Welcome‎ > ‎


Data analysis competition Biomag 2016 (The 20th International Conference on Biomagnetism)

Title: Single-trial classification of event-related fields: detection of happy faces

Organizers: Hubert Cecotti, Jose Sanchez Bornot, and Girijesh Prasad

Northern Ireland Functional Brain Mapping (NIFBM) Facility NIFBM
Intelligent Systems Research Centre, School of Computing and Intelligent Systems
Ulster University, Magee Campus, Londonderry~Derry, Northern Ireland, UK.


In this data analysis competition, we consider an experiment using MEG recordings where subjects did pay attention to a stream of images (presentation rate=1 Hz  ( stimulus onset asynchrony=1000 ms, stimulus duration=333 ms)) (each block of 12 images contains faces from the same person, but with different facial expressions corresponding to 6 different classes: anger, disgust, fear, neutrality, sadness, and happiness). The FACES database was used for this experiment (http://colab.mpdl.mpg.de/mediawiki/Faces). The goal of the task was to detect the presence of faces with happiness, by pressing a button. The goal of the data analysis is to detect the presence of a face with happiness by using only the MEG signal and the stimulus onsets.

The data from four healthy adult subjects (mean age=33.8, 3 males) is provided and pre-processed using bandpass filtering and maxfilter.
The original data was acquired at 1 kHz with a 306 channel MEG system (Elekta Neuromag), which is comprised of 204 planar gradiometers and 102 magnetometers. The signal was bandpassed between 0.1 and 41 Hz, and downsampled to 125 Hz.
The data is processed and available directly in the Matlab format as a matrix containing:
  •     the signal (planardat)
  •     the stimulus onsets from the images and the behavioral responses (triggers)
    •  t1: Anger (non-target)
    •  t2: Disgust (non-target)
    •  t3: Fear (non-target)
    •  t4: Happiness (target)
    •  t5: Neutrality (non-target)
    •  t6: Sadness (non-target)
    • behavioral responses
    • test: triggers relative to the test data (unknown labels)
For the first part, only half of the labels, for each subject, will be available for training a classifier, the second half of the labels will be used to assess the methods. The evaluation will be performed by using the area under the ROC curve (AUC).

Participants will have to provide a complete description of the methods.
Therefore, the output should be a vector of real numbers corresponding to the classifier outputs for the test triggers (in the same order).
Participants are therefore not required to provide the program or code but they are required to provide a detailed description of the method that was used for both training and the test.  The winner of the competition will obtain an official certificate from the organizers. The winner of the competition will be also invited as a co-author to write about the results of the competition.

To download the data, please follow the links thereafter:

Additional information (05/09/2016):

You may use the position of the labels:
Position of the channels.

  • Team 1: Alexandre Barachant, Jean-Remi King, France.
  • Team 2: Emanuele Olivetti and Paolo Avesani, NeuroInformatics Laboratory (NILab), Bruno Kessler Foundation, Center for Mind and Brain Sciences (CIMeC), University of Trento, Italy
  • Team 3: Cristian Grozea, Fraunhofer Institute FOKUS, Berlin, Germany.
  • Team 4: Mohammed Abdulaal, School of EEE The University of Manchester, UK.
  • Team 5: Zafer İşcan, Centre for Cognition and Decision Making, National Research University, Higher School of Economics, Russian Federation.
  • Team 6: Andrea Vitale, Institute for advanded study (IUSS), Pavia, Italy and Christian Salvatore, National Research Council (CNR) at the Institute of Molecular Bioimaging and Physiology (IBFM), Milano, Italy.

Results (04/10/2016):

AUC for each subject.

  s1 s2 s3 s4 Mean SD Position
Team 1 0.982 0.893 0.962 0.986 0.956 0.043 Gold
Team 2 0.940 0.778 0.819 0.930 0.867 0.081 Silver
Team 3 0.808 0.655 0.734 0.890 0.772 0.101 Bronze
Team 4 0.524 0.498 0.545 0.466 0.508 0.034 6th
Team 5 0.756 0.559 0.577 0.745 0.659 0.106 4th
Team 6 0.511 0.520 0.618 0.505 0.538 0.054 5th

Congratulations to all the participants to have invested time and effort for the analysis of the data, and more importantly to the winning team: Alexandre Barachant and Jean-Remi King.

Data with the complete ground truth (coming soon, after the publication of the results.)


If there is a problem to parse the data and/or use the data, you may contact:
Hubert Cecotti: h.cecotti@ulster.ac.uk
Jose Miguel Sanchez Bornot: bornot@gmail.com
Girijesh Prasad: g.prasad@ulster.ac.uk