Context

This dataset consists of EEG recordings from 16 infants subjects. The subjects were presented with a visual animation: a black approaching rotating circle consisting of four smaller circles one-third of the size, (red, green, blue, yellow) rotating within the black circle on white background. This virtual object is called Loom. The loom rotates with a constant angular velocity of 300 degrees per second and starts at a virtual distance of 43.1 meters giving a visual angle of 5 degrees (6 cm diameter); it then grows to a maximum size of 131 degrees ( 350 cm diameter). The loom approaches the subjects constant accelerations at three different speeds with a duration of 2 seconds (−21.1 ㎨ ),3 seconds (−9.4 ㎨ ), and 4 seconds (−5.3 ㎨ ). The looming object moves the same distance in all three cases, as well as in reverse. This looming object is known for eliciting a response on the occipital area of the Brain. Such areas are recorded through O1, Oz, O2 electrodes according to the standard 81-electrode configuration in the 10-10 system. The response to a visual stimulus like Loom is known as a visual evoked potential or VEP.

This dataset was acquired by the Developmental Neuroscience Laboratory (NU-Lab) at Dragvoll, NTNU, Norway. For more information about infant data and the research of the NU-Lab see the Literature section.

Content

LOOMING_DATA Folder

In this folder, you find the same data of the RAW folder, but not the data has been sliced to have a constant array size (the difference in the size of the raw arrays are a couple of milliseconds which can be safely disregarded.).
The data is also organized with clear separation of looming stimulus duration, as well as subject and session.

The data is now represented as a CSV file.
No further processing is done to the data. So one would need to perform pre-processing steps such as removing the 50Hz noise component that comes as noise from the mains.

The folder and filename structure defines the characteristics of each datum. For example

./LOOMING_DATA/SUBJECT_##/SESSION_#/L#/##.csv

The folder SUBJECT_## specifies the subject where ## is a number from 1 to 16.
The folder SESSION_# specifies the recording session where # is either 1 or 2.
The folder L# specifies the looming type where # is either 2,3 or 4.
The filename is ##..csv and it defined by the sample number for each looming trial. This number varies as it not possible to get acquire always the same number of trials per infant subject.

RAW Folder

There are two folders in the dataset. The RAW folder has *.evt and *.mul files.
The evt files are column separated text files, which marks events using timestamps (TMU) columns.
The comment indicates the type of stimulus: 200s is the 2 seconds duration looming, 300s the 3 seconds duration stimulus, 400s is the 4 seconds duration stimulus, revd is the reverse stimulus. The beginning of the stimulus is also marked with the stm+ comment while stm- marks the end of the stimulus.

The filenames for both *.mul and *.evt files use the same naming rules:
For the following filename SUBJECT_XX_SESSION_#
The name of the file is determined by the subject id XX (a number from 1 to 16).
The name of the file is also determined by the session recording number #. Every subject has two recording sessions.

Acknowledgements

⚠ IN THE CASE YOU PUBLISH MATERIAL BASED ON THIS DATASET THEN IN YOUR ACKNOWLEDGEMENTS, PLEASE MENTION THE BASE WORK AS A REFERENCE

@inproceedings{BarbosaNIKT15,
  author    = {Igor Barros Barbosa and Kenneth Vilhelmsen and Audrey van der Meer and Ruud van der Weel and Theoharis Theoharis},
  title     = {{EEG} Biometrics: On the Use of Occipital Cortex Based Features from
               Visual Evoked Potentials},
  booktitle = {28th Norsk Informatikkonferanse, {NIK} 2015, H{\o}gskolen i {\AA}lesund, {\AA}lesund, Norway, November 23-25, 2015},
  publisher = {Bibsys Open Journal Systems, Norway},
  year      = {2015},
  url       = {http://ojs.bibsys.no/index.php/NIK/article/view/243},
}

Inspiration

Can you automatically detect VEP responses? See the aforementioned literature for more information on VEP.
Can the data be used to classify the subject?
Can the data be used to classify the session thought all the subjects?

Literature

The following is a list of reference work on Looming/VEP studies on infants, performed by the NU-Lab

Igor Barros Barbosa, Kenneth Vilhelmsen, Audrey van der Meer, Ruud van der Weel, and Theoharis Theoharis. “EEG Biometrics: On the Use of Occipital Cortex Based Features from Vi-
sual Evoked Potentials.” In: 28th Norsk Informatikkonferanse, NIK 2015, Høgskolen i Ålesund. Bibsys Open Journal Systems, Norway, Nov. 2015. Url: http://ojs.bibsys.no/index.php/NIK/article/view/243

Van Der Weel, F. R. R., & Van Der Meer, A. L. H. (2009). Seeing it coming: Infants’ brain responses to looming danger. Naturwissenschaften, 96(12), 1385–1391. https://doi.org/10.1007/s00114-009-0585-y
Van Der Meer, A. L. H., Svantesson, M., & Van Der Weel, F. R. R. (2013). Longitudinal study of looming in infants with high-density EEG. Developmental Neuroscience, 34(6), 488–501. https://doi.org/10.1159/000345154

Agyei, S. B., Van der Weel, F. R. R., & Van der Meer, A. L. H. (2016). Development of Visual Motion Perception for Prospective Control: Brain and Behavioral Studies in Infants. Frontiers in Psychology, 7(February), 1–14. https://doi.org/10.3389/fpsyg.2016.00100
Vilhelmsen, K., Agyei, S. B., Van der Weel, F. R. R., & Van der Meer, A. L. H. (2019). A high-density EEG study of differentiation between two speeds and directions of simulated optic flow in adults and infants. Psychophysiology, e13281. https://doi.org/10.1111/psyp.13281