Advances in Data Analysis and Classification

, Volume 7, Issue 1, pp 109–119

Regularized logistic discrimination with basis expansions for the early detection of Alzheimer’s disease based on three-dimensional MRI data

Regular Article

DOI: 10.1007/s11634-013-0127-5

Cite this article as:
Araki, Y., Kawaguchi, A. & Yamashita, F. Adv Data Anal Classif (2013) 7: 109. doi:10.1007/s11634-013-0127-5


In recent years, evidence has emerged indicating that magnetic resonance imaging (MRI) brain scans provide valuable diagnostic information about Alzheimer’s disease. It has been shown that MRI brain scans are capable of both diagnosing Alzheimer’s disease itself at an early stage and identifying people at risk of developing Alzheimer’s. In this article, we have investigated statistical methods for classifying Alzheimer’s disease patients based on three-dimensional MRI data via L2-type regularized logistic discrimination with basis expansions. Preceding studies adopted an open approach when applying three-dimensional data analysis. Our proposed classification model with dimension reduction techniques offers discriminant functions with excellent prediction performance in terms of sensitivity and specificity.


Regularized logistic discriminationThree-dimensional MRI dataAlzheimer’s diseaseDimension reductionTuning parameter selection

Mathematics Subject Classification (2000)


Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Yuko Araki
    • 1
  • Atsushi Kawaguchi
    • 1
  • Fumio Yamashita
    • 2
  1. 1.Biostatistics CenterKurume UniversityKurumeJapan
  2. 2.Division of Ultrahigh Field MRIInstitute for Biomedical Sciences, Iwate Medical UniversityMoriokaJapan