Comparing Hemodynamic Models with DCM
Author Information
Author(s): Stephan Klaas, Weiskopf Nikolaus, Drysdale Peter M., Robinson Peter A., Friston Karl J.
Primary Institution: Wellcome Trust Centre for Neuroimaging, Institute of Neurology, University College London
Hypothesis
Which formulation is the best model for empirically measured BOLD responses?
Conclusion
The best model is a non-linear model with a revised form for the coefficients, treating ε as a free parameter.
Supporting Evidence
- The study compared eight different hemodynamic models using Bayesian model selection.
- Results indicated that allowing the parameter ε to vary improved model performance.
- The non-linear model with revised coefficients outperformed all other models tested.
Takeaway
This study looked at different ways to model brain activity and found that a more flexible model works best when analyzing brain scans.
Methodology
The study used dynamic causal modeling (DCM) to compare different hemodynamic models based on fMRI data.
Limitations
The results may not be generalizable across different brain regions and data sets.
Participant Demographics
Twelve subjects were included in the study.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website