Task-induced frequency modulation features for brain-computer interfacing

Jayaram, Vinay; Hohmann, Matthias R.; Just, Jennifer; Schölkopf, Bernhard; Grosse-Wentrup, Moritz

doi:10.1088/1741-2552/aa7778

Cited by 8 publications

(8 citation statements)

References 49 publications

(66 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Most importantly, they induce modulation of neural activity in broad areas of the cortex, which could make them particularly suitable for use with consumer-grade EEG equipment with few sensors. We could show that both BCI control strategies could be used in a laboratory study to induce significant differences in neural activity [3], [11], [30], [4]. Now, we complement the laboratory results with MYND to investigate whether these strategies can be executed in an uncontrolled, realistic environment, by utilizing a consumer-grade EEG, and if repeated execution affects performance.…”

Section: Methodsmentioning

confidence: 94%

MYND: Unsupervised Evaluation of Novel BCI Control Strategies on Consumer Hardware

Hohmann¹,

Konieczny²,

Hackl³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Neurophysiological studies are typically conducted in laboratories with limited ecological validity, scalability, and generalizability of findings. This is a significant challenge for the development of braincomputer interfaces (BCIs), which ultimately need to function in unsupervised settings on consumer-grade hardware. We introduce MYND: A framework that couples consumer-grade recording hardware with an easy-to-use application for the unsupervised evaluation of BCI control strategies. Subjects are guided through experiment selection, hardware fitting, recording, and data upload in order to selfadminister multi-day studies that include neurophysiological recordings and questionnaires. As a use case, we evaluate two BCI control strategies ("Positive memories" and "Music imagery") in a realistic scenario by combining MYND with a four-channel electroencephalogram (EEG). Thirty subjects recorded 70.4 hours of EEG data with the system at home. The median headset fitting time was 25.9 seconds, and a median signal quality of 90.2% was retained during recordings. Neural activity in both control strategies could be decoded with an average offline accuracy of 68.5% and 64.0% across all days. The repeated unsupervised execution of the same strategy affected performance, which could be tackled by implementing feedback to let subjects switch between strategies or devise new strategies with the platform. Significance StatementUsing thoughts to interact with our environment could be the next frontier of medical and personal technology. The deployment of brain-computer interfaces (BCIs) on affordable hardware in daily life is the ultimate goal of this field, and both researchers and users need novel ways to evaluate conceptual systems under realistic conditions. We present a framework for the unsupervised evaluation of BCI control strategies on a consumer-grade electroencephalogram. We show that combining real-time guidance for hardware administration with a set of control strategies that induce broad spatial modulations in neural activity may be a viable basis for research on accessible BCI usage in daily life. M. R. H., M. G-W., and B. S. had the idea of building a platform for large-scale EEG recordings. M. R. H. designed and developed the platform, conducted experiments, analysed and interpreted the data, performed a literature review, and wrote the paper. M. H. and B. W. assisted with the design of the platform. L. K., M. L. and B. W. assisted with experiments. T. Z. and R. E. assisted with software development. M. G-W. and B. S. provided advice on the platform development and supervised the work.

show abstract

Section: Methodsmentioning

confidence: 94%

MYND: Unsupervised Evaluation of Novel BCI Control Strategies on Consumer Hardware

Hohmann¹,

Konieczny²,

Hackl³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…We utilized a set of stimulus-free two-task BCI control strategies that were previously evaluated in laboratory experiments with healthy subjects. It was shown that both BCI control strategies could be used in a laboratory study to induce significant differences in neural activity [20,19,27,16]. Now, we complement the laboratory results with MYND to investigate whether these strategies can be executed in an uncontrolled, realistic environment, by utilizing a consumer-grade EEG, and if repeated execution affects performance.…”

Section: Choice Of Laboratory-based Bci Control Strategiesmentioning

confidence: 92%

“…Self-referential thoughts have shown to modulate activity in the theta-(3-7 Hz) alpha-(8-13 Hz) and beta-bands (17-30 Hz) of the human EEG [12,36]. Further, it was found that the dominant frequency of the EEG spectrum, the "alpha peak frequency," is modulated by these strategies as well [27]. In figure 3, we used the laboratory data from previous studies [20,16] to illustrate the induced band-power modulations parietal and prefrontal cortex with the two BCI control strategies.…”

Section: Choice Of Laboratory-based Bci Control Strategiesmentioning

confidence: 99%

“…In the "Positive memories" strategy, subjects are asked to switch between thinking about a positive memory of their past or consecutively subtract a small number from a larger number [20,19,27]. Changing between daydreaming and a task that requires attention modulates activity in the "default mode network," a large-scale network that is involved in selfreferential processes [13,39].…”

Section: Choice Of Laboratory-based Bci Control Strategiesmentioning

confidence: 99%

“…Muscular artifacts in the laboratory studies were removed from the data with an independent component analysis (ICA) before band-power computations, as described in [19]. We extracted four features: theta-(3-7 Hz) alpha-(8-13 Hz) and beta-log-band-power (17-30 Hz), as well as the dominant frequency of the EEG spectrum as described in [27]. Then, we used mean and standard-deviation to normalize band-powers and the dominant frequency for every subject across the whole session in the laboratory study, and within each day in the at-home study.…”

Section: Control Strategy Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

Mynd

Hohmann

Konieczny

Hackl

et al. 2020

Proceedings of the 33rd Annual ACM Symposium on User Interface Software and Technology

Self Cite

View full text Add to dashboard Cite

Neurophysiological laboratory studies are often constraint to immediate geographical surroundings and access to equipment may be temporally restricted. Limitations of ecological validity, scalability, and generalizability of findings pose a significant challenge for the development of brain-computer interfaces (BCIs), which ultimately need to function in any context, on consumer-grade hardware. We introduce MYND: An open-source framework that couples consumer-grade recording hardware with an easy-to-use application for the unsupervised evaluation of BCI control strategies. Subjects are guided through experiment selection, hardware fitting, recording, and data upload in order to self-administer multi-day studies that include neurophysiological recordings and questionnaires at home. As a use case, thirty subjects evaluated two BCI control strategies ("Positive memories" and "Music imagery") by using a four-channel electroencephalogram (EEG) with MYND. Neural activity in both control strategies could be decoded with an average offline accuracy of 68.5% and 64.0% across all days.

show abstract