Advertisement

Deep Learning for Cerebellar Ataxia Classification and Functional Score Regression

  • Zhen Yang
  • Shenghua Zhong
  • Aaron Carass
  • Sarah H. Ying
  • Jerry L. Prince
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8679)

Abstract

Cerebellar ataxia is a progressive neuro-degenerative disease that has multiple genetic versions, each with a characteristic pattern of anatomical degeneration that yields distinctive motor and cognitive problems. Studying this pattern of degeneration can help with the diagnosis of disease subtypes, evaluation of disease stage, and treatment planning. In this work, we propose a learning framework using MR image data for discriminating a set of cerebellar ataxia types and predicting a disease related functional score. We address the difficulty in analyzing high-dimensional image data with limited training subjects by: 1) training weak classifiers/regressors on a set of image subdomains separately, and combining the weak classifier/regressor outputs to make the decision; 2) perturbing the image subdomain to increase the training samples; 3) using a deep learning technique called the stacked auto-encoder to develop highly representative feature vectors of the input data. Experiments show that our approach can reliably classify between one of four categories (healthy control and three types of ataxia), and predict the functional staging score for ataxia.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Jung, B.C., Choi, S.I., Du, A.X., Cuzzocreo, J.L., Geng, Z.Z., Ying, H.S., Perlman, S.L., Toga, A.W., Prince, J.L., Ying, S.H.: Principal component analysis of cerebellar shape on mri separates sca types 2 and 6 into two archetypal modes of degeneration. The Cerebellum 11(4), 887–895 (2012)CrossRefGoogle Scholar
  2. 2.
    Jung, B.C., Choi, S.I., Du, A.X., Cuzzocreo, J.L., Ying, H.S., Landman, B.A., Perlman, S.L., Baloh, R.W., Zee, D.S., Toga, A.W., et al.: Mri shows a region-specific pattern of atrophy in spinocerebellar ataxia type 2. The Cerebellum 11(1), 272–279 (2012)CrossRefGoogle Scholar
  3. 3.
    Leow, A.D., Klunder, A.D., Jack Jr., C.R., Toga, A.W., Dale, A.M., Bernstein, M.A., Britson, P.J., Gunter, J.L., Ward, C.P., Whitwell, J.L., et al.: Longitudinal stability of mri for mapping brain change using tensor-based morphometry. Neuroimage 31(2), 627–640 (2006)CrossRefGoogle Scholar
  4. 4.
    Ashburner, J.: Computational anatomy with the spm software. Magnetic Resonance Imaging 27(8), 1163–1174 (2009)CrossRefGoogle Scholar
  5. 5.
    Fan, Y., Shen, D., Gur, R.C., Gur, R.E., Davatzikos, C.: Compare: classification of morphological patterns using adaptive regional elements. IEEE Transactions on Medical Imaging 26(1), 93–105 (2007)CrossRefGoogle Scholar
  6. 6.
    Batmanghelich, N.K., Taskar, B., Davatzikos, C.: Generative-discriminative basis learning for medical imaging. IEEE Transactions on Medical Imaging 31(1), 51–69 (2012)CrossRefGoogle Scholar
  7. 7.
    Bengio, Y.: Learning deep architectures for AI. Foundations and Trends in Machine Learning 2(1), 1–127 (2009)CrossRefzbMATHMathSciNetGoogle Scholar
  8. 8.
    Subramony, S., May, W., Lynch, D., Gomez, C., Fischbeck, K., Hallett, M., Taylor, P., Wilson, R., Ashizawa, T., et al.: Measuring friedreich ataxia: interrater reliability of a neurologic rating scale. Neurology 64(7), 1261–1262 (2005)CrossRefGoogle Scholar
  9. 9.
    Dale, A.M., Fischl, B., Sereno, M.I.: Cortical Surface-Based Analysis I: Segmentation and Surface Reconstruction. NeuroImage 9(2), 179–194 (1999)CrossRefGoogle Scholar
  10. 10.
    Breiman, L.: Random forests. Machine Learning 45(1), 5–32 (2001)CrossRefzbMATHGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Zhen Yang
    • 1
  • Shenghua Zhong
    • 1
  • Aaron Carass
    • 1
  • Sarah H. Ying
    • 2
  • Jerry L. Prince
    • 1
  1. 1.Johns Hopkins UniversityBaltimoreUSA
  2. 2.Johns Hopkins School of MedicineBaltimoreUSA

Personalised recommendations