Interests

• Random Field Theory
• Brain Imaging
• Multiple Testing
• Selective Inference

About me

I am a 4th year PhD student at the Department of Statistics at the University of Oxford on the Oxford-Warwick Statistics Programme (OxWaSP) supervised by Professors Thomas Nichols and Armin Schwartzman and am affiliated with the Big Data Institute. I work on developing statistical methods with applications in Neuroimaging, Astrophysics and beyond. For more details see my CV. If you're interested in collaborating or have a question about my work please feel free to email me (samuel.davenport(AT)stats.ox.ac.uk).

Research Interests

My interest in Random Field Theory (RFT) stems from the huge potential for application across a variety of fields and the beauty of the underlying mathematics. Recently this has been a matter of controversy in the neuroimaging community, primarily because RFT methods have been applied widely without proper regard for the underlying assumptions. However it turns out that a number of these assumptions are unnecessary and can be dropped (by modifying the theory). My work focuses on developing RFT (and other) methods that provide correct control of the false positive rates and allow for better, more powerful inference.

Big Data Validation

In general I am keen to encourage rigorous testing of statistical methods in brain imaging using large datasets. My work on selective inference has looked at providing selection corrected estimates for the effect size in fMRI and VBM data. This has been tested using a validation framework that uses data from the UK biobank and involves setting aside a large number of subjects to calculate a ground truth and dividing the remaining subjects into small groups (of sizes representative of those used in typical brain imaging studies) in order to compare methods. It is vital that all existing and emerging statistical methods be validated in this manner so that we can ensure that our methods work as well as is claimed. For more details see our paper.