Levels of inference in psychophysiology: pointwise, cluster, and latent analyses of inhibition-task theta after fatigue-inducing protocols

Objectives: This study examined whether extending a confirmatory, single-channel analysis of frontal theta (Fz) to the whole scalp (32 channels), using permutation-based approaches with distinct, non-overlapping assumptions, influences conclusions about how different fatigue protocols affect inhibition-task theta.

Design: Thirty-six adults completed a within-subject design contrasting two prolonged Stroop protocols: a high-workload adaptive and a low-workload monotonous protocol. Both produced comparable subjective fatigue but differed in workload, sleepiness, and boredom. Inhibitory control was then assessed using Flanker and Go/No-Go tasks.

Methods: EEG was recorded and artefacts reduced via independent component analysis. Theta power was extracted via filter-Hilbert transform, and expressed as decibel change from a −200 to 0 ms pre-stimulus baseline. Theta difference scores (Flanker: incongruent minus congruent; Go/No-Go: no-go minus go) were quantified across all 32 channels from 200 to 600 ms post-stimulus. Three Monte Carlo approaches (10,000 iterations), all maxT-corrected, compared post-protocol inhibition-task theta at successively less constrained levels of inference: pointwise, cluster, and latent.

Results: For Flanker, no approach revealed reliable differences. For Go/No-Go, the pointwise approach revealed only a focal effect localized in space and time (greater inhibition-task theta after the high-workload protocol). The cluster-level test revealed a frontal cluster spanning a longer time window. Latent analysis revealed a component, with bootstrap-stable loadings forming a distributed scalp pattern broader than frontal-midline topography.

Conclusions: Despite comparable subjective fatigue, the post-protocol inhibition-task theta differed, with spatial-temporal extent depending on the level of inference. No single analysis revealed a definitive extent; each returned a different description of the same difference.