- Email: [email protected]
Prefrontal lateralization during a verbal fluency task predicts neuropsychological changes in healthy elderly controls: a fNIRS study
Alzheimer’s Imaging Consortium: IC-P-Poster Imaging (CC) but normal neuropsychological test performance (Saykin, 2006). The current study was designed to assess RS-fMRI patterns in CC compared with MCI, AD and healthy controls (HC). Methods: To date, 13 CC, 9 HC, 4 MCI and 3 AD participants were scanned at rest with eyes closed on a Siemens 3T. RS-fMRI was analyzed using FSL, AFNI and SPM8. For each individual, the sum of amplitude of low frequency fluctuation (ALFF; 0.01-0.1 Hz) was calculated at each voxel (Biswal, 2010). Using PCC seed ROIs adapted from Fox et al (2005) voxelwise cross-correlation maps were generated for each subject. Group comparisons and covariate analyses were performed using SPM8 with age as a covariate. Results: Compared to HC, MCI/AD showed decreased ALFF in the PCC (p < 0.01, corrected), but increased ALFF in bilateral hippocampi (p < 0.01). The CC group consistently showed intermediate changes. ROI analyses indicated differences in ALFF of PCC (HC > CC > MCI/AD, p < 0.05, effect size: 0.61), and ALFF of hippocampus (HC < CC < MCI/AD, p < 0.01, effect size: 0.75). ALFF of PCC was positively correlated with neuropsychological performance (MMSE, DRS and CVLT; r ¼ 0.45 to 0.56, p < 0.01), while hippocampal ALFF was negatively correlated with performance (r ¼ -0.48 to -0.67, p < 0.01). PCC seeded cross-correlation maps showed decreased hippocampal connectivity in MCI/AD compared to HC or CC (p < 0.01). Conclusions: RS-fMRI appears sensitive to early prodromal neurodegenerative changes in regions associated with AD, notably including pre-MCI individuals with CC. While there is decreased functional connectivity between PCC and hippocampus, regionally increased ALFF in hippocampus may indicate a compensatory mechanism in early prodromal AD.
Subjects
PREDICTION OF MCI TO ALZHEIMER’S CONVERSION WITH HIPPOCAMPAL SHAPE FEATURES AND SUPPORT VECTOR MACHINE
Jonathan Young1, Gerard Ridgway2, Kelvin Leung2, Josephine Barnes2, Sebastien Ourselin1, 1University College London, London, United Kingdom; 2UCL Institute of Neurology, London, United Kingdom. Background: Hippocampal atrophy is a marker of the onset of Alzheimer’s disease (AD) and hippocampal volumetry has been used in a number of studies to provide early diagnosis of AD and predict conversion of Mild Cognitive Impairment (MCI) to AD. However rates of atrophy are not uniform across the hippocampus, making shape a potentially more sensitive biomarker than volume. This study investigated the utility of SPHARM parameterizations of hippocampal shape coupled with a support vector machine (SVM) classifier to predict conversion from MCI to AD. Methods: Left and right hippocampi from a total of 330 images of MCI patients from the ADNI database were automatically segmented. The resulting volumes were then processed with to produce a description of each shape as a weighted sum of spherical harmonic basis functions. The weights of each shape were taken as a feature vector for
Training 1
Testing 2
82 MCI-converter, 138 MCI-stable 74% 0.8 (68%, 80%)
41 MCI-converter, 69 MCI-stable 68% 0.74 (58%, 77%)
Area under curve Classification accuracy 95% confidence interval on accuracy (1): Accuracy and AUC calculated from ten-fold cross validation on all 220 subjects. Best accuracy found when using 143 hippocampal shape features for classification, best AUC found when using 170 hippocampal shape features for classification. (2): Parameters yielding best AUC in training were used to build an SVM classifier from all 220 training subjects. This was then applied to the hitherto unused 110 remaining subjects to test the performance of the classifier on new data.
SVM classification. Separate classifiers were built for right and left hippocampal shape features and for a concatenation of the two. For each experiment, two thirds of the subjects were used to build a classifier. The shape features were ranked according to their ability to separate MCI converter and MCI stable subjects, and then a ten fold cross validation was performed for each classifier to find the number of hippocampal features and SVM parameters yielding the best classification results. The resulting classifiers were then applied to an unseen test set comprising the remaining third of the subjects. Results: Performance with features from the right hippocampus only was superior to left only or both together. As AUC is considered a better measure of classification efficacy than accuracy, the parameters yielding the best AUC were used when classifying unseen data. In this case classification accuracy was 68% with a binomial (exact) 95% confidence interval of 58% to 77% and an AUC of 0.74. Details are given in the table. Conclusions: This method shows promise in discriminating between MCI stable and MCI converter subjects. The best classification results are obtained with a somewhat biased classifier, meaning that specificity was substantially better than sensitivity, however the AUC shows that this classifier can potentially be both accurate and balanced.
