This "one-way ANOVA" model is the most basic analysis-of-variance model assuming one dependent variable and one between-subjects (grouping) factor with two or more levels (e.g. a group of healthy participant and one or more patient groups). To use this model for fMRI data, a single condition (beta) or a calculated contrast need to be extracted from the first-level GLM results as the dependent variable. The single-factor ANOVA is also an important model to analyze results from various other analysis tools resulting in a single dependent variaqble per subject. Examples of such analysis tools are cortical thickness measurements that usually result in a thickness surface map per subject. Other examples are second-level analyses of ICA, FA or Granger Causality maps. While it is possible to compare two groups also with the "Combine VMPs/SMPs" tools, the single-factor ANOVA model allows analysis of designs with multiple experimental groups.
The single-factor ANOVA model is selected automatically when a multi-subject VMP or multi-subject SMP file is selected in the Input tab of the ANCOVA dialog. Note that each subject must have exactly one map for this design. To decide whether the single-factor model is appropriate, the program checks the names of the individual maps to determine how many condition maps exist per subject. If the number of condition maps per subject is 1, the single-factor model is selected, if the number is larger than 1, the single-factor repeated measures model is selected as the default design. If you have a multi-factorial design, more within or between factors need to be specified using the No. of within-subjects factors or No. of between-subjects factors spin boxes.
The procedure to specify the design and to run the model is identical to the case of the single-factor ANCOVA model except that the latter adds a covariate. For further instructions, follow the description in topic single-factor ANCOVA ignoring the extra steps for adding the covariate.
Copyright © 2020 Rainer Goebel. All rights reserved.