Design Creativity: EEG Dataset in Loosely Controlled Modified TTCT-F Creativity Experiments

"Creativity Dataset" includes electroencephalography (EEG) signals recorded while participants performed creativity tasks. The participants in this study consisted of 28 graduate students. All participants were right-handed and had normal or corrected vision, reported good mental health, and had no history of medical or psychiatric issues or treatment. The EEG cap was positioned according to the 10-10 international standard system. EEG signals were recorded using a 64-channel BrainVision actiCHamp and sampled at 500 Hz. The creativity experiment followed a modified figural Torrance Test of Creative Thinking (TTCT-F) test. Each creativity experiment included three incomplete sketches shown to the participants. For each sketch, participants were asked to go through three creativity tasks including idea generation, idea evolution, and idea evaluation. The idea generation section instructed participants to intuitively complete a sketch based on their initial perception of the image. In idea evolution (IDE), volunteers were tasked with creating a drawing that significantly diverged from the previous sketch. In idea evaluation (IDR), participants assessed the difficulty of thinking and drawing in the idea generation and evolution sections. All the tasks were self-paced and open-ended, lasting up to three minutes. Rest periods (RST1 and RST2) of three minutes were included at the beginning and end of each experiment. The dataset is structured with 28 (.mat) files corresponding to the 28 participants. Each participant's file contains three trials and in each trial, participants engaged in three creativity tasks: idea generation (IDG), idea evolution (IDE), and idea evaluation (IDR). We used [dataset name]_[participant number]_[trial number]_[creativity task] to name the data records. Dataset _name is “Creativity” for all the recordings. Trial number can be 1,2, or 3 as there were three creativity problems (trials). Creativity task can be IDG, or IDE, or IDR. Rest periods are named RST1 and RST2. The EEG signals are represented in 2D matrices, where rows correspond to the number of channels (63 in this dataset, as one reference channel (Cz) was removed during preprocessing), and columns represent the length of the EEG signals. EEG signals were preprocessed using the Harvard Automated Processing Pipeline for Electroencephalography (HAPPE). This involved a bandpass filter (1 to 50 Hz), identification and isolation of globally bad EEG channels, removal of artifacts through independent components analysis (ICA) and wavelet transform, detection and isolation of local bad channels within 2-second segmented epochs, and interpolation of bad channels. The EEG data was then re-referenced to an average reference. The EEG signals were segmented by aligning them with video recordings of the experiments. The preprocessed and segmented EEG data were stored in the creativity dataset. To ensure participant privacy, we omitted videos that may reveal their identity.

创造力数据集（Creativity Dataset）收录了被试完成创造力任务时记录的脑电图（electroencephalography, EEG）信号。本研究的被试共28名研究生，均为右利手，视力正常或经矫正后正常，自述心理健康状况良好，无内科或精神疾病病史及治疗史。脑电图帽按照国际10-10系统（10-10 international standard system）进行定位，采用64通道BrainVision actiCHamp设备采集脑电信号，采样率设置为500 Hz。
本创造力实验采用改良版图形托兰斯创造性思维测验（figural Torrance Test of Creative Thinking, TTCT-F）。每项实验会向被试展示三幅未完成的素描作品，针对每幅素描，要求被试完成三项创造力任务：想法生成、想法拓展与想法评估。其中，想法生成环节（IDG）指导被试基于对图像的初始感知，直观完成素描；想法拓展环节（idea evolution, IDE）要求被试创作一幅与初始素描差异显著的全新画作；想法评估环节（idea evaluation, IDR）则要求被试对想法生成与拓展环节中思考与绘图的难度进行评估。所有任务均为自定进度的开放性任务，单任务最长持续时长为3分钟。每项实验的起始与结束阶段均设置了时长3分钟的休息时段（RST1与RST2）。
本数据集共包含28个.mat格式文件，一一对应28名被试。每名被试的文件中包含3组试次，每组试次内被试需完成三项创造力任务：想法生成（IDG）、想法拓展（IDE）与想法评估（IDR）。数据记录采用"数据集名称_被试编号_试次编号_创造力任务类型"的命名规则，所有记录的数据集名称均为"Creativity"。试次编号可取1、2、3，对应三项创造力问题（试次）；创造力任务类型可为IDG、IDE或IDR；休息时段的命名为RST1与RST2。
脑电信号以二维矩阵形式存储：行对应通道数量（本数据集共63通道，因预处理阶段移除了参考通道Cz），列对应脑电信号的时长。脑电信号采用哈佛自动化脑电处理流水线（Harvard Automated Processing Pipeline for Electroencephalography, HAPPE）进行预处理，具体流程包括：1~50 Hz带通滤波、识别并剔除全局坏道、通过独立成分分析（ICA）与小波变换去除伪影、在2秒分段时段内检测并剔除局部坏道、对坏道进行插值。随后将脑电数据重新参考至平均参考。将脑电信号与实验的视频记录对齐后进行分段。预处理并完成分段的脑电数据被存入本创造力数据集。为保护被试隐私，所有可能泄露被试身份的视频均已被移除。