IC-P-080 IC-P-079
S41
PREFRONTAL LATERALIZATION DURING A VERBAL FLUENCY TASK PREDICTS NEUROPSYCHOLOGICAL CHANGES IN HEALTHY ELDERLY CONTROLS: A FNIRS STUDY
Julia Zeller1, Thomas Polak1, Andreas Fallgatter2, 1University W€urzburg, W€urzburg, Germany; 2University T€ubingen, T€ubingen, Germany. Background: Near-Infrared Spectroscopy (NIRS) is a non-invasive optical method measuring brain activation by assessing changes in the concentration of oxy- [O2Hb] and deoxygenated [HHb] haemoglobin in the cerebral cortex. During a verbal fluency task (VFT) healthy controls display a typical pattern of activation in prefrontal regions with a distinct lateralisation to the left dorso-lateral prefrontal cortex (DLPFC). This lateralisation decreases with age indicating compensatory processes. We expected the amount of lateralisation during the VFT to predict neuropsychological changes in healthy elderly controls over a one year interval. Methods: 44 elderly healthy controls (aged 55-82 years) were measured during two versions of a VFT (phonological and semantical) using 44-channel functional NIRS. MMST and DemTect were performed at the time of the fNIRS
S42
Alzheimer’s Imaging Consortium: IC-P-Poster Imaging
measurement and again after one year. Results: During both versions of the VFT subjects displayed a typical activation pattern in the DLPFC with a distinct increase in O2Hb. Subjects’ lateralisation did not correlate with age. The amount of lateralisation during the letter version could account for a significant amount of the changes in MMST (16%) and DemTect (9%) after one year. Subjects displaying less distinct or no lateralisation showed a decrease in neuropsychological performance whereas subjects with a clearly lateralized activation did not change in MMST or DemTect. Conclusions: fNIRS seems to be a suitable tool for the study of long term changes in elderly subjects when assessing brain activation and cognitive decline. As alternations in oxygenation probably occur even before atrophy in affected brain regions, fNIRS could add to the development of methods for the early detection of mild cognitive impairment or Alzheimer’s disease.
IC-P-081
NEAR-INFRARED SPECTROSCOPY: A RELIABLE TOOL TO ASSESS CORTICAL CHANGES IN ALZHEIMER’S DISEASE AND THEIR PROGRESSION OVER TIME
Julia Zeller1, Thomas Polak1, Andreas Fallgatter2, 1University W€urzburg, W€ urzburg, Germany; 2University, T€ubingen, Germany. Background: The technique of Near-Infrared Spectroscopy (NIRS) is a non-invasive optical method allowing the in vivo measurement of brain activation via changes in the concentration of oxy- [O2Hb] and deoxygenated [HHb] haemoglobin in the cerebral cortex. Neuroimaging studies show widespread activation deficits in patients suffering from Alzheimer’s disease (AD) during tasks targeting the visuo-spatial and executive domains located in prefrontal and parietal areas. Using NIRS the present study assessed changes in prefrontal and parietal oxygenation in AD patients and healthy controls and their progression over a one-year interval. Methods: 71 Patients suffering from AD and 68 elderly healthy controls were measured during two versions of a verbal fluency task (VFT, letter and category) and a line orientation task (LO) using 44-channel (VFT) and 52-channel (LO) fNIRS. 14 patients and 49 controls completed a follow-up measurement after one year. Results: During the letter version of the VFT and the LO AD-patients displayed a reduced activation pattern in the prefrontal (VFT) and the parietal cortex (LO) as compared to healthy controls. Increases in O2Hb in parietal regions correlated with prefrontal activity in AD-patients but remained independent in healthy controls. Prefrontal activation decreased significantly over the one year interval in ADpatients. Healthy controls showed no changes at follow up. Activation decreases in the prefrontal cortex were correlated with the decrease of the MMSE score, with greater decrease indicating lower MMSE score at the follow-up. Conclusions: fNIRS seems to be a well suited instrument to measure disease related changes in the cortex of AD patients and their progression over time. Considering the fact that we did not observe any changes in healthy controls without any cognitive decline fNIRS could be a promising device to be used in prospective studies of AD and other dementias.
IC-P-082
3-D RECONSTRUCTION OF CT IMAGES BY LABVIEW
Pravin Jogdand, VJTI, Mumbai, India.
