Dream Database from Donders

======================================== Dream EEG and Mentation (DREAM) data set ======================================== Data set information -------------------- - Common name: Dream Database from Donders - Full name: N/A - Authors: Cagatay Demirel, Jarrod Gott, Martin Dresler - Location: Donders Centre for Cognitive Neuroimaging - Year: 2019 - 2021 - Set ID: 5 - Amendment: 0 - Corresponding author ID: 12 Previous publications: Konkoly, K. R., Appel, K., Chabani, E., Mangiaruga, A., Gott, J., Mallett, R., ... & Paller, K. A. (2021). Real-time dialogue between experimenters and dreamers during REM sleep. Current Biology, 31(7), 1417-1427. Correspondence: [e-mail] Metadata -------- - Key ID: 6 - Date entered: 2023-02-21T12:48:54+00:00 - Number of samples: 7 - Number of subjects: 6 - Proportion REM: 14% - Proportion N1: 14% - Proportion N2: 43% - Proportion W: 0% - Proportion experience: 100% - Proportion no-experience: 0% - Proportion healthy: 100% - Provoked awakening: Yes - Time of awakening: Morning - Form of response: Structured - Date approved: 2023-02-21T12:47:51+00:00 How to decode data files ------------------------ * EEG files: All the EEG data are already transformed into ".edf" format from original Brainvision files. * Dream reports: Microsoft Word files in .docx format are used to format the dream reports in our study. Initially, we captured dream reports as .wav files and then meticulously refined the transcribed text to extract clear and accurate dream reports. ### Treatment group codes ### Codes correspond to the study numbers described below. Experimental description ------------------------ The database consists of data from three projects: * Real-time dialogue during lucid dreaming (study no: 0): Subjects communicated with the researchers while lucid dreaming, and answered very basic calculations with eye signals (e.g "what is 8 minus 6? --> "answer is 2 by moving ocular muscles on both sides to count the amount of eye signalling). * Motor-decoding during lucid dreaming (study no: 1): The goal is to induce hand-clenching during lucid dreams. In this experiment, participants are instructed to provide LRLR eye signals between each hand-clenching event to differentiate between the occurrences. * Other dream data: No experimental description, just some high-density EEG data with the dream content (study no: 2). Note: Please note that the majority of "dream recall" moments, clear phasic REM stages, and lucid dreaming (in some individuals) occurred around midday during our study. It is important to mention that all participants arrived at the EEG lab with high REM pressure around 7:00 a.m., and the events were observed between 11:00-12:00. As a result, our dream segments are more likely to be perceived as "day awakenings". Note: The data was collected under blanket ethical approval, and informed consent for participation was obtained from subjects. ### DREAM categorization procedure ### N/A ### Data acquisition ### The equipment used in this study included: - Easycap 128-channel EEG device (using the 10/05 EEG layout) with passive electrodes, which included EMG, EOG, and ECG. - actiCAP 64-channel EEG device (using the 10/10 EEG layout). * During study no. 1 and no. 2, certain EEG channels are converted into EOG and skin EMG signals through the use of adhesive holders. To avoid confusion, all modified channels are given updated names to reflect their current function. * ExG box with additional passive and bipolar EOG and EMGs are utilized. ### Data preprocessing ### The EEG data in the .edf format has not been preprocessed and remains in its raw form. However, since the data was originally collected from Brainvision files, the 3D electrode layout information (using the 10/10 system) is already embedded in the .edf files. As a result, when loading the data into either FieldTrip or MNE platforms, the layout information will be automatically included, and there is no need to search for an external EEG layout to integrate the data structure into MATLAB or Python platforms. ### Recommended data preprocessing for the whole data ### Various preprocessing pipelines could be utilized for the intended analysis; however, this is the most general preprocessing pipeline identified for our specific dataset. 1) Channel type assignment 2) Notch filter at 50 Hz 3) 0.1 - 49 Hz band-pass filter (in case above 50 Hz are not interested). 4) Noisy-channel tagging & interpolation 5) Optionally, EEG channel interpolation can be performed to generate artificial EEG signals on channels that have been converted to EOG and EMG. 6) Signal-space projection (SSP): https://mne.tools/0.16/manual/preprocessing/ssp.html

======================================== 梦境脑电图与心理活动（DREAM）数据集 ========================================
数据集基本信息
--------------------
- 通用名称：Donders梦境数据库
- 完整名称：无
- 作者：卡加泰·德米雷尔（Cagatay Demirel）、贾罗德·戈特（Jarrod Gott）、马丁·德雷斯勒（Martin Dresler）
- 所属机构：Donders认知神经成像中心
- 采集年份：2019-2021
- 数据集ID：5
- 修订版本：0
- 通讯作者ID：12
- 已发表相关文献：Konkoly K R, Appel K, Chabani E, Mangiaruga A, Gott J, Mallett R, 等. 快速眼动睡眠期间实验者与梦者的实时对话[J]. 当代生物学, 2021, 31(7): 1417-1427.
- 通讯邮箱：[e-mail]

元数据
--------------------
- 元数据键ID：6
- 录入日期：2023-02-21T12:48:54+00:00
- 样本量：7
- 受试者人数：6
- 快速眼动睡眠（Rapid Eye Movement, REM）占比：14%
- N1睡眠期占比：14%
- N2睡眠期占比：43%
- 清醒状态（Wakefulness, W）占比：0%
- 有梦境回忆占比：100%
- 无梦境回忆占比：0%
- 健康受试者占比：100%
- 唤醒方式：主动诱发唤醒
- 唤醒时间：晨间
- 应答形式：结构化应答
- 审核通过日期：2023-02-21T12:47:51+00:00

数据文件解码说明
--------------------
* EEG文件：所有脑电图数据已从原始Brainvision文件转换为EDF（.edf）格式。
* 梦境报告：本研究采用.docx格式的Microsoft Word文档存储梦境报告。最初我们以.wav格式录制梦境口述报告，后经细致转录整理，提取得到清晰准确的梦境文本内容。

### 实验组编码 ###
编码对应下文所述的研究编号。

实验设计说明
--------------------
本数据集包含三个项目的数据：
* 清醒梦期间的实时对话（研究编号：0）：受试者在清醒梦状态下可与研究人员进行沟通，并通过眼球运动信号完成基础算术作答（例如“8减6等于多少？→ 通过双侧眼肌运动计数作答，答案为2”）。
* 清醒梦期间的运动解码（研究编号：1）：实验目标为在清醒梦中诱发手部握拳动作。本实验中，受试者需在每次手部握拳动作之间通过LRLR眼球信号进行区分标记。
* 其他梦境数据（研究编号：2）：无额外实验设计，仅包含带有梦境内容的高密度脑电图数据。

注意事项：本研究中，多数“梦境回忆”时段、清晰的阶段性快速眼动（phasic REM）阶段以及部分个体的清醒梦事件均发生在午间时段。需特别说明的是，所有受试者均于清晨7:00左右抵达脑电图实验室，此时REM睡眠压力较高，实验观测时段为11:00-12:00，因此本数据集的梦境片段更易被视为“日间唤醒”所得结果。

另需说明：本数据采集在统一伦理审批框架下完成，受试者均已签署知情同意书。

### DREAM分类流程 ###
无

数据采集
--------------------
本研究使用的设备包括：
- 128通道Easycap脑电图设备（采用10/05电极布局），搭配无源电极，同步采集肌电（Electromyography, EMG）、眼电（Electrooculography, EOG）及心电（Electrocardiography, ECG）信号。
- 64通道actiCAP脑电图设备（采用10/10电极布局）。
* 在研究编号0和1的实验中，部分脑电图通道通过黏贴式转接器转换为眼电及皮肤肌电信号。为避免混淆，所有修改后的通道均已更新命名，以反映其当前采集功能。
* 额外配备ExG盒，用于采集附加的无源双极眼电及肌电信号。

数据预处理
--------------------
本数据集的EDF（.edf）格式脑电图数据未经预处理，保留原始采集状态。由于原始数据源自Brainvision文件，.edf文件中已嵌入基于10/10系统的三维电极布局信息。因此，当将数据导入FieldTrip或MNE平台时，电极布局信息将自动加载，无需额外检索外部脑电图布局文件即可适配MATLAB或Python平台的数据结构。

全数据集推荐预处理流程
--------------------
针对本数据集的分析可采用多种预处理流程，以下为适配本数据集的通用标准预处理方案：
1）通道类型标注
2）50Hz陷波滤波
3）0.1-49Hz带通滤波（若无需关注50Hz以上频段）
4）噪声通道标记与插值
5）可选：针对已转换为眼电及肌电的通道，可通过插值生成人工脑电图信号
6）信号空间投影（Signal-Space Projection, SSP）：https://mne.tools/0.16/manual/preprocessing/ssp.html