MEMO27

该数据集是一个基于五名受试者的fMRI数据构建的细粒度多标签情感解码数据集，包含了针对2196个激发不同情感的视频的大脑活动响应。情感评分基于34个类别，并设有构建情感标签矩阵的阈值。平均每个视频的情感标签数量为4.64，标签密度为0.172。数据集包含了五名受试者在2196个视频刺激下的BOLD响应，其任务是进行大脑活动情感解码。

This dataset is a fine-grained multi-label emotion decoding dataset built upon fMRI data collected from five subjects, encompassing brain activity responses to 2196 videos that elicit various emotions. Emotion annotations are based on 34 categories, with thresholds set to construct the emotion label matrix. The average number of emotion labels per video is 4.64, yielding a label density of 0.172. The dataset also contains BOLD responses from the five subjects during exposure to these 2196 video stimuli, with the core task being brain activity-based emotion decoding.