Background: CT scan imaging is popular and plays a unique role in clinical diagnosis and treatment because it is a noninvasive and real-time. Diseases like Alzheimer’s need to be diagnosed at early stages only so that their cure is possible. For this resolution up to micron level is required for analyzing 2D and 3D image. If we try to improve image resolution then cost of equipment increases. So there is trade of between cost of equipment andresolution of image. Methods: This trade of can be overcome by using LabVIEWwith Biomedical start up kit 3.0. In this the 2D slices of DIOCOM images obtained from CT scanmachine is processed for increasing resolution. This process does not require to change the module of equipment. The LabVIEW is a graphical programming environment, which can be used to built data acquisition and instrument control application. Results: This paper introduces the experiment of 3D reconstruction CT images via Virtual Instrumentation - LabVIEW. The main idea is based on marching cubes algorithm and image processing implemented by module with the help of NI Biomedical Startup Kit.The two dimensional images shot by the CT scan device provide information about the surface properties of human body. There is implemented algorithm which can be used for 3D reconstruction of CT images in biomedical application. Images of array 2D slices from CT scan machine can be saved in CD or hard disc and can be imported to study 3D module of image for diagnosis. Conclusions: In this paper we have presented method for 3D reconstruction of biomedical images from CT. The presented method is implemented in program LabVIEW. Presented CT images of head were obtained from Visible Human. The main addition of this work is implementation of mentioned method in program LabVIEW and Biomedical Start up kit 3.0(Biomedical Workbench). Experimental result in LabVIEW is comparable to conventional program as Matlab, Visualization tool kit(VTK) with respect to resolution and image handling capability. The disadvantage of mentioned method is processing time is more as compared to other algorithm and software.
IC-P-083
BETA-AMYLOID BINDING OF [3]AZD4694 AND [3H]AV45 IN VITRO
Fredrik Jeppsson, Anders Jureus, Johan Sandell, Britt-Marie Swahn, Zsolts Cselenyi, Lars Farde, Samuel Svensson, AstraZeneca R&D, Sodertalje, Sweden. Background: The presence of ß-amyloid plaques in the brain is a hallmark of Alzheimer’s disease (AD) and serves as biomarker for confirmation of diagnosis postmortem. Several positron emission tomography (PET) radioligands that bind selectively to ß-amyloid are currently under clinical development for in vivo imaging of brain amyloid. [18F]AV45 (florbetapir F 18) and [18F]Bay94-9172 (florbetaben) are stilbene derivatives, where as [18F]GE067 (F-PIB, flutemetamol) and [18F]AZD4694 have been derived from Thioflavin t. Few studies have evaluated the mutual binding interaction between these two classes of PET amyloid imaging agents in a head-to-head comparison. Methods: b-Amyloid binding capability and cross competition will be evaluated by [3H]AZD4694 and [3H]AV45 binding to synthetic b-amyloid fibrils as well cortical sections from human AD brain. Selectivity of binding to b-amyloid plaques and degree of nonspecific interactions will further be explored using autoradiography in cortical sections from human AD brain as well tg2576 mice. Results: In the present study we will evaluated and compare imaging properties emphasizing on affinity and degree of nonspecific interaction of AZD4694 and AV45 in vitro. Conclusions: Accumulated results show that build up of brain b-amyloid deposits is a very early event in preclinical AD. The ability to successfully determine early changes in b-amyloid build up in the living brain will require PET
Subjects
PREDICTION OF MCI TO ALZHEIMER’S CONVERSION WITH HIPPOCAMPAL SHAPE FEATURES AND SUPPORT VECTOR MACHINE
Jonathan Young1, Gerard Ridgway2, Kelvin Leung2, Josephine Barnes2, Sebastien Ourselin1, 1University College London, London, United Kingdom; 2UCL Institute of Neurology, London, United Kingdom. Background: Hippocampal atrophy is a marker of the onset of Alzheimer’s disease (AD) and hippocampal volumetry has been used in a number of studies to provide early diagnosis of AD and predict conversion of Mild Cognitive Impairment (MCI) to AD. However rates of atrophy are not uniform across the hippocampus, making shape a potentially more sensitive biomarker than volume. This study investigated the utility of SPHARM parameterizations of hippocampal shape coupled with a support vector machine (SVM) classifier to predict conversion from MCI to AD. Methods: Left and right hippocampi from a total of 330 images of MCI patients from the ADNI database were automatically segmented. The resulting volumes were then processed with to produce a description of each shape as a weighted sum of spherical harmonic basis functions. The weights of each shape were taken as a feature vector for
Training 1
Testing 2
82 MCI-converter, 138 MCI-stable 74% 0.8 (68%, 80%)
41 MCI-converter, 69 MCI-stable 68% 0.74 (58%, 77%)
Area under curve Classification accuracy 95% confidence interval on accuracy (1): Accuracy and AUC calculated from ten-fold cross validation on all 220 subjects. Best accuracy found when using 143 hippocampal shape features for classification, best AUC found when using 170 hippocampal shape features for classification. (2): Parameters yielding best AUC in training were used to build an SVM classifier from all 220 training subjects. This was then applied to the hitherto unused 110 remaining subjects to test the performance of the classifier on new data.
