T1-weighted MRI data of 10 normal subjects (SIEMENS 1.5T, image size:
256×256×128, resolution: 1.17mm × 1.17mm× 1.25mm) were used to validate adMCFC. Fig. 8 showed the results from both adMCFC and FCM on the MR image of one of the 10 subjects. The bias field and corresponding misclassifications can be easily detected at the top part in the original MR image and segmentation results of FCM. But adMCFC yielded a much better result. Such improvement can also be demonstrated with 3D rendering of the segmented WM in Fig 9 where WM loss is very significant at the top area in the results of FCM.
5 Discussions and Conclusion
In this work we have qualitatively described the requirements of our LTD model and presented a improved MCFC method to separated brain tissue in T1-weighted MR images with more accuracy than MCFC as well as other related methods in the condi- tion of bias field and biophysical properties variations. It is difficult for a fixed con- text size to guarantee the assumptions of LTD in all contexts because of the complex- ity of the brain. While adMCFC can determine the size of a context according to its anatomic position to result in a lower misclassification rate than original MCFC can do.
There are several issues to study further to improve the performance of adMCFC, such as the relationship between context size and enlarging coefficient, the shape and size of the mask to enlarge context. Please note that, we corrected a minor mistake in the MCFC algorithm so that we obtained slightly different results in Table 1 and Fig.
2 from those in [11].
202 C.Z. Zhu et al.
Fig. 8. Segmentation results. Original images (first row); FCM results (second row) and adMCFC results (third row)
Fig. 9. 3D rendering of segmented WM from two view angles (top and bottom row).
FCM results (left column) and adMCFC results (right column)
Anatomy Dependent Multi-context Fuzzy Clustering 203
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Arnold, J. B., Liow, J.-S., Schaper, K. A., et al. Qualitative and Quantitative Evaluation of Six Algorithms for Correcting Intensity Nonuniformity Effects. NeuroImage (2001) 13:
931–934.
Clarke, L. P., Velthuizen, R. P., Camacho, M. A., Heini, J. J., Vaidyanathan, M, Hall, L.
O., Thatcher, R. W. and Silbiger, M. L. MRI Segmentation: Methods and Applications.
Mag. Res. Imag. (1995) 13: 343-368.
Collins, D. L., Zijdenbos, A. P., Kollokian, V., Sled, J. G.., Kabani, N.J., Holmes,C.J., Evans, A.C. Design and Construction of a Realistic Digital Brain Phantom. IEEE Trans.
Med. Imag. (1998) 17: 463–468.
Guillemaud, R. and Brady, M. Estimating the Bias Field of MR Images. IEEE Trans.
Med. Imag. (1997). 16: 238-251.
Likar B., Viergever, M. A. and Pernus F., Retrospective Correction of MR Intensity In- homogeneity by Information Minimization, IEEE Trans. Med. Imag. (2001). 20: 1398- 1410.
Pham, D. L. and Prince J. L. Adaptive Fuzzy Segmentation of Magnetic Resonance Im- ages. IEEE Trans. Med. Imag. (1999) 18: 737-752.
Shattuck, D. W., Sandor-Leahy, S. R., Schaper, K. A., Rottenberg, D. A. and Leahy, R.
M. Magnetic Resonance Image Tissue Classification Using a Partial Volume Model.
NeuroImage. (2001) 13: 856-876.
Sled J.G., Zijdenbos A.P., and EvansA.C., A Nonparametric Method for Automatic Cor- rection of Intensity Nonuniformity in MRI Data, IEEE Trans. Med. Imag. (1998). 17: 87- 97.
Suri, J. S., Setarehdan, S. K. and Singh, S. Advanced Algorithmic Approaches to Medical Image Segmentation: state-of-the-art applications in cardiology, neurology, mammogra- phy and pathology. Spring-Verlag London Berlin Heidelberg, (2002).
Worth, A. J., Markis, N., Caviness, V. S. and Kennedy, D. N. et al. Neuroanatomical segmentation in MRI: Technological Objectives. International J. of Pattern Recognition and Artificial Intelligent (1997) 11:1161-1187.
Zhu C.Z.and Jiang T. Z., “Multicontext fuzzy clustering for separation of brain tissues in magnetic resonance images”, NeuroImage, Volume 18, Issue 3, (2003) Pages 685-696.
Visual Search in Alzheimer’s Disease
— fMRI Study
Jing Hao1, Kun-cheng Li1, Ke Li2, De-xuan Zhang3, Wei Wang1, Bin Yan2, Yan-hui Yang1, Yan Wang3, Qi Chen3, Bao-ci Shan2, and Xiao-lin Zhou3
1 Department of Radiology, Xuanwu Hospital, Capital University of Medical Sciences, Beijing 100053, China
2 Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100039, China
3 Department of Psychology, Peking University, Beijing 100034, China
Abstract. The aim was to investigate the neural basis of visual at- tention deficits in Alzheimer’s disease (AD) patients using functional MRI. Thirteen AD patients and 13 age-matched controls participated in the experiment of two visual search tasks, one was a pop-out task, the other was a conjunction task. The fMRI data were collected on a 1.5T MRI system and analyzed by SPM99. Both groups revealed almost the same networks engaged in both tasks, including the superior pari- etal lobule (SPL), frontal and occipito-temporal cortical regions (OTC), primary visual cortex and some subcortical structures. AD patients have a particular impairment in the conjunction task. The most pronounced differences were more activity in the SPL in controls and more activity in the OTC in AD patients. These results imply that the mechanisms controlling spatial shifts of attention are impaired in AD patients.
1 Introduction
Alzheimer’s disease (AD) was considered as a dementia characterized by global cognitive impairment. Amnesia has long been recognized as a primary manifes- tation and is the core symptom for the clinical diagnosis of probable AD [1].
However, there has been a suggestion that attention is impaired early in the course of AD [2]. Until recently, there has been a relative paucity of experimen- tal studies about attentional functions in AD. Attentional impairments have been revealed in many studies of attentional capacities including both auditory [3] and visual [4] selective processing, visual search [5] and attention shifting [6].
However, other studies have showed no marked deficits in detecting, shifting to and engaging target items [7,8]. There is therefore a continuing debate concern- ing the status of attentional functions in AD patients. In the present study, we used computer-presented visual search task to examine whether an attentional deficit exists in AD, We also intended to investigate the neural basis of visual attention deficits with functional magnetic resonance imaging (fMRI).
G.-Z. Yang and T. Jiang (Eds.): MIAR 2004, LNCS 3150, pp. 204–212, 2004.
© Springer-Verlag Berlin Heidelberg 2004
Visual Search in Alzheimer’s Disease — fMRI Study 205
2 Method