Related Experiment Video
Updated: Aug 29, 2026
Brain Imaging Investigation of the Neural Correlates of Observing Virtual Social Interactions
Published on: July 6, 2011
Model assessment and model building in fMRI
Mehrdad Razavi1, Thomas J Grabowski, Walter P Vispoel
1Department of Neurology, University of Iowa, Iowa City, Iowa, USA. Mehrdadrazavi@hotmail.com
Abstract:
Model quality is rarely assessed in fMRI data analyses and less often reported. This may have contributed to several shortcomings in the current fMRI data analyses, including: (1) Model mis-specification, leading to incorrect inference about the activation-maps, SPM[t] and SPM[F]; (2) Improper model selection based on the number of activated voxels, rather than on model quality; (3) Under-utilization of systematic model building, resulting in the common but suboptimal practice of using only a single, pre-specified, usually over-simplified model; (4) Spatially homogenous modeling, neglecting the spatial heterogeneity of fMRI signal fluctuations; and (5) Lack of standards for formal model comparison, contributing to the high variability of fMRI results across studies and centers. To overcome these shortcomings, it is essential to assess and report the quality of the models used in the analysis. In this study, we applied images of the Durbin-Watson statistic (DW-map) and the coefficient of multiple determination (R(2)-map) as complementary tools to assess the validity as well as goodness of fit, i.e., quality, of models in fMRI data analysis. Higher quality models were built upon reduced models using classic model building. While inclusion of an appropriate variable in the model improved the quality of the model, inclusion of an inappropriate variable, i.e., model mis-specification, adversely affected it. Higher quality models, however, occasionally decreased the number of activated voxels, whereas lower quality or inappropriate models occasionally increased the number of activated voxels, indicating that the conventional approach to fMRI data analysis may yield sub-optimal or incorrect results. We propose that model quality maps become part of a broader package of maps for quality assessment in fMRI, facilitating validation, optimization, and standardization of fMRI result across studies and centers. Hum. Brain Mapping 20:227-238, 2003.
Insights
Assessing functional magnetic resonance imaging (fMRI) model quality is crucial. Using Durbin-Watson (DW-map) and R(2) maps improves fMRI analysis validity and standardization, preventing suboptimal results.
Area of Science:
- Neuroimaging
- Statistical modeling
- Brain mapping
Background:
- Functional magnetic resonance imaging (fMRI) analysis often lacks rigorous model quality assessment, leading to potential mis-specification and incorrect inferences.
- Current practices may result in suboptimal model selection, under-utilization of systematic model building, and spatially homogenous modeling, contributing to result variability across studies.
- Lack of standardized model comparison exacerbates inconsistencies in fMRI findings.
Purpose of the Study:
- To introduce and evaluate the utility of Durbin-Watson statistic (DW-map) and R(2)-map as complementary tools for assessing fMRI model quality.
- To demonstrate how assessing model quality can overcome common shortcomings in fMRI data analysis.
- To propose the integration of model quality maps into standard fMRI quality assessment protocols.
Main Methods:
- Application of Durbin-Watson (DW-map) and R(2)-map images to assess model validity and goodness of fit in fMRI data.
- Utilizing classic model building techniques to construct higher-quality models based on reduced models.
- Investigating the impact of including appropriate versus inappropriate variables on model quality and the number of activated voxels.
Main Results:
- Higher quality models, built using appropriate variables, improved model validity and goodness of fit.
- Model mis-specification, through the inclusion of inappropriate variables, adversely affected model quality.
- Higher quality models sometimes decreased activated voxels, while lower quality or inappropriate models could artificially inflate them, highlighting potential flaws in conventional fMRI analysis.
Conclusions:
- Assessing and reporting fMRI model quality using tools like DW-maps and R(2)-maps is essential for accurate inference.
- Model quality assessment facilitates validation, optimization, and standardization of fMRI results across different studies and centers.
- Integrating model quality maps into fMRI analysis pipelines can enhance the reliability and reproducibility of neuroimaging findings.
