moabb.datasets.MartinezCagigal2023Pary#
- class moabb.datasets.MartinezCagigal2023Pary(conditions=('2', '3', '5', '7', '11'))[source]#
P-ary m-sequence-based c-VEP dataset from Martínez-Cagigal et al. (2023)
Dataset summary
#Subj
16
#Chan
16
#Trials / class
2-30
Trials length
5.3/6.7/10.3/4.0/10.0 s
Freq
256 Hz
#Sessions
5
#Trial classes
16
#Epochs classes
2-11
#Epochs / class
6200-19220
Codes
p-ary m-sequence
Presentation rate
120 Hz
Dataset Description
This dataset was originally recorded for study [1], which evaluated different non-binary encoding strategies. Specifically, five different conditions were tested in a 16-command speller. Each condition used a different p-ary m-sequence to encode the commands via circular shifting. One command was encoded using the original m-sequence, while the remaining commands were encoded using shifted versions of that sequence [2].
A p-ary m-sequence means it contains p different events, which were encoded using different shades of gray. For example, in the binary case (p=2), events 0 and 1 were encoded using white and black flashes, respectively. For p=3, black, white, and mid-gray flashes were used [1].
The evaluated conditions were:
Base 2: GF(2^6) m-sequence of 63 bits
Base 3: GF(3^4) m-sequence of 80 bits
Base 5: GF(5^3) m-sequence of 124 bits
Base 7: GF(7^2) m-sequence of 48 bits
Base 11: GF(11^2) m-sequence of 120 bits
The dataset includes recordings from 16 healthy subjects performing a copy-spelling task under each condition. The evaluation was conducted in a single session, during which each participant completed:
A calibration phase consisting of 30 trials using the original m-sequence (divided into six recordings of five trials each), and
An online copy-spelling task of 32 trials (divided into two recordings of 16 trials each).
Each trial consisted of 10 cycles (i.e., repetitions of the same code). Additionally, participants completed questionnaires to assess satisfaction and perceived eyestrain for each m-sequence condition. Questionnaire results are available in [3].
The encoding was displayed at a 120 Hz refresh rate. EEG signals were recorded using a g.USBamp amplifier (g.Tec, Guger Technologies, Austria) with 16 active electrodes and a sampling rate of 256 Hz. Electrodes were placed at: F3, Fz, F4, C3, Cz, C4, CPz, P3, Pz, P4, PO7, PO8, Oz, I1, and I2; grounded at AFz and referenced to the earlobe.
Note
Recordings of user “zdvm” for bases 2, 3, 5, and 7 had a sampling rate of 600 Hz. The rest of recordings have all a sampling rate of 256 Hz.
The experimental paradigm was executed using the MEDUSA© software [4].
- Parameters:
conditions (tuple of str, optional) – Which conditions to load. Default is all conditions: (“2”, “3”, “5”, “7”, “11”). Each condition corresponds to a different p-ary m-sequence base.
References
[1] (1,2)Martínez-Cagigal, V., Santamaría-Vázquez, E., Pérez-Velasco, S., Marcos-Martínez, D., Moreno-Calderón, S., & Hornero, R. (2023). Non-binary m-sequences for more comfortable brain-computer interfaces based on c-VEPs. Expert Systems with Applications, 232, 120815. https://doi.org/10.1016/j.eswa.2023.120815
[2]Martínez-Cagigal, V., Thielen, J., Santamaría-Vázquez, E., Pérez-Velasco, S., Desain, P., & Hornero, R. (2021). Brain-computer interfaces based on code-modulated visual evoked potentials (c-VEP): A literature review. *Journal of Neural Engineering, 18*(6), 061002. https://doi.org/10.1088/1741-2552/ac38cf
[3]Martínez-Cagigal, V. (2025). Dataset: Non-binary m-sequences for more comfortable brain-computer interfaces based on c-VEPs. https://doi.org/10.35376/10324/70945
[4]Santamaría-Vázquez, E., Martínez-Cagigal, V., Marcos-Martínez, D., Rodríguez-González, V., Pérez-Velasco, S., Moreno-Calderón, S., & Hornero, R. (2023). MEDUSA©: A novel Python-based software ecosystem to accelerate brain-computer interface and cognitive neuroscience research. Computer Methods and Programs in Biomedicine, 230, 107357. https://doi.org/10.1016/j.cmpb.2023.107357
Notes
Although the dataset was recorded in a single session, each condition is stored as a separate session to match the MOABB structure. Within each session, eight runs are available (six for training, two for testing).
Added in version 1.2.0.