As the size and complexity of datasets increases, both visual-ization systems and their users are put under more pressure to oer quickand thorough insights about patterns hidden in this ocean of data. Whilenovel visualization techniques are being developed to better cope withthe various data contexts, users nd themselves increasingly often undermental bottlenecks that can induce a variety of emotions. In this paper,we execute a study to investigate the eectiveness of various emotion-triggered  adaptation  methods  for  visualization  systems.  The  emotionsconsidered are boredom and frustration, and are measured by means ofbrain-computer interface technology. Our ndings suggest that less intru-sive adaptive methods perform better at supporting users in overcomingemotional states with low valence or arousal, while more intrusive onestend to be misinterpreted or perceived as irritating.