Farouk Salim Nathoo

PhD, Statistics, Simon Fraser University

MMath, Statistics, University of Waterloo 

BSc, Combined Honours, Mathematics and Statistics, University of British Columbia  


 
                                    
                         

Associate Professor

Canada Research Chair in Biostatistics (Tier 2, 2013 - 2023)

Department of Mathematics and Statistics, University of Victoria

Victoria, BC, Canada, V8W 3P4

 

Adjunct Professor

Department of Statistics and Actuarial Science, Simon Fraser University

Burnaby, BC, Canada V5A 1S6

Curriculum Vitae

Short Professional Biography

Office: DTB (formerly SSM) A-545
Phone: 250-472-4693
 Fax: 250-721-8962
 nathoo at math dot uvic dot ca

Research Program:

We focus generally on biostatistics and specifically on the development of statistical methods for the analysis of neuroimaging data and the joint analysis of neuroimaging and genetic data. Within this context we apply and develop approaches for statistical modelling and computation.  A secondary focus is the development of Bayesian methods for cognitive science. Our group is receptive to starting new collaborations with researchers involved with complex biomedical data problems requiring statistical expertise in the analysis of existing data, the development of new methods for novel data or statistical problems, and in the design of experiments.

Editorial Service:

2016 - present: Associate Editor, Canadian Journal of Statistics

2013: Guest Co-Editor, Statistical Methods in Medical Research, Special Issue on Statistics for Imaging. This special issue contains five papers focussed on neuroimaging statistics: Bayesian nonparametric modeling of fMRI data; predicting brain activity with positron-emission tomography using a Bayesian spatial model; Bayesian computation for neuroimaging within the  context of the electromagnetic inverse problem; spectral analysis of resting-state magnetoencephalography data with space-varying regression; constructing shape features and detecting hippocampal shape changes in early Alzheimer’s using support vector machines.

Publications:

    Selected Papers:

       Trainees indicated with '*'.
      

    Selected Software:

  1. R package 'bgsmtr' for Bayesian regression analysis of imaging genetics data. This R package implements the methodology in Greenlaw et al. (2017) and Song et al. (2019). The package has options for both spatial and non-spatial error covariance models and as options for computational implementation based on both MCMC and mean-field variational Bayes. Bayesian false discovery rate procedures are also implemented in the latest version of the package. 
  2. R software 'Pottsmix' for fitting the latent Gaussian spatiotemporal mixture model for solving the electromagnetic inverse problem based on combined MEG and EEG data.  This implements the iterated conditional modes algorithm for simultaneous point estimation and model selection developed in Song, Nathoo and Babul (2019).
  3. Christoph Huber has written a nice Shiny App implementing the Bayes factor approximations for repeated measures designs developed in Nathoo and Masson (2015).
  4. Matlab codes for fitting the spatial spike-and-slab model for solving the electromagnetic inverse problem based on combined MEG (or EEG) and fMRI data. The software implements the mean-field variational Bayes algorithm developed in Nathoo et al. (2014).
  5. C++ Codes for Hamiltonian Monte Carlo, mean-field variational Bayes and integrated nested Laplace approximation approaches for fitting Log-Gaussian Cox processes based on Teng, Nathoo and Johnson (2017).
  6. R software and examples for mixed effects repeated measures designs with binary response developed in Song, Nathoo and Masson (2017).
  7. Python codes for feature selection methods (t-test, lasso, PCA, stacked autoencoder) discussed in Shi and Nathoo (2018).

Recent Lab News:

I am currently recruiting PhD and MSc students to work on a number of projects involving Bayesian methods and neuroimaging data. Potential applicants should have very strong programming skills (we emphasize R and Matlab but some students use Python), a good undergraduate math background. Interested candidates should contact Farouk Nathoo (nathoo@uvic.ca), send me your undergraduate and graduate transcripts, your publications if any, and research interests. A formal application to the Department of Mathematics and Statistics, University of Victoria is required in all cases.

Student Supervision:

    Current Lab Members:



 Yin presenting an e-poster at JSM 2018.





Li presenting an e-poster at JSM 2018.

Teaching:

In Fall 2018 I am teaching MATH/STAT 452/552 (Stochastic Processes/Applied Stochastic Models; course outline) and STAT 460/560 (Bayesian Statistics; course outline). Course material will be distributed in class.

CANSSI Collaboration:

I am a co-leader of the Canadian Statistical Sciences Institute Collaborative Research Team: "Joint Analysis of Neuroimaging Data: High-dimensional Problems, Spatio-Temporal Models and Computation" 2016 - 2019.

  

    

Conference Organization: