Abstract

Speaker: Rik Henson, MRC CBU , Cambridge, UK.

Title: Efficient fMRI design for activations and patterns.

Abstract: How to design an efficient fMRI experiment for activation analysis, beta-series regression (BSR) and multi-voxel pattern analysis (MVPA).

I will run through a Matlab livescript (ie interactive notebook) that demonstrates the basic ideas behind statistical power (efficiency) for (within-subject) fMRI analyses. The first part reprises what is already available here: https://imaging.mrc-cbu.cam.ac.uk/imaging/DesignEfficiency, ie explains how to optimise the time between trials (or components of trials) and the ordering of those trials (eg randomised/pseudorandomised/blocked) in order to detect a particular hypothesis (contrast) regarding activation differences between trial-types in a single voxel/ROI. This will be explained from the perspectives of signal processing, mathematics and correlations (whichever you find more intuitive). The second part explains how optimal estimation of activation for single-trials (using LSA , regularised LSA or LSS , eg for BSR ) depends on the ratio of trial variability to scan noise, and any temporal structure in your trial order. The final part explains how optimal estimation of patterns of activity over voxels (for MVPA ) also depends on the spatial covariance of trial variability and scan noise.

Venue: MRC CBU West Wing Seminar Room and Zoom https://us02web.zoom.us/j/82385113580?pwd=RmxIUmphQW9Ud1JBby9nTDQzR0NRdz09 (Meeting ID: 823 8511 3580; Passcode: 299077)