED TO CORTICAL THINNING
J. Alex Becker1, Jeremy Carmasin1, Bruce Fischl1, Doug Greve1, Amy DeLuca1, Pete LaViolette1, Jacqueline O’Brien1, Kelly O’Keefe1, Alan Fischman1, Dorene Rentz2, Reisa Sperling2, Keith Johnson1, 1 Massachusetts General Hospital, Boston, MA, USA; 2Brigham and Women’s Hospital, Boston, MA, USA. Contact e-mail: [email protected] Background: A clear relationship between amyloid plaque burden and neuronal loss or regional atrophy has not been demonstrated in autopsy studies. To examine this relationship in vivo, we tested the hypothesis that regional fibrillar amyloid burden, as measured by 11C Pittsburgh Compound B (PiB), is associated with reduced cortical thickness. Methods: C11-PIB PET data and 3T MRI-MPRAGE were acquired in 50 subjects who were CDR0 (N⫽30), CDR0.5 (N⫽12) or CDR1(N⫽8). PiB retention was expressed as the distribution volume ratio (DVR) at each voxel as determined by the Logan graphical method with cerebellar cortex as reference region. Parametric DVR datasets were corrected for the partial-volume effect using Meltzer’s twocompartment method. MPRAGE data were processed by FreeSurfer to extract gray/white segments and derive cortical gray matter thickness at each vertex. Freesurfer was used to transform each subject’s cortical surface into standardized space, and cortical thickness and average cortical partial-volume-corrected PiB DVR were calculated at each vertex of this standardized surface for each subject. Differences in cortical thickness and in DVR at each vertex as a contrast of the CDR0 group versus the combined CDR0.5/1 group were assessed by ANCOVA with age as a covariate. Linear regression was used to investigate the relationship of thickness and DVR at each vertex across the entire sample, with age as a covariate. Results: Amyloid deposition and cortical thinning were significantly greater in the CDR0.5/1 group compared to the CDR0 group in superior temporal, lateral parietal, precuneus, posterior cingulate, and medial frontal areas (p⬍0.001). When PiB and thickness data were regressed locally, i.e., at each vertex, greater amyloid was significantly related to cortical thinning in posterior cingulate/precuneus and temporal areas. Conclusions: Amyloid deposition is associated locally with cortical thinning, predominantly in posterior cingulate/precuneus and temporal cortices, areas where both dysfunction and atrophy are commonly observed in AD. Our findings suggest that damage to neuronal populations, whether due to soluble or insoluble amyloid or other factors, occurs at sites of amyloid deposition. IC-P2-078
ALTERED BRAIN ACTIVATION DURING VERBAL WORKING MEMORY IN MILD COGNITIVE IMPAIRED SUBJECTS
Arun Bokde1,2, Michaela Karmann2, Christine Born2, Stefan J. Teipel3,2, Muamer Omerovic2, Maximilian F. Reiser2, Hans-Juergen Moller2, Harald Hampel1,2, 1Trinity College, Univ of Dublin, Dublin, Ireland; 2Ludwig-Maximilian-University, Munich, Germany; 3University of Rostock, Rostock, Germany. Contact e-mail: [email protected] Background: Mild Cognitive Impairment (MCI) indicates a high risk for conversion to dementia and a clinical diagnosis of Alzheimer’s disease
T39
(AD). The objective of this study was to delineate the differences in brain activation between 8 MCI subjects and 8 health controls (HC) during a verbal working memory task (delay match-to-sample design). Methods: Previous studies of working memory in MCI examined an n-back task. The MCI group composed of subjects with at least memory impairment. Brain activation was measured using functional magnetic resonance imaging (fMRI). Groups were matched for performance. The task was analyzed as an event-related design. Results: Both groups activated a wide network in the posterior and frontal areas of the brain while at the same time deactivating the default network. There was higher activation in the parietal and frontal lobes in the MCI compared to the HC group during the encoding and maintenance stages. The HC group had higher activation than the MCI group during the encoding and retrieval stages. Both groups deactivated the default network. Conclusions: The activation differences between groups indicate compensatory mechanisms within the MCI group for the effects of the putative AD neuropathology. This has been the first study that has examined the different stages of verbal working memory in MCI subjects. IC-P2-079
NEW CORTICAL THICKNESS ESTIMATION METHOD FOR ATROPHY CHARACTERIZATION OF ALZHEIMER’S DISEASE PATIENTS
