When humans shift between tasks, they initially show slower responses in the new task than in the previous one. Persisting attentional settings are increasingly recognized as a source for these shifting costs. However, the extent to which specific mechanisms underlying information selection and interference control contribute to this phenomenon remains less clear. Here, we use time-frequency analyses of human electroencephalogram (EEG) data to explore the aftereffects of two such mechanisms: target amplification and distracter inhibition. Participants completed a set-shifting task in which interference during switch trials could either result from the persisting amplification of previous target colors or the persisting inhibition of previous distracter colors. In a first set of analyses, we focused on frontal midline theta (FMT) as a time-continuous marker of overall interference. Compared to a control condition, we found transient peaks of FMT in both experimental conditions that matched the effects of persisting target amplification and distracter inhibition predicted by a computational model of the task. In a second set of analyses we used steady-state visually evoked potentials (SSVEPs) as a direct measure of the attentional resources allocated to target and distracter colors. However, SSVEP amplitudes did not differ reliably between stimulation frequencies during switch trials, preventing us from drawing further conclusions on the origins of the interference processes reflected in FMT dynamics. Implications for theories of selective attention and potential limitations of frequency tagging in the context of mental set-shifting research are discussed.