Epoch versus impulse models in the analysis of parametric fMRI studies

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Data Analysis

Background:

• Parametric functional magnetic resonance imaging (fMRI) commonly uses the general linear model (GLM) to relate hemodynamic response amplitude to behavioral or electrophysiological parameters.
• Analyzing single-trial response time (RT) offers a method to better isolate task-specific activation in fMRI studies.

Purpose of the Study:

• To investigate methods for using single-trial response time (RT) within the general linear model (GLM) framework to enhance the isolation of task-specific activation in parametric fMRI.
• To compare the effects of different explanatory variable (EV) strategies, including conventional and parametrically informed approaches, on fMRI activation patterns.

Main Methods:

• fMRI and RT data were collected from 21 subjects performing a visual-oddball task.
• Four EVs were generated: constant impulse, average RT-informed constant epoch, single-trial RT-informed variable impulse, and single-trial RT-informed variable epoch.
• EVs were analyzed individually and as orthogonalized pairs to assess their impact on BOLD signal detection.

Main Results:

• Individual EVs revealed similar extensive activation patterns.
• Orthogonalized EVs showed correlations with BOLD signal variance primarily in sensorimotor and parietal areas.
• Orthogonalizing the variable epoch EV to the constant epoch EV successfully isolated the dorsal frontoparietal attention network.

Conclusions:

• For short event durations, individual EVs yield comparable activation patterns.
• Orthogonalized, parametrically informed EVs enhance the isolation of task-specific BOLD signal changes.
• Varying parametric timing integration methods in fMRI analyses can significantly influence, and potentially refine or confound, study findings.