Habib Ganjgahi

Excited to share our new publication!🎉 We developed a probabilistic ML method to better understand the dimensions of MS and how the disease evolves. A multidisciplinary effort with MS specialistis, @Novartis and @Roche. With Dieter Haering, @hwiendl, @cholmesuk #MSresearch

🚨 Paradigm shift in #MultipleSclerosis 🚨 New study in Nature Medicine shows MS is not fixed subtypes but a dynamic continuum of states across 4 dimensions: disability, brain damage, relapses & silent inflammation. 📖 Read: nature.com/articles/s4159… #MS #Neurology #AI

Last week in Barcelona our 38th Plenary meeting was held with great contribution from all our core and affiliate centres. We are very grateful to all participants, congratulate all authors of finished studies and look forward to soon hearing results of ongoing projects!

Excited to be presenting work on Posterior Bootstrap for robust #Bayesian inference later today at #isba2021! @ISBA_events

Have you ever been curious to know how your fancy NN behaves at a single prediction? Then check out our (@TerLucile*,@karlado,@cholmesuk) paper where we improve upon local explanation models (i.e. SHAP) by imposing a neighbourhood distribution as reference arxiv.org/pdf/2106.14648…

Excited to present our work (@ganjgahi, @ten_photos and @cholmesuk) on structured Bayesian variable selection for brain lesion data at ISBA in a few minutes! Come join the Q&A session or the poster session tomorrow! #ISBA2021 @ISBA_events

Come visit Anna Menacher present on a scalable Bayesian approach to modelling lesion data! #OHBM2021 Poster 1341, right now in ohbm.sparkle.space/in/poster1341 w/ @ganjgahi

Coordinate Based Meta Analysis needs meta-regression! ... to compare groups or to control for moderators. See Yifan Yu's work on a practical point process approach for CBMA #OHBM2021 Poster 1557 right now ohbm.sparkle.space/in/poster1557

Stop by poster 2378 at #OHBM2021 to learn about our work on nonparametric causal analysis of brain and cognition, applied to developmental neuroimaging. ohbm.sparkle.space/in/poster2378 Joint work with @arokem, @jdyeatman , @StatsSimon and @richiehalford

Hi #OHBM2021 people. Check out our work (poster #1341) on a Bayesian method for binary lesion regression modeling by @MenacherAnna and @ten_photos that is scalable to large-scale studies like UK Biobank. ohbm.sparkle.space/in/poster1341

FSLeyes 0.30.0 is out, with support for 2D images (.bmp, .png, .jpg etc), and for RGB(A) NIfTI images. Get it from pypi, conda-forge, or users.fmrib.ox.ac.uk/~paulmc/fsleye…!

Preprint with @rikhens @SimonWhite83 and Linda Geerligs now out on bioRxiv. We use Bayesian ERGMs to characterise the brain's connectivity structure at a group level, modelling multiple brain networks simultaneously. Comments and feedback v welcome! biorxiv.org/content/10.110…

Our paper is finally published in NeuroImage. Ensemble learning with 3D convolutional neural networks for functional connectomes. sciencedirect.com/science/articl… … Pre-print: arxiv.org/abs/1809.06219

"Most of the common statistical models (t-test, correlation, ANOVA; chi-square, etc.) are special cases of linear models or a very close approximation. This beautiful simplicity means that there is less to learn." lindeloev.github.io/tests-as-linea…

There is no single functional atlas even for a single individual: Parcellation of the human brain is state dependent | #bioRxiv biorxiv.org/content/10.110…

Excited about the symposium now in Sinopoli Hall on How Current Predictive Connectivity Models Meet Clinician's needs ! #OHBM2019 starting with @GaelVaroquaux now! 😃 @R3RT0

“All models are wrong but a linear mixed model will do the job most of the time” — attributed to Robin Thompson by Peter Visscher.

At ISBA in Edinburgh for 2 days. Learning from the always amazing Judith Rousseau, about the asymptotic behaviour of credible regions.

@NeuroFidelity on GWA of @uk_biobank neuroimaging phenotypes @OHBM #OHBM2018