Experimental setup: behavioral task, pharmacological manipulation and physiological responses.

A) Schematic representation of the behavioral task. Participants responded to the orientation (CW/CCW) of unilaterally presented Gabor stimuli that were embedded in noise (bilaterally presented). The likely location of the Gabor stimulus was cued (horizontal dash presented 0.33° left/right from fixation) with 80% validity before stimulus onset. B) Schematic overview of experimental sessions. Participants came to the lab on four occasions: one intake session and three experimental sessions. On the experimental sessions, participants received either placebo (PLC, data in orange), donepezil (DNP, 5mg, data in green) or atomoxetine (ATX, 40mg, data in blue). Drug order was counterbalanced across participants. C) Time schedule of experimental sessions. Participants received a pill on two moments in each session, one at the beginning of the session and a second pill two hours later. The first pill contained either placebo (PLC and ATX session) or donepezil (DNP session), the second pill was either a placebo (PLC and DNP session) or atomoxetine (ATX session). Behavioral testing started 4 hours after administration of the first pill. D) Baseline pupil diameter was measured before onset of the behavioral task. Participants fixated while the background luminance of the monitor was dimmed (for 15s) and then brightened (for 15s) to establish the pupil size in dark and bright circumstances. E) Pupil diameter during the dark (left) and bright (right) measurement windows for each drug condition separately. G-H) Effects of drug on heart rate (panel G) and mean arterial blood pressure (MAP; panel H, see Methods). Measurements were baseline-corrected to the first measurement taken right before ingestion of the first pill. I-J) Effects of drug on subjective ratings of alertness (panel I) and (panel J) calmness, derived from Visual Analogue Scale (VAS, see Methods). Abbr.: ATX: atomoxetine, PLC: placebo, DNP: donepezil

Behavioral results.

A) Signal detection theoretic sensitivity (d’), separately per drug and cue validity. B) As A, but for reaction time (RT). C) Schematic of the drift diffusion model (DDM), accounting for behavioral performance and reaction times. The model describes behavior on the basis of various latent parameters, including drift rate (v), boundary separation (a) and non-decision time (t0). D-E) As A, but for drift rate and non-decision time separately for the three drug conditions.

Evidence accumulation is affected by cue validity and drug, indexed by changes in centroparietal positivity (CPP).

A) Response-locked CPP for correct and incorrect answers, B) for trials with fast and slow RTs and C) for participants with overall high drift rate. D) Modulation of response-locked CPP by drug and cue validity. The horizontal black line indicates the time-window for which CPP slope was calculated (linear regression from -250ms to 0ms pre-response). The topographic map shows activation at the moment of the response, with white markers indicating the centro-parietal ROI used for the CPP analyses (channels CP1, CP2, CPz). E) Peak CPP amplitude, separately for drug and cue validity. F) CPP slope for all drug and cue validity conditions.

Modulation of perceptual processing by drug and attention, indexed by changes in occipito-temporal activity.

A) Stimulus-locked ERPs (0ms = target presentation) over bilateral occipito-temporal ROIs (left hemifield: P7/P9 and right hemifield: P8/P10; see white markers in topographic map), showing activity over regions contralateral (solid) and ipsilateral (dashed) to the cue for each drug condition (averaged over cue validity). To isolate temporal clusters of interest we performed 3×2×2 (drug x cue validity x hemisphere) factor rmANOVAs across time (resolution of∼8ms) and controlled for multiple comparisons using cluster-based permutation testing 44. Topographic map shows cue direction effect (contra – ipsi) across drug and attention conditions. Black dots indicate significant effects of rmANOVAs over time. B) Average activity within the first two temporal clusters from panel A related to cue direction. C) Activity within the late cluster of panel C for each of the three drug conditions. D) Same as panel A, but now split up for contralateral (solid) and ipsilateral (dashed) to the target stimulus (averaged over cue validity) and drug condition. Topographic map shows overall activity (no contrast) across drug and attention conditions. The drug cluster in this panel is identical to the cluster in panel A. E) Activity in late cluster of panel D related to target selection.

Attentional and catecholaminergic modulation of prestimulus cortical activity.

A) Cue-locked ERP (μV/cm2) over occipital regions contralateral and ipsilateral to the predictive cue split up for three drug conditions. Black horizontal bars; rmANOVA main effects cluster corrected for multiple comparisons. Topographic map shows the contrast between cue directions (left vs. right) during the early lateralization cluster. White markers indicate spatial ROI used for analyses. B) Average activity within the early temporal cluster related to cue-locked hemispheric lateralization. C) Cue-locked modulations of lateralized (contra-ipsi) prestimulus time-frequency power (dB/cm2). The significant TF cluster (highlighted by the solid black line) is derived from cluster-based permutation testing controlling for multiple comparisons. Green dotted box indicates alpha-band time-frequency ROI that is used to compute panel D. D) Cue-locked prestimulus alpha power (dB/cm2) over occipital regions contralateral and ipsilateral to the predictive cue split up for three drug conditions. A cluster-base corrected 3×2 rmANOVA revealed that alpha power was modulated by hemisphere (lateralization effect) and drug. Topographic map shows the contrast between cue directions (left vs. right) during the cluster in which we observed a main effect of drug. E) Average alpha power within the late temporal cluster in which drug effects on alpha power were observed, split up for drug conditions and hemisphere.