基于多模态相关向量回归机的老年痴呆症临床变量预测*

南京大学学报(自然科学版) ›› 2012, Vol. 48 ›› Issue (2): 140–146.

基于多模态相关向量回归机的老年痴呆症临床变量预测*

程波，张道强**

出版日期:2015-05-21 发布日期:2015-05-21
作者简介: (南京航空航天大学计算机科学与技术系，南京，210016)
基金资助:
南京航空航天大学基本科研业务费

Predicting clinical variables in Alzheimer’s disease
based on multimodal relevance vector regression

Cheng Bo，Zhaug Dao一Qiang

Online:2015-05-21 Published:2015-05-21
About author: (Department of Computer Science and Engineering, Nanjing University of
Aeronautics and Astronautics, Nanjing，210016，China)

摘要/Abstract

摘要： 老年痴呆症(Alzheimer’s disease, AD)的临床变量值和多模态特征都是对其内在致病病理的外在反映.木文提出一种多模态相关向量回归机，通过对多模态特征的学习来预测临床变量值.首先采
用核方法将多模态数据融合成一个混合核矩阵，然后使用相关向量回归机对临床变量简易精神状态检查(mini mental state examination, MMSE)和老年痴呆症评定量表(Alzheimer’ s disease assessment
scale, ADAS-Cog)建立回归模型，最后用相关系数和平方根均方误差来验证算法的性能.在标准数据集ADNI上的实验结果表明，木文提出的多模态方法的预测性能优于单模态方法.

Abstract: Recently, effective and accurate diagnosis disease stage of Alzheimer’s disease(AD) or mild cognitive impammcnt(MCl)has attracted more and more attention. Numerous studies have demonstrated that clinical
variables and multimodal features of AD arc external reflections of the intrinsic disease pathology.This paper proposes a multimodal regression method for estimating disease stage and predicting clinical progression from three
modalities of biomarkers，i, e.，magnetic resonance imaging(MRI)，fluoro-deoxy-glucose-positron emission tomography(FDG-PET)，and cerebrospinal fluid(CSF)biomarkers. Specifically, our multimodal regression
framework includes three key steps; firstly, we use the specific application tool to orginal MRl and FDU-PET images data from the 202 Alzheimer’s disease ncuroimaging initiative(ADNI)subjects. For each preprocessed
original MR or FIWrPF Timage, 93 regions of interest(ROIs) are labeled by an atlas warping algorithm. And then, for each MR or FDU-PET imagc，93 volumetric features are extracted from the 93 ROIs.Therefore, for each
subject，the last features come from 93 features from the MRl image，another 93 features from the PET imagc，and 3 features from the CSF biomarkers which original values are directly used as features. Secondly, multi-modal data
(MRI，PET，and CSF) arc fused into a mixed kernel matrix through the kernel method，and then regression model is constructed by the relevant vector machine regression (RVR) for the clinical variables including mini mental state
examination(MMSE) and Alzheimer’s disease assessment scale(ADAS-Cog). Finally, our multimodal regression method compared with single-modal approaches and bimodal approaches. Moreover, the performance of our
regression model is validated by the correlation coefficient (CORR) and the square root of the mean square error (RMSE).These regression experiment scheme arc tested on the ADNI dataset by 10-fold cross-validation.
Experimental results on the ADNI database show that the prediction performance of our multi-modal RVR approach is superior to the corresponding single-modal approaches and bimodal approaches.

程波，张道强**. 基于多模态相关向量回归机的老年痴呆症临床变量预测*
[J]. 南京大学学报(自然科学版), 2012, 48(2): 140–146.

Cheng Bo，Zhaug Dao一Qiang
. Predicting clinical variables in Alzheimer’s disease
based on multimodal relevance vector regression
[J]. Journal of Nanjing University(Natural Sciences), 2012, 48(2): 140–146.

参考文献

[1]Ye J P，Chen K W, Wu T，et al. Heterogene- ous data fusion for Alzheimer’s disease study. Knowledge Discovery and Data Mining Confer ence. ACM, 2008，1025一1033.
[2]Hinrichs C，Singh V，Xu G et al. MKI. for robust multi-modality AD classification; Ler ture Notes in Computer Science. Springer-Ver- lag,2009，786一79’1.
[3]Davatzikos C，Fan Y，Wu X, et al. Detection of prodromal Alzheimer’s disease via pattern classification of magnetic resonance imaging. Neurobiology of Aging, 2008，29:514一523.
[4]Batmanghclich N, Taskar B, Davatzikos C. A general and unifying framework for feature con- struction, in imag}based pattern classification. Prince J L，Pham D L，Mycrs K J, Information
Processing in Medical imaging. Lecture Notes in Computer Science, Springer-Verlag, 2009， 5636:423一434.
[5]Davatzikosa C，Bhatt P, Shaw L M, et al. Pre- diction of MCl to AD conversion, via MRl，CSF biomarkcrs，and pattern classification. Ncurobi- ology of Aging, 2011，32(12):2322, e19一 2322. e27.
[6]Zhang D Q, Shen D G. Semi-supervised multi- modal classification of Alzheimer’s disease. In- ternational Symposium on Biomedical imaging, 2011，1628一1631.
[7]Zhang D Q, Wang Y，Zhou L，et al. Multimo- dal classification of Alzheimer’s disease and mild cognitive impairment. Neurolmage, 2011,55(3):856~867
[8]Filipovych R，Davatzikos C. Semi-supervised pattern classification of medical images:Appli- canon to mild cognitive impairment(MCI). NeuroImage,2011,55(3):1109~1119
[9]Caselli R J，Dueck A C，Osborne D, et al. Longitudinal modeling of age-rclated memory decline and the APOE epsilon4 effect.The New England Journal of Medicine, 2009，361:255一 263.
[10]Reiman E M, Chen K，Liu X，et al. Fibrillar amyloid-beta burden in cognitively normal peo- ple at 3 levels of genetic risk for Alzheimer’s disease. Proceedings of the National Academy of Sciences of the United States of America, 2009，106:6820一6825.
[11]Twamley E W，Ropacki S A，Yondi M W. Ncuropsychological and ncuroimaging changes in preclinical Alzheimer’s disease. Journal of the international Ncuropsychological Society, 2006，12:707一735.
[12]Stonnington C M, Chu C，Kloppcl S, et al. Predicting clinical scores from magnetic reso nance scans in Alzheimer’s disease. Neurolm age，2010，51(4):1405一1413.
[13]Roselli F,Tartaglione B, Federico F, et al.Rate of MMSE score change in Alzheimer’s dis ease influence of education and vascular risk factors. Clinical Ncurology and Ncurosurgery,2009,111:327~330
[14]Fan Y，Kaufer D, Shen D G. Joint estimation of m ultiple clinical variables of ncurolo gical dis- cases from imaging patterns, IEEE international Symposium on Biomedical imaging. Rotterdam, 2010，852一855.
[15]Wang Y，Fan Y，Yhatta P，et al. High-dimen- sional pattern regression using machine learn- ing; From medical images to continuous clinical variables. Ncurolmagc，2010，50(4): 1519一1535.
[16]Zhang X G, Zhang X G, Wu X B, et al. Waves lets support vector machines robust estimate method to regression function. Journal of Nan- jing University(Natural Sciences)，2006，12 (5): 528-534.(张晓光，张兴敢，吴行标等.
回归函数的小波支持向量机鲁棒估计法.南京大学学报(自然科学),2006,42(5):528-534).
[17]Tipping M E. Sparse Baycsian learning and the relevance vector machine. Journal of Machine Learning R escarch，2001，1;211一244.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed