Background
In this project the Bag-of-Lies data set is used for predicting if participants lied or not. The data set is multimodal, consisting of video, audio, EEG and Eye Gaze data from 35 unique subjects describing a selection of images. The finalized data set has a total of 325 annotated recordings consisting of 162 lies and 163 truths, meaning that the data is balanced on the target variable.
The data includes 10 female and 25 male participants, with each subject being shown 6–10 images from a selected image set and asked to provide a description of the image. [1] Participants were free to choose between giving a honest description or lie about the image shown. This affirms that the results are not directly dependent on the complexity or intelligibility of the figures shown. During the experiment EEG and gaze data was collected.
In this project, the main focus is on the EEG data. Electroencephalography (EEG) is an electrophysiological monitoring method to record electrical activity of the brain. EEG measures voltage fluctuations resulting from ionic current within the neurons of the brain. [2] The EEG recordings are obtained by placing electrodes on the scalp and measuring the electrical activity in each region. As a result mapped time series data of all measured regions is acquired.
The EEG data, which is collected for our data set, has 13 channels (F3, FC5, F7, T7, P7, O1, O2, P8, T8, F8, AF4, FC6, F4), sampled at 2048 Hz internally and output filtered to 128 Hz. [1] For each channel the quality of the reading is also included.
