Experimental materials and methods. a, Experimental paradigm. Two macaques were trained to exert force on an isometric joystick using the left hand. The force applied on the x- and y-axes of the joystick was used to control the position of a cursor (white dot) on a monitor. During a period of static force application, sudden task-irrelevant auditory stimuli were delivered through a beeper placed behind the monitor. b, Task timeline. The task began with the presentation of a target (white circle) on the center of the screen. Monkeys had 2 s to bring the cursor inside the target and were required to hold the cursor there for a variable time interval (ranging between 7 and 10 s). In 33% of trials, auditory stimuli were unexpectedly delivered during this interval. If the cursor remained inside the target, the trial was considered successful, and a liquid reward was given. Trials were separated by a 2–2.5 s (jittered) interval (during which monkeys were not required to exert force and therefore the cursor was likely to be back to the start position). c, EEG and LFP recording. In Experiment 1, we recorded EEG signals using 29 active electrodes (black dots) and 2 “zero-reference” electrodes (gray dots), mounted on custom-made EEG caps tailored to fit each animal’s head. In Experiment 2, LFPs were recorded from the right dorsolateral prefrontal cortex (Brodmann area 9), using a five-channel multiple-electrode array system for extracellular recording.