For within subject analysis the scanning paradigm was specified in SPM and a first level model estimation was performed.
For group analyses, individual contrast images were entered into a second level model using a flexible factorial ANOVA to assess (i) disease-related (between-subject factor Group, levels controls/Parkinson's disease OFF); and (ii) drug-induced (within-subject factor Medication, ON/OFF) effects on the factor Condition (i.e. Tapping/ Free/Intern/Extern.
To do this, we modified the second-level model to include community and pet ownership (p) specific intercepts (A cp) as follows: We then introduced a third-level model of the form where ϕ(dog) and ϕ(nodog) denote the between-community effects of ambient air pollution (centered by the overall mean pollution levels) by individual pet (e.g., dog) ownership status.
The 6 contrasts were evaluated in a second-level model with 1-sample t-tests.
These results were corroborated by an additional analysis based on a second-level model where regression maps from each emotion category were included as 9 separate conditions.
Statistical parametric maps were computed for each contrast of interest (the correct trials for each condition), and these contrast maps were entered into a second-level model treating individual subjects as a random effect.
A t-statistic was calculated for every voxel and contrast images were inserted into a random effects second level model (one-sample t-test) for group analyses.
In a second analysis, another first level model was created to investigate differences between old items in the different recollection tasks based on whether the participant had related a word to themselves or Obama during study, irrespective of whether the experimenter or participant had spoken the word out loud.
As in any Bayesian analysis, we define a third level of the model so that the variance parameters that are involved in the second-level equations (Equation 3) are themselves treated as unknown and given (hyper prior distributions.
In the subject-specific first level model, each stimulus onset was modeled by a canonical hemodynamic response function and its temporal derivative.
The crucial analysis step of the fixation-related approach was realized during the subject-specific first level model specification.
