In this project, we investigated whether the instruction-based congruency effect could be adjusted to the appearance of a cue. The associated work can be found at Munin.

Humans display an ability to rapidly adapt behaviour to novelly instructed relations. Instructions typically bind known stimuli and responses in novel ways that must be retained for future implementation, which can induce unintentional effects on a secondary task. This effect can be adjusted based on the degree of preparation for the retained instruction. Cues can be used to explicitly signal upcoming task demands, such that the degree of preparation for the retained instructions can be adjusted. To investigate whether preparation can be adjusted to an explicit cue, I used a novel design implemented in the instruction-based literature. A cue was presented a couple of trials before the trial related to the retained instruction was presented. I hypothesized that the effect on the secondary task would be reduced or absent before the cue but present after the cue. The analysis revealed the expected pattern when considering time as an interacting factor in the analysis. Thus, only the later parts of the experiment indicated the expected pattern, while the earlier and middle parts of the experiment did not. The evidence suggests that practice is necessary to attenuate preparation before the cue, in line with prior research

All data and materials used in this thesis is freely and publicly available in the main Open Science Foundation (OSF) repository, and this GitHub repository. Under the "Analyses - R" folder, you can find all material related to the data analyses. Under the "Task - jsPsych", you can find the task coded in JavaScript using the jsPsych library. Both experiments in the thesis were preregistered at OSF:

• Experiment 1
• Experiment 2

Each experiment has an associated analyses script in the "Analyses - R/reports" folder. They can be used to generate the outputs used for the thesis. Additional analysis scripts can be found under src.