fMRI data of subjects reading Harry Potter

该数据集包含了8名受试者在阅读《哈利·波特与魔法石》第九章时的功能性磁共振成像（fMRI）数据，旨在通过这些数据来解释自然语言处理（NLP）模型中的词汇和序列嵌入。此外，该数据已经使用SPM8进行了切片时间和运动校正，去趋势化，并使用3毫米全宽半最大（FWHM）的核进行了平滑处理。数据规模涉及8名受试者，每名受试者拥有25000至31000个皮质体素。研究任务是通过脑成像数据来解释自然语言处理模型。

This dataset collects functional magnetic resonance imaging (fMRI) data from 8 participants while they read Chapter 9 of *Harry Potter and the Philosopher's Stone*. The primary purpose of this dataset is to interpret lexical and sequential embeddings within natural language processing (NLP) models using the acquired neuroimaging data. Preprocessing of the raw data was performed using SPM8, including slice-timing correction, motion correction, detrending, and spatial smoothing with a 3-mm full-width at half-maximum (FWHM) kernel. The dataset comprises 8 subjects, with each subject containing 25,000 to 31,000 cortical voxels. The research task of this dataset is to interpret natural language processing models via brain imaging data.