Interpolation of diffusion weighted imaging datasets Tim B. Dyrby a, ⁎, Henrik Lundell a , Mark W. Burke b , Nina L. Reislev a , Olaf B. Paulson a,c , Maurice Ptito a,d,e , Hartwig R. Siebner a a Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark b Department of Physiology and Biophysics, Howard University, Washington, DC, USA c Neurobiology Research Unit, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark d School of Optometry, University of Montreal, Montreal, Canada e Department of Neuroscience and Pharmacology, Copenhagen University, Copenhagen, Denmark abstract article info Article history: Accepted 3 September 2014 Available online 16 September 2014 Keywords: Cortical layers Hippocampus Histology Diffusion MRI DTI Tractography Validation Regularisation Image resolution Diffusion weighted imaging (DWI) is used to study white-matter fibre organisation, orientation and structural connectivity by means of fibre reconstruction algorithms and tractography. For clinical settings, limited scan time compromises the possibilities to achieve high image resolution for finer anatomical details and signal-to-noise-ratio for reliable fibre reconstruction. We assessed the potential benefits of interpolating DWI datasets to a higher image resolution before fibre reconstruction using a diffusion tensor model. Simulations of straight and curved crossing tracts smaller than or equal to the voxel size showed that conventional higher- order interpolation methods improved the geometrical representation of white-matter tracts with reduced partial-volume-effect (PVE), except at tract boundaries. Simulations and interpolation of ex-vivo monkey brain DWI datasets revealed that conventional interpolation methods fail to disentangle fine anatomical details if PVE is too pronounced in the original data. As for validation we used ex-vivo DWI datasets acquired at various image resolutions as well as Nissl-stained sections. Increasing the image resolution by a factor of eight yielded finer geometrical resolution and more anatomical details in complex regions such as tract boundaries and cortical layers, which are normally only visualized at higher image resolutions. Similar results were found with typical clinical human DWI dataset. However, a possible bias in quantitative values imposed by the interpolation method used should be considered. The results indicate that conventional interpolation methods can be successfully applied to DWI datasets for mining anatomical details that are normally seen only at higher resolutions, which will aid in tractography and microstructural mapping of tissue compartments. © 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Introduction Practical restrictions in the available scanning time impose major constrains on clinically applied diffusion weighted imaging (DWI) in terms of image resolution and signal-to-noise ratio (SNR). A large voxel size and under-sampled data sets hamper robust fibre and tract reconstruction. Severe partial volume effects (PVE) in the reconstruc- tion render it difficult to correctly resolve crossing and bending fibres as well as interfacing fibre bundles. This limits the analysis of microstructural features e.g. diffusion tensor analysis and reliability of tractography (Oouchi et al., 2007). Several super resolution (SR) approaches have been suggested to improve the anatomical detail beyond the resolution of the originally acquired DWI data while maintaining SNR. The main idea behind SR is to obtain partially overlapping low resolution image volumes of the same object which then are interpolated and combined (reconstructed) to provide a single higher image resolution. Greenspan et al. (2002) acquired multiple image volumes with thick but overlapping slices and could thereby reconstruct thinner slices from the overlapping infor- mation while maintaining SNR. Similar attempts have recently been presented where multiple images with anisotropic voxels are acquired with multiple orientations (Poot et al., 2012; Scherrer et al., 2012). Such super resolution methods use adapted acquisition protocols and non-standard reconstruction methods, and can therefore not be applied to already existing DWI data, which limits their use in clinical settings. In contrast to SR techniques, image regularisation techniques do not change the final image resolution of the acquired data. Image regularisation reconstruction techniques use spatial dependence between neighbouring voxels to obtain a more robust result in terms of SNR. For example, in regions with coherent structures such as brain WM fibre bundles, image regularisation is a promising technique for improving the noise properties in DTI (Arsigny et al., 2006; Castaño-Moraga et al., 2004). Interpolation of DWIs is commonly used NeuroImage 103 (2014) 202–213 ⁎ Corresponding author at: Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital Hvidovre, Denmark. Fax: +45 3647 0302. E-mail address: timd@drcmr.dk (T.B. Dyrby). http://dx.doi.org/10.1016/j.neuroimage.2014.09.005 1053-8119/© 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). Contents lists available at ScienceDirect NeuroImage journal homepage: www.elsevier.com/locate/ynimg