j.p.piatkowski at sms.ed.ac.uk
10 Crichton St
EH8 9AB, Edinburgh, UK
I am currently awaiting formal PhD graduation following the approval of my thesis corrections. I have been funded through the Neuroinformatics and Computational Neuroscience Doctoral Training Centre (DTC) and a member of the Institute of Adaptive and Neural Computation (IANC) at the University of Edinburgh.
What I did (PhD)
In paticular, I was interested in estimating the volume of these pathways with precision greater than the acquisition voxels. This requires the removal of ambiguity associated with partial volume effects, where more than one tissue type is present in a voxel.
In order to make this estimation as robust and efficient as possible, I looked into several classes of Bayesian machine learning methods:
- Markov chain Monte Carlo sampling techniques,
- models that facilitate the sharing of information between neighbouring sites and
- methods for doing inference on the diffusion equation itself, effectively changing the forward model.
Given that the brain images we work with typically comprise 106 voxels, I used distributed computation and ran my code on Eddie.
List of publications
Mark E. Bastin, Jakub P. Piatkowski, Amos J. Storkey, Laura J. Brown, Alasdair M.J. MacLullich, and Jonathan Clayden, "Tract shape modelling provides evidence of topological change in corpus callosum genu during normal ageing", NeuroImage, 2008, 43(1), 20-28. [Elsevier]
Peer-reviewed conference posters
Jakub P. Piatkowski, Mark E. Bastin, Amos J. Storkey, "Estimating white matter tract volume in partial volume voxels with diffusion MRI" at the International Society for Magnetic Resonance in Medicine (ISMRM) 17th International Meeting and Exhibition, 2009.
Jakub P. Piatkowski, Mark E. Bastin, Amos J. Storkey, "Discovering white matter structure beyond anisotropy mapswith diffusion MRI" at the International Society for Magnetic Resonance in Medicine (ISMRM) and European Society for Magnetic Resonance in Medicine and Biology (ESMRMB) Joint Annual Meeting, 2010.
Machine Learning and Pattern Reckognition (MLPR), see course homepage.
Introductory Applied Machine Learning (IAML), see course homepage.
Software Engineering with Objects and Components (SEOC), see course homepage.
On behalf of the Informatics Graduate School, I co-organised the biannual Firbush outing for postgraduate students and members of staff.