fMRI Methods

To download software developed by the Cusack Lab, click here


Ultra-high Field Strength MRI

with Jacob Matthews and Joe Gati

We are lucky to have one of a few 7T MRI scanners in the world and a group of excellent MR physicist to work with at the Robarts Research Institute. Using ultra-high resolution structural and functional imaging, we are investigating how different layers of auditory cortex code sounds during perception and cognition, and how this changes over the first year of life as thalamocortical connections emerge and the subplate disappears. 

Click here for information about the first Robarts High-Field MRI Winter school. 


Optimizing EPI protocols 

In order to optimize BOLD SNR for auditory fMRI in infants, children and adults we have undergone thorough testing of different cutting-edge EPI sequences (such as multiband and multiecho EPI) and protocols (such as sparse vs. continuous data acquisitions).


I am currently investigating the impact of motion - the most ubiquitous problem for fMRI in non-compliant populations such as infants and children - on highly accelerated EPI sequences.  


Motion artifacts in a 3-month old infant (MPRAGE)

Motion artifacts in a 3-month old infant (MPRAGE)

Multi-voxel Pattern Analysis

In contrast to conventional modeling of fMRI data, multi-voxel pattern analysis allows studying informational content of activity patterns evoked by different conditions. I am using one MVPA method - Representational Similarity Analysis - to simultaneously probe many features of complex sounds simultaneously in order to investigate how neural representations in auditory cortex of the change when people engage in different cognitive task.

See our recent chapter in Attention & Performance for more detail! 



Amazon Mechanical Turk 

We make extensive use of amazon mechanical turk to test stimuli and run simple experiments and pilot studies. This is a great resource that has made our lives easier and accelerated experimental design many times!


Automatic Analysis Toolbox

Developed by Rhodri Cusack and constantly in use and being updated by the lab to automate and parallelise fMRI analysis; integrates with SPM; univariate and multivariate fMRI analysis.