NSD

该数据集名为神经影像与人工智能桥梁的NSD，是最大的同类数据集，包含了8名被试在观看不同自然场景时密集采集的功能性磁共振成像（fMRI）数据。每个被试的数据集都分为训练集和测试集，其中包含了特定数量的图像刺激和fMRI试验。这些图像刺激源自COCO数据集。规模上，该数据集涉及8名被试、9000至10000个不同的自然场景，以及22000至30000次的重复。其任务是利用大脑活动进行图像重建。

This dataset, named NSD (Neural Imaging and AI Bridge), is the largest dataset of its kind. It contains densely acquired functional magnetic resonance imaging (fMRI) data from 8 participants while they viewed various natural scenes. Each participant's dataset is split into a training set and a test set, which include a specific number of image stimuli and fMRI trials. These image stimuli are derived from the COCO dataset. In terms of scale, this dataset involves 8 participants, 9,000 to 10,000 distinct natural scenes, and 22,000 to 30,000 repeated trials. The core task of this dataset is image reconstruction using brain activity.