PCA2 (deliberate task focus)

Contributed by brontemckeown on June 7, 2023

Collection: Experience sampling reveals the role that covert goal states play in task-relevant behavior.

Description: Task-based analyses were carried out using FSL. A model was set up including 6 explanatory variables (EVs). EVs 1 and 2 modeled ‘vigilance’ and 'target detection’ periods. EV 3 modeled activity 6s prior to each mDES probe. Finally, EVs 4, 5, and 6 modeled the 3 thought components identified through PCA, with a time period of 6s prior to the mDES probes and the scores for the relevant component as a parametric regressor. EVs were mean-centered within each run and no thresholding was applied to the EVs. Standard and extended motion parameters were included as confounds. This was convolved with a hemodynamic response function using FSL’s gamma function. We chose to use the same 6s interval as used in Turnbull et al. 23. Contrasts were included to assess brain activity that related to each of the two task events (vigilance and target detection) and that related to each component of thought during the 6s period prior to the probe. The three runs were included in a fixed-level analysis to average across the activity within an individual. The averaged-run individual-level unthresholded (z-stat) contrast maps were used in the state-space analyses described below. Group-level analyses followed best practice 50. Specifically, we used FLAME, as implemented by FSL. Figure 2A and Figure 3C show the unthresholded group-level maps in MNI space while Figure 2B and Figure 3D show scatterplots representing how the BOLD activity in the maps’ parcels is distributed along the state-space dimensions. The unthresholded group-level maps were used in the NeuroSynth analysis described in the next section.

Task View 3D View