The NBD dataset consists of three parts: - fMRI data in folder (/fMRI/), - text & meta-data (in folder /text_metadata/), - supplementary annotation (in fonder /annotation/). # fMRI data It consists of 24 folders (/S01/,..., /S24/) -- one for each subject. Each subject folder is divided in 6 run folders (/run1/,/run2/,/run3/,/run4/,/run5/, and /run6/) containing .nii volume images constituting the magnetic resonance recording during the presentation of the stimuli. # Text \& meta-data It consists of three .txt tab separated files - Narrative_1_wordtiming.txt, - Narrative_2_wordtiming.txt, - Narrative_3_wordtiming.txt, containing the text of the three narratives stimuli presented to the subjects. Each row in the file corresponds to one word (or word + punctuation) of one of the three narrative text stimuli accompanied with the temporal parameters of each word of the textual stimuli with regard to the experimental paradigm described above. These consist of word onset, offset and duration in seconds. # Supplementary annotations Annotations consist of part of speech (PoS) tags, and computational measures assigned to each word of the text stimuli. - Narrative_1_annotation.txt, - Narrative_2_annotation.txt, - Narrative_3_annotation.txt), Each column of the annotation file contains: the word and its PoS, word perplexity, PoS perplexity, average phonetic perplexity, word frequency, PoS frequency, average phonetic frequency, word entropy, word semantic association.