To test this, we asked new participants to write captions describing the EEG stimuli (e.g. “A city seen through a forest”) and used Google Sentence Encoder to transform captions into semantic embeddings. We compared the RDMs of this semantic model to the super/typical brain RDMs