2.8. Response modeling and statistical analysis

A general linear model was used with an HD-DOT-derived adult hemodynamic response function (Hassanpour et al., 2014) to generate within-participant beta values of stimulus contrasted against rest for each run. Mean beta value maps for each participant were generated by averaging beta values from all runs for a given task type. Cluster-extent based thresholding was performed in SPM12 (Wellcome Trust Centre for Neuroimaging, London, UK) using the mean beta value maps for each participant as input. Cluster-extent based thresholds (k) were calculated with the Gaussian random field (GRF) method. The residual images were used to estimate intrinsic smoothness for the task vs. rest effects and the task contrast effects (e.g., OV vs. CV) using a primary voxel-level p < 1×10⁻⁴. One sample t-tests were calculated for the task vs. rest effects and paired t-tests were calculated for the task contrast effects, with p < 0.05 for each tail. As a total, 24 tails were assessed (positive and negative for each of the twelve comparisons for each of ΔHbO₂ and ΔHbR). Only clusters with p < 0.0021 (family wise error critical value, FWEc) were included as significant (0.05 ÷ 24 = 2.1 × 10⁻³).

To describe anatomical regions and Brodmann areas for each cluster, SPM outputs were entered into the WFU PickAtlas 3.0.5b (Maldjian et al., 2004; Maldjian et al., 2003) (Tables 2, ​,3,3, ​,4,4, ​,5).5). For each region, we then also calculated the Cohen’s-d to assess the effect size of the task response relative to rest, as well as for the contrasts between the paradigms. Finally, we estimated the temporal profile of the ΔHbO₂ contrast hemodynamic activity for each stimulus type within a variety of implicated regions of interest using a finite impulse response model (Glover, 1999; Hassanpour et al., 2014; Miezin et al., 2000). This procedure allowed us to both evaluate the time course of the evoked responses to each task type to ensure they were physiologically plausible and to highlight spatio-temporal characteristics differentiating brain function underlying these language tasks.

ΔHbO₂ Task vs. rest effects

RW, Reading Words; CV, Covert Verb Generation; OV, Overt Verb Generation; L, left; R, right; k, Cluster-extent based threshold

ΔHbR Task vs. rest effects

RW, Reading Words; CV, Covert Verb Generation; OV, Overt Verb Generation; L, left; R, right; k, Cluster-extent based threshold.

ΔHbO₂ Task vs. task effects

RW, Reading Words; CV, Covert Verb Generation; OV, Overt Verb Generation; L, left; R, right; k, Cluster-extent based threshold

ΔHbR Task vs. task effects

RW, Reading Words; CV, Covert Verb Generation; OV, Overt Verb Generation; L, left; R, right; k, Cluster-extent based threshold