In recent years, many studies have been done on brain signals to be used to interpret people's mental states or intentions. Electrical brain signals or electroencephalogram (EEG) signals are the first choice in many studies on brain-computer interface. One of the most commonly used BCI systems is spellers which enable a person to spell letters using their brain signal without the need for muscle movement. These systems are appropriate for patients with neurological disorders. Accuracy and speed of spelling are two important criteria in evaluating these types of systems.
Event-Relate Potentials (ERPs) and Steady-State Evoked Potentials (SSEPs) are the two most important patterns of EEG signals used in BCI systems.
Event-related potentials (ERP) are transient or short signals that are characterized by voltage deviations in the EEG and are generated by external stimuli or cognitive processes resulting from external events. When a certain stimulus catches a person's attention, a change becomes apparent in his/her brain signals potential. Changes resulted from ERP in brain signals can be detected by processing performed in BCI systems (including feature extraction and classification procedures). So the user can control the BCI by focusing on a proper stimulus. One of the important components in ERP patterns is the P300 wave, which occurs between 300 and 1000 milliseconds after stimulation.
Steady-State Evoked Potentials (SSEPs) are oscillatory potentials that are generated by a repetitive external stimulus.
Steady-State Visual Evoked Potentials (SSVEP), as the most widely used type of this pattern, is an oscillatory signal that is generated by off-on visual stimulation frequency in the brain. SSVEP is widely used in BCI applications, and the pattern of stimulation of these potentials is as below: an image flashes at a certain frequency, when the person stares at the image, the effect of the stimulation frequency and its harmonics on the brain signal increases.
The proposed data for the Fourth National Brain-Computer Interface Competition are related to designing a new letter spelling system.
The objectives of the Competition:
Some objectives of the competition are as follows:
The Mode of Holding the Competition in this Course:
In the Fourth National Brain-Computer Interface Competition, the competition will be conducted on the registered data in a spelling system, in two stages. In the first stage, the participants will be provided with some labeled data. The data have been recorded from different people. Such data are used to train model parameters. Unlabeled data will then be provided to participants to evaluate the methods. The participants will announce evaluation data labels and based on the accuracy and precision results, the selected teams will be invited to in person(face-to-face) competition (second stage). In the face-to-face stage, a number of unlabeled data will be provided to the participants, in which both the accuracy and the rate of information transfer may affect the final result.
Terms of Participation in the Competition:
The competition is open to all researchers.
Event Place: National Brain Mapping Laboratory (NBML)