Pierrick Bourgeat1, Oscar Acosta1, Jurgen Fripp1, Sebastien Ourselin2,1, Collin Masters3,4, David Ames5, Peter Hudson6, Christopher Rowe7, Victor Villemagne3,8, Olivier Salvado1, 1CSIRO ICT Centre, Brisbane, Australia; 2University College London, London, United Kingdom; 3The Mental Health Research Institute, University of Melbourne, Parkville, Australia; 4Centre for Neurosciences, University of Melbourne, Parkville, Australia; 5National Ageing Research Institute, Parkville, Australia; 6CSIRO - CMHT, Parkville, Australia; 7Department of Nuclear Medicine and Centre for PET, and Department of Medicine University of Melbourne, Austin hospital, Melbourne, Australia; 8 Department of Nuclear Medicine and Centre for PET, and Department of Medicine University of Melbourne, Austin hospital, Parkville, Australia. Contact e-mail: [email protected] Background: Accurate cortical thickness estimation is important for the study of many neurodegenerative diseases, including Alzheimer’s disease (AD). We propose a novel voxel based method which is both accurate and computationally efficient. The novelty of the approach is to integrate the partial volume information when computing the thickness using the Laplacian definition of thickness. Methods: MR images were classified into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) in their native space by using an expectation maximisation (EM) approach. Fractional content of GM for the voxels along tissue interfaces was computed by modeling mixture of tissues and performing a maximum a posteriori classification. The cortical thickness was computed using the Laplacian approach which divides the cortex into a set of equipotential sublayers. Normalised gradient of the Laplace solution provides non intersecting streamlines between the WM and CSF. The length of these streamlines defines the thickness. A subvoxel estimate of the location of the boundary was obtained through ray-casting using both partial volume estimation and the direction of the streamlines. This estimate was used to initialise the boundary conditions when computing the length of the streamlines, measured using a set of partial differential equations. Finally, regional statistical analysis was performed using the automated anatomical labelling atlas after extraction and smoothing of the cortical thickness map projected along the WM/GM boundary. Results: The technique was applied to a database of 3D MPRAGE MRI (Table 1). Regional analysis (Table 2.) showed large differences in the hippocampus, parahippocampus and in the temporal and frontal lobes between normal controls, and Alzheimer’s disease patient. Examples of the thickness map are presented Figure 1.
AMYLOID DEPOSITION RELATED TO CORTICAL THINNING
J. Alex Becker1, Jeremy Carmasin1, Bruce Fischl1, Doug Greve1, Amy DeLuca1, Pete LaViolette1, Jacqueline O’Brien1, Kelly O’Keefe1, Alan Fischman1, Dorene Rentz2, Reisa Sperling2, Keith Johnson1, 1 Massachusetts General Hospital, Boston, MA, USA; 2Brigham and Women’s Hospital, Boston, MA, USA. Contact e-mail: [email protected] Background: A clear relationship between amyloid plaque burden and neuronal loss or regional atrophy has not been demonstrated in autopsy studies. To examine this relationship in vivo, we tested the hypothesis that regional fibrillar amyloid burden, as measured by 11C Pittsburgh Compound B (PiB), is associated with reduced cortical thickness. Methods: C11-PIB PET data and 3T MRI-MPRAGE were acquired in 50 subjects who were CDR0 (N⫽30), CDR0.5 (N⫽12) or CDR1(N⫽8). PiB retention was expressed as the distribution volume ratio (DVR) at each voxel as determined by the Logan graphical method with cerebellar cortex as reference region. Parametric DVR datasets were corrected for the partial-volume effect using Meltzer’s twocompartment method. MPRAGE data were processed by FreeSurfer to extract gray/white segments and derive cortical gray matter thickness at each vertex. Freesurfer was used to transform each subject’s cortical surface into standardized space, and cortical thickness and average cortical partial-volume-corrected PiB DVR were calculated at each vertex of this standardized surface for each subject. Differences in cortical thickness and in DVR at each vertex as a contrast of the CDR0 group versus the combined CDR0.5/1 group were assessed by ANCOVA with age as a covariate. Linear regression was used to investigate the relationship of thickness and DVR at each vertex across the entire sample, with age as a covariate. Results: Amyloid deposition and cortical thinning were significantly greater in the CDR0.5/1 group compared to the CDR0 group in superior temporal, lateral parietal, precuneus, posterior cingulate, and medial frontal areas (p⬍0.001). When PiB and thickness data were regressed locally, i.e., at each vertex, greater amyloid was significantly related to cortical thinning in posterior cingulate/precuneus and temporal areas. Conclusions: Amyloid deposition is associated locally with cortical thinning, predominantly in posterior cingulate/precuneus and temporal cortices, areas where both dysfunction and atrophy are commonly observed in AD. Our findings suggest that damage to neuronal populations, whether due to soluble or insoluble amyloid or other factors, occurs at sites of amyloid deposition. IC-P2-078
