Does higher sampling rate (multiband + SENSE) improve group statistics - An example from social neuroscience block...

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Magnetic Resonance Imaging (MRI)

Background:

• Multiband (MB) or Simultaneous Multi-Slice (SMS) acquisition techniques enable simultaneous MRI signal acquisition from multiple spatial locations, increasing efficiency.
• Previous evaluations of MB acquisition primarily focused on resting-state fMRI or simple tasks, limiting understanding of its utility in complex cognitive neuroscience applications.
• The commercial availability of MB acquisition necessitates evaluating its performance in more representative experimental paradigms.

Purpose of the Study:

• To assess whether the benefits of MB acquisition generalize to group-level analyses in functional MRI (fMRI) using a complex cognitive task.
• To compare the performance of different MB acceleration factors (MB1-MB4) and SENSE factors on a Philips 3T scanner.
• To determine optimal MB acquisition parameters for group analyses in cognitive neuroscience research.

Main Methods:

• Twenty-three subjects underwent fMRI scanning on a Philips 3T scanner using five sequences with varying MB (1-4) and SENSE (up to 2) factors, resulting in TRs from 2.45s to 0.63s.
• Participants viewed movie blocks depicting complex actions with hand-object interactions and control blocks without such interactions.
• Data processing utilized the SPM12 pipeline, including unified segmentation and canonical hemodynamic response function (HRF) modeling, followed by random-effects group-level voxel-wise analysis.

Main Results:

• All tested sequences successfully identified the action observation network associated with the task.
• MB4 acceleration yielded the highest group-level t-values for both the cognitive task and resting-state networks (RSNs).
• MB1 sequences required a 50% larger voxel volume to achieve comparable t-statistics to MB4 sequences, and even then, performance did not match MB4.

Conclusions:

• MB4 acceleration, combined with SENSE factors of 1.5 or 2 on a Philips scanner, is recommended for group-level analyses of cognitive block-design fMRI tasks, particularly with voxel sizes around cortical thickness (e.g., 2.7 mm isotropic).
• While not drastically altering results, multiband acquisition provides a moderate but significant benefit for fMRI group analyses.
• Higher MB acceleration factors enhance statistical power in group analyses, making them advantageous for cognitive neuroscience applications.