Multimodal surface matching: Fast and generalisable cortical registration using discrete optimisation

Emma C. Robinson, Saad Jbabdi, Jesper Andersson, Stephen Smith, Matthew F. Glasser, David C. Van Essen, Greg Burgess, Michael P. Harms, Deanna M. Barch, Mark Jenkinson

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

26 Citations (Scopus)


Group neuroimaging studies of the cerebral cortex benefit from accurate, surface-based, cross-subject alignment for investigating brain architecture, function and connectivity. There is an increasing amount of high quality data available. However, establishing how different modalities correlate across groups remains an open research question. One reason for this is that the current methods for registration, based on cortical folding, provide sub-optimal alignment of some functional sub-regions of the brain. A more flexible framework is needed that will allow robust alignment of multiple modalities. We adapt the Fast Primal-Dual (Fast-PD) approach for discrete Markov Random Field (MRF) optimisation to spherical registration by reframing the deformation labels as a discrete set of rotations and propose a novel regularisation term, derived from the geodesic distance between rotation matrices. This formulation allows significant flexibility in the choice of similarity metric. To this end we propose a new multivariate cost function based on the discretisation of a graph-based mutual information measure. Results are presented for alignment driven by scalar metrics of curvature and myelination, and multivariate features derived from functional task performance. These experiments demonstrate the potential of this approach for improving the integration of complementary brain data sets in the future.

Original languageEnglish
Title of host publicationInformation Processing in Medical Imaging - 23rd International Conference, IPMI 2013, Proceedings
Number of pages12
Publication statusPublished or Issued - 12 Jul 2013
Externally publishedYes
Event23rd International Conference on Information Processing in Medical Imaging, IPMI 2013 - Asilomar, CA, United States
Duration: 28 Jun 20133 Jul 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7917 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other23rd International Conference on Information Processing in Medical Imaging, IPMI 2013
CountryUnited States
CityAsilomar, CA

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this