ALTERED BRAIN ACTIVATION DURING VERBAL WORKING MEMORY IN MILD COGNITIVE IMPAIRED SUBJECTS
Arun Bokde1,2, Michaela Karmann2, Christine Born2, Stefan J. Teipel3,2, Muamer Omerovic2, Maximilian F. Reiser2, Hans-Juergen Moller2, Harald Hampel1,2, 1Trinity College, Univ of Dublin, Dublin, Ireland; 2Ludwig-Maximilian-University, Munich, Germany; 3University of Rostock, Rostock, Germany. Contact e-mail: [email protected] Background: Mild Cognitive Impairment (MCI) indicates a high risk for conversion to dementia and a clinical diagnosis of Alzheimer’s disease
T39
(AD). The objective of this study was to delineate the differences in brain activation between 8 MCI subjects and 8 health controls (HC) during a verbal working memory task (delay match-to-sample design). Methods: Previous studies of working memory in MCI examined an n-back task. The MCI group composed of subjects with at least memory impairment. Brain activation was measured using functional magnetic resonance imaging (fMRI). Groups were matched for performance. The task was analyzed as an event-related design. Results: Both groups activated a wide network in the posterior and frontal areas of the brain while at the same time deactivating the default network. There was higher activation in the parietal and frontal lobes in the MCI compared to the HC group during the encoding and maintenance stages. The HC group had higher activation than the MCI group during the encoding and retrieval stages. Both groups deactivated the default network. Conclusions: The activation differences between groups indicate compensatory mechanisms within the MCI group for the effects of the putative AD neuropathology. This has been the first study that has examined the different stages of verbal working memory in MCI subjects. IC-P2-079
NEW CORTICAL THICKNESS ESTIMATION METHOD FOR ATROPHY CHARACTERIZATION OF ALZHEIMER’S DISEASE PATIENTS
Pierrick Bourgeat1, Oscar Acosta1, Jurgen Fripp1, Sebastien Ourselin2,1, Collin Masters3,4, David Ames5, Peter Hudson6, Christopher Rowe7, Victor Villemagne3,8, Olivier Salvado1, 1CSIRO ICT Centre, Brisbane, Australia; 2University College London, London, United Kingdom; 3The Mental Health Research Institute, University of Melbourne, Parkville, Australia; 4Centre for Neurosciences, University of Melbourne, Parkville, Australia; 5National Ageing Research Institute, Parkville, Australia; 6CSIRO - CMHT, Parkville, Australia; 7Department of Nuclear Medicine and Centre for PET, and Department of Medicine University of Melbourne, Austin hospital, Melbourne, Australia; 8 Department of Nuclear Medicine and Centre for PET, and Department of Medicine University of Melbourne, Austin hospital, Parkville, Australia. Contact e-mail: [email protected] Background: Accurate cortical thickness estimation is important for the study of many neurodegenerative diseases, including Alzheimer’s disease (AD). We propose a novel voxel based method which is both accurate and computationally efficient. The novelty of the approach is to integrate the partial volume information when computing the thickness using the Laplacian definition of thickness. Methods: MR images were classified into gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF) in their native space by using an expectation maximisation (EM) approach. Fractional content of GM for the voxels along tissue interfaces was computed by modeling mixture of tissues and performing a maximum a posteriori classification. The cortical thickness was computed using the Laplacian approach which divides the cortex into a set of equipotential sublayers. Normalised gradient of the Laplace solution provides non intersecting streamlines between the WM and CSF. The length of these streamlines defines the thickness. A subvoxel estimate of the location of the boundary was obtained through ray-casting using both partial volume estimation and the direction of the streamlines. This estimate was used to initialise the boundary conditions when computing the length of the streamlines, measured using a set of partial differential equations. Finally, regional statistical analysis was performed using the automated anatomical labelling atlas after extraction and smoothing of the cortical thickness map projected along the WM/GM boundary. Results: The technique was applied to a database of 3D MPRAGE MRI (Table 1). Regional analysis (Table 2.) showed large differences in the hippocampus, parahippocampus and in the temporal and frontal lobes between normal controls, and Alzheimer’s disease patient. Examples of the thickness map are presented Figure 1.