SVM classification. Separate classifiers were built for right and left hippocampal shape features and for a concatenation of the two. For each experiment, two thirds of the subjects were used to build a classifier. The shape features were ranked according to their ability to separate MCI converter and MCI stable subjects, and then a ten fold cross validation was performed for each classifier to find the number of hippocampal features and SVM parameters yielding the best classification results. The resulting classifiers were then applied to an unseen test set comprising the remaining third of the subjects. Results: Performance with features from the right hippocampus only was superior to left only or both together. As AUC is considered a better measure of classification efficacy than accuracy, the parameters yielding the best AUC were used when classifying unseen data. In this case classification accuracy was 68% with a binomial (exact) 95% confidence interval of 58% to 77% and an AUC of 0.74. Details are given in the table. Conclusions: This method shows promise in discriminating between MCI stable and MCI converter subjects. The best classification results are obtained with a somewhat biased classifier, meaning that specificity was substantially better than sensitivity, however the AUC shows that this classifier can potentially be both accurate and balanced.
IC-P-080 IC-P-079
S41
PREFRONTAL LATERALIZATION DURING A VERBAL FLUENCY TASK PREDICTS NEUROPSYCHOLOGICAL CHANGES IN HEALTHY ELDERLY CONTROLS: A FNIRS STUDY
Julia Zeller1, Thomas Polak1, Andreas Fallgatter2, 1University W€urzburg, W€urzburg, Germany; 2University T€ubingen, T€ubingen, Germany. Background: Near-Infrared Spectroscopy (NIRS) is a non-invasive optical method measuring brain activation by assessing changes in the concentration of oxy- [O2Hb] and deoxygenated [HHb] haemoglobin in the cerebral cortex. During a verbal fluency task (VFT) healthy controls display a typical pattern of activation in prefrontal regions with a distinct lateralisation to the left dorso-lateral prefrontal cortex (DLPFC). This lateralisation decreases with age indicating compensatory processes. We expected the amount of lateralisation during the VFT to predict neuropsychological changes in healthy elderly controls over a one year interval. Methods: 44 elderly healthy controls (aged 55-82 years) were measured during two versions of a VFT (phonological and semantical) using 44-channel functional NIRS. MMST and DemTect were performed at the time of the fNIRS
S42
Alzheimer’s Imaging Consortium: IC-P-Poster Imaging
measurement and again after one year. Results: During both versions of the VFT subjects displayed a typical activation pattern in the DLPFC with a distinct increase in O2Hb. Subjects’ lateralisation did not correlate with age. The amount of lateralisation during the letter version could account for a significant amount of the changes in MMST (16%) and DemTect (9%) after one year. Subjects displaying less distinct or no lateralisation showed a decrease in neuropsychological performance whereas subjects with a clearly lateralized activation did not change in MMST or DemTect. Conclusions: fNIRS seems to be a suitable tool for the study of long term changes in elderly subjects when assessing brain activation and cognitive decline. As alternations in oxygenation probably occur even before atrophy in affected brain regions, fNIRS could add to the development of methods for the early detection of mild cognitive impairment or Alzheimer’s disease.
