Advertisement

Detecting Outlying Subjects in High-Dimensional Neuroimaging Datasets with Regularized Minimum Covariance Determinant

  • Virgile Fritsch
  • Gael Varoquaux
  • Benjamin Thyreau
  • Jean-Baptiste Poline
  • Bertrand Thirion
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6893)

Abstract

Medical imaging datasets used in clinical studies or basic research often comprise highly variable multi-subject data. Statistically-controlled inclusion of a subject in a group study, i.e. deciding whether its images should be considered as samples from a given population or whether they should be rejected as outlier data, is a challenging issue. While the informal approaches often used do not provide any statistical assessment that a given dataset is indeed an outlier, traditional statistical procedures are not well-suited to the noisy, high-dimensional, settings encountered in medical imaging, e.g. with functional brain images. In this work, we modify the classical Minimum Covariance Determinant approach by adding a regularization term, that ensures that the estimation is well-posed in high-dimensional settings and in the presence of many outliers. We show on simulated and real data that outliers can be detected satisfactorily, even in situations where the number of dimensions of the data exceeds the number of observations.

Keywords

Outlier detection Minimum Covariance Determinant regularization robust estimation neuroimaging fMRI 

References

  1. 1.
    Aljabar, P., Wolz, R., Srinivasan, L., Counsell, S., Boardman, J.P., Murgasova, M., Doria, V., Rutherford, M.A., Edwards, A.D., Hajnal, J.V., Rueckert, D.: Combining morphological information in a manifold learning framework: application to neonatal MRI. Med. Image Comput. Comput. Assist. Interv. 13, 1 (2010)Google Scholar
  2. 2.
    Gardner, A., Krieger, A., Vachtsevanos, G., Litt, B.: One-class novelty detection for seizure analysis from intracranial EEG. J. Mach. Learn. Res. 7, 1025 (2006)MathSciNetzbMATHGoogle Scholar
  3. 3.
    Gerber, S., Tasdizen, T., Joshi, S., Whitaker, R.: On the manifold structure of the space of brain images. Med. Image Comput. Comput. Assist. Interv. 12, 305 (2009)Google Scholar
  4. 4.
    Grubel, R.: A minimal characterization of the covariance matrix. Metrika 35, 49 (1988)MathSciNetCrossRefzbMATHGoogle Scholar
  5. 5.
    Kherif, F., Flandin, G., Ciuciu, P., Benali, H., Simon, O., Poline, J.B.: Model based spatial and temporal similarity measures between series of functional magnetic resonance images. Med. Image Comput. Comput. Assist. Interv., 509 (2002)Google Scholar
  6. 6.
    Penny, W.D., Kilner, J., Blankenburg, F.: Robust bayesian general linear models. Neuroimage 36, 661 (2007)CrossRefGoogle Scholar
  7. 7.
    Perrot, M., Rivière, D., Tucholka, A., Mangin, J.F.: Joint bayesian cortical sulci recognition and spatial normalization. Inf. Process. Med. Imaging 21, 176–187 (2009)CrossRefGoogle Scholar
  8. 8.
    Rousseeuw, P.J.: Least median of squares regression. J. Am. Stat. Ass. 79, 871 (1984)MathSciNetCrossRefzbMATHGoogle Scholar
  9. 9.
    Rousseeuw, P.J., Hubert, M.: Robust statistics for outlier detection. WIREs Data Mining Knowl. Discov. 1, 73 (2011)CrossRefGoogle Scholar
  10. 10.
    Rousseeuw, P.J., Van Driessen, K.: A fast algorithm for the minimum covariance determinant estimator. Technometrics 41(3), 212 (1999)CrossRefGoogle Scholar
  11. 11.
    Segata, N., Blanzieri, E.: Fast and scalable local kernel machines. J. Mach. Learn. Res. 11, 1883 (2009)MathSciNetzbMATHGoogle Scholar
  12. 12.
    Woolrich, M.: Robust group analysis using outlier inference. Neuroimage 41, 286 (2008)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Virgile Fritsch
    • 1
    • 2
  • Gael Varoquaux
    • 3
    • 1
    • 2
  • Benjamin Thyreau
    • 2
  • Jean-Baptiste Poline
    • 2
    • 1
  • Bertrand Thirion
    • 1
    • 2
  1. 1.Parietal TeamINRIA Saclay-Île-de-FranceSaclayFrance
  2. 2.CEA, DSV, I2BMGif-Sur-YvetteFrance
  3. 3.Inserm, U992Gif-Sur-YvetteFrance

Personalised recommendations