For AI to be useful in high-stakes applications, humans need to be able to recognize when they can rely on AI and when not. However, in many cases, humans tend to overrely on AI, i.e. people adopt AI recommendations even when they are not adequate. Explanations of AI outputs are meant to help users calibrate their reliance in AI to appropriate levels, but often even increase overreliance. In this microproject, we investigated whether and how overreliance on AI and the effect of explanations depend on the perceived decision difficulty. We ran an AI-assisted decision-making experiment where participants were asked to solve a series of decision tasks with support by an AI model, both with and without explanations. Along with participants' reliance behavior, we measured perceived decision difficulty for each task and participant through EEG data. We found that people overrely significantly more in difficult decisions, and that explanations can in principle be helpful for easy decisions, but less so for difficult decisions.