IC-P-081
NEAR-INFRARED SPECTROSCOPY: A RELIABLE TOOL TO ASSESS CORTICAL CHANGES IN ALZHEIMER’S DISEASE AND THEIR PROGRESSION OVER TIME
Julia Zeller1, Thomas Polak1, Andreas Fallgatter2, 1University W€urzburg, W€ urzburg, Germany; 2University, T€ubingen, Germany. Background: The technique of Near-Infrared Spectroscopy (NIRS) is a non-invasive optical method allowing the in vivo measurement of brain activation via changes in the concentration of oxy- [O2Hb] and deoxygenated [HHb] haemoglobin in the cerebral cortex. Neuroimaging studies show widespread activation deficits in patients suffering from Alzheimer’s disease (AD) during tasks targeting the visuo-spatial and executive domains located in prefrontal and parietal areas. Using NIRS the present study assessed changes in prefrontal and parietal oxygenation in AD patients and healthy controls and their progression over a one-year interval. Methods: 71 Patients suffering from AD and 68 elderly healthy controls were measured during two versions of a verbal fluency task (VFT, letter and category) and a line orientation task (LO) using 44-channel (VFT) and 52-channel (LO) fNIRS. 14 patients and 49 controls completed a follow-up measurement after one year. Results: During the letter version of the VFT and the LO AD-patients displayed a reduced activation pattern in the prefrontal (VFT) and the parietal cortex (LO) as compared to healthy controls. Increases in O2Hb in parietal regions correlated with prefrontal activity in AD-patients but remained independent in healthy controls. Prefrontal activation decreased significantly over the one year interval in ADpatients. Healthy controls showed no changes at follow up. Activation decreases in the prefrontal cortex were correlated with the decrease of the MMSE score, with greater decrease indicating lower MMSE score at the follow-up. Conclusions: fNIRS seems to be a well suited instrument to measure disease related changes in the cortex of AD patients and their progression over time. Considering the fact that we did not observe any changes in healthy controls without any cognitive decline fNIRS could be a promising device to be used in prospective studies of AD and other dementias.
IC-P-082
3-D RECONSTRUCTION OF CT IMAGES BY LABVIEW
Pravin Jogdand, VJTI, Mumbai, India.
Background: CT scan imaging is popular and plays a unique role in clinical diagnosis and treatment because it is a noninvasive and real-time. Diseases like Alzheimer’s need to be diagnosed at early stages only so that their cure is possible. For this resolution up to micron level is required for analyzing 2D and 3D image. If we try to improve image resolution then cost of equipment increases. So there is trade of between cost of equipment andresolution of image. Methods: This trade of can be overcome by using LabVIEWwith Biomedical start up kit 3.0. In this the 2D slices of DIOCOM images obtained from CT scanmachine is processed for increasing resolution. This process does not require to change the module of equipment. The LabVIEW is a graphical programming environment, which can be used to built data acquisition and instrument control application. Results: This paper introduces the experiment of 3D reconstruction CT images via Virtual Instrumentation - LabVIEW. The main idea is based on marching cubes algorithm and image processing implemented by module with the help of NI Biomedical Startup Kit.The two dimensional images shot by the CT scan device provide information about the surface properties of human body. There is implemented algorithm which can be used for 3D reconstruction of CT images in biomedical application. Images of array 2D slices from CT scan machine can be saved in CD or hard disc and can be imported to study 3D module of image for diagnosis. Conclusions: In this paper we have presented method for 3D reconstruction of biomedical images from CT. The presented method is implemented in program LabVIEW. Presented CT images of head were obtained from Visible Human. The main addition of this work is implementation of mentioned method in program LabVIEW and Biomedical Start up kit 3.0(Biomedical Workbench). Experimental result in LabVIEW is comparable to conventional program as Matlab, Visualization tool kit(VTK) with respect to resolution and image handling capability. The disadvantage of mentioned method is processing time is more as compared to other algorithm and software.
IC-P-083
BETA-AMYLOID BINDING OF [3]AZD4694 AND [3H]AV45 IN VITRO
Fredrik Jeppsson, Anders Jureus, Johan Sandell, Britt-Marie Swahn, Zsolts Cselenyi, Lars Farde, Samuel Svensson, AstraZeneca R&D, Sodertalje, Sweden. Background: The presence of ß-amyloid plaques in the brain is a hallmark of Alzheimer’s disease (AD) and serves as biomarker for confirmation of diagnosis postmortem. Several positron emission tomography (PET) radioligands that bind selectively to ß-amyloid are currently under clinical development for in vivo imaging of brain amyloid. [18F]AV45 (florbetapir F 18) and [18F]Bay94-9172 (florbetaben) are stilbene derivatives, where as [18F]GE067 (F-PIB, flutemetamol) and [18F]AZD4694 have been derived from Thioflavin t. Few studies have evaluated the mutual binding interaction between these two classes of PET amyloid imaging agents in a head-to-head comparison. Methods: b-Amyloid binding capability and cross competition will be evaluated by [3H]AZD4694 and [3H]AV45 binding to synthetic b-amyloid fibrils as well cortical sections from human AD brain. Selectivity of binding to b-amyloid plaques and degree of nonspecific interactions will further be explored using autoradiography in cortical sections from human AD brain as well tg2576 mice. Results: In the present study we will evaluated and compare imaging properties emphasizing on affinity and degree of nonspecific interaction of AZD4694 and AV45 in vitro. Conclusions: Accumulated results show that build up of brain b-amyloid deposits is a very early event in preclinical AD. The ability to successfully determine early changes in b-amyloid build up in the living brain will require PET