Brainwave Entrainment Beats (BWEB)

Brainwave entrainment beats detection has become an important topic due to the ability of these beats to change human brain waves to decrease anxiety, help focus attention, improve memory, improve mood, enhance creativity, reduce pain, help with meditation, enhance mental flexibility, and enhance sleep quality. However, listening to it can cause unwanted side effects as it can increase feelings of depression, anxiety, anger, and confusion in some people.

According to our understanding and previous literature assessment, no other researcher worked on how to detect the presence or absent of brainwave entrainment beats inside audio recordings in real time. As all the work that has been done were based on the classification of electromagnetic waves of the volunteers while they were listening to the audio files. Because of this, there is no data set that contains the brainwave entrainment beats inside the audio files.  Therefore, this dataset is the first dataset of its kind

Given the goal of creating a dataset for brainwave entrainment beats, we identified all kinds of brainwave entrainment beats and determined their effect on brain waves. The structure of the dataset is consisted of two categories which are, brainwave entrainment and no entrainment. A total of 150 tracks are available in each category. Within each category, there are files of various sorts. This collection of files was created to make the dataset more diversified and to include all the available options for each category. A balance has been kept between these different types by adding an equal number of files for each sort. The three varieties of brainwave entrainment are monaural, binaural, and isochronic, and they are defined by the sorts of beats inside the audio files. Delta, theta, alpha, beta, and gamma are the five various varieties of these three categories, which are categorized based on their effect on brain waves. The tracks in the No entrainment category were compiled from audio files belonging to various music genres as well as natural sounds.

The dataset samples were extracted from audio files from SoundCloud and YouTube servers. The audio samples were chosen for download based on audience feedback. Each track is 30 seconds long and has a sampling rate of 44100 samples per second at 32 bits. Samples were taken from each audio file in an orderly manner, as they were taken in a way that covers each audio file in terms of the beginning, middle and end.