Optimal experimental designs for fMRI when the model matrix is uncertain.

Area of Science:

• Neuroscience
• Statistics
• Experimental Design

Background:

• Designing functional magnetic resonance imaging (fMRI) experiments is complex when the statistical model depends on both the stimulus sequence and uncertain subject feedback.
• Evaluating the quality of fMRI designs is challenging because the information matrix is not fully determined during the design phase.

Purpose of the Study:

• To propose an easy-to-use optimality criterion for evaluating fMRI designs.
• To develop an efficient approach for obtaining optimal fMRI designs under uncertainty.

Main Methods:

• Development of a novel optimality criterion for fMRI design evaluation.
• Implementation of an efficient computational approach to identify optimal designs.

Main Results:

• The proposed method provides an accessible way to assess fMRI design quality.
• The new approach significantly reduces the computing time needed to achieve statistically efficient fMRI designs compared to previous methods.

Conclusions:

• The developed criterion and method offer a practical solution for optimizing fMRI experimental designs with uncertain subject feedback.
• This work enhances the efficiency and reduces the computational burden in creating high-quality fMRI studies.