Turning negative into positives! Exploiting ‘negative’ results in Brain–Machine Interface (BMI) research

Lotte, Fabien; Jeunet, Camille; Chavarriaga, Ricardo; Bougrain, Laurent; Thompson, David E.; Scherer, Reinhold; Mowla, Rakibul; Kübler, Andrea; Grosse-Wentrup, Moritz; Dijkstra, Karen; Dayan, Natalie

doi:10.1080/2326263x.2019.1697143

Cited by 13 publications

(6 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results of the present study suggest that findings of adverse effects of interference or an early boost on motor skill performance cannot readily be generalized to MI NF skill acquisition. As previously shown, the interplay of MI (NF) and performance increase is complex (Dickhaus, Sannelli, Müller, Curio, & Blankertz, 2009;Lotte et al, 2019;Vidaurre & Blankertz, 2010). Context factors are likely contributing to this complexity.…”

Section: Resultsmentioning

confidence: 97%

Motor Imagery EEG neurofeedback skill acquisition in the context of declarative interference and sleep

Daeglau

Zich

Welzel

et al. 2020

Preprint

View full text Add to dashboard Cite

Motor imagery (MI) practice in combination with neurofeedback (NF) is a promising supplement to facilitate the acquisition of motor abilities and the recovery of impaired motor abilities following brain injuries. However, the ability to control MI NF is subject to a wide range of inter-individual variability. A substantial number of users experience difficulties in achieving good results, which compromises their chances to benefit from MI NF in a learning or rehabilitation context. It has been suggested that context factors, that is, factors outside the actual motor task, can explain individual differences in motor skill acquisition. Retrospective declarative interference and sleep have already been identified as critical factors for motor execution (ME) and MI based practice. Here, we investigate whether these findings generalize to MI NF practice.Three groups underwent three blocks of MI NF practice each on two subsequent days. In two of the groups, MI NF blocks were followed by either immediate or delayed declarative memory tasks. The control group performed only MI NF and no specific interference tasks. Two of the MI NF blocks were run on the first day of the experiment, the third in the morning of the second day. Significant within-block NF gains in mu and beta frequency event-related desynchronization (ERD) where evident for all groups. However, effects of sleep on MI NF ERD were not found. Data did also not indicate an impact of immediate or delayed declarative interference on MI NF ERD.Our results indicate that effects of sleep and declarative interference context on ME or MI practice cannot unconditionally be generalized to MI NF skill acquisition. The findings are discussed in the context of variable experimental task designs, inter-individual differences, and performance measures.

show abstract

Section: Resultsmentioning

confidence: 97%

Motor Imagery EEG neurofeedback skill acquisition in the context of declarative interference and sleep

Daeglau

Zich

Welzel

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Most studies proved that classifying affective states from EEG remains really challenging since results hardly go over chance level accuracy. Some studies were even unable to obtain better than chance results when reproducing previous works with statistically rigorous evaluation methods [30]. Finally, confounding factors due to electromyography (e.g., facial muscles activity during emotion expression) have likely played a role in the performances obtained in many studies.…”

Section: B Affective States Estimation From Eegmentioning

confidence: 96%

Modern Machine-Learning Algorithms: For Classifying Cognitive and Affective States From Electroencephalography Signals

Appriou

Cichocki

Lotte

2020

IEEE Syst. Man Cybern. Mag.

Self Cite

View full text Add to dashboard Cite

Estimating cognitive or affective states from brain signals is a key but challenging step in the creation of passive brain-computer interface (BCI) applications. So far, estimating mental workload or emotions from EEG signals is only feasible with modest classification accuracies, thus leading to unreliable neuroadaptive applications. However, recent machine learning algorithms, notably Riemannian geometry based classifiers (RGC) and convolutional neural networks (CNN), have shown to be promising for other BCI systems, e.g., motor imagery-BCIs. However, they have not been formally studied and compared together for cognitive or affective states classification. This paper thus explores such machine learning algorithms, proposes new variants of them, and benchmarks them with classical methods to estimate both mental workload and affective states (Valence/Arousal) from EEG signals. We study these approaches with both subject-specific and subject-independent calibration, to go towards calibration-free systems. Our results suggested that a CNN obtained the highest mean accuracy, although not significantly so, in both conditions for the mental workload study, followed by RGCs. However, this same CNN underperformed in both conditions for the emotion data set, a data set with small training data. On the contrary, RGCs proved to have the highest mean accuracy with the Filter Bank Tangent Space classifier (FBTSC) we introduced in this paper. Our results thus contributed to improve the reliability of cognitive and affective states classification from EEG. They also provide guidelines about when to use which machine learning algorithm.

show abstract

“…To be able to build such models, it should be stressed that we need both positive and negative results, i.e. to know both what works and what does not work to improve training [291].…”

Section: On the Need For Models Of Mt-bci User Trainingmentioning

confidence: 99%

A review of user training methods in brain computer interfaces based on mental tasks

et al. 2021

Self Cite

View full text Add to dashboard Cite

Mental-Tasks based Brain-Computer Interfaces (MT-BCIs) allow their users to interact with an external device solely by using brain signals produced through mental tasks. While MT-BCIs are promising for many applications, they are still barely used outside laboratories due to their lack of reliability. MT-BCIs require their users to develop the ability to self-regulate specific brain signals. However, the human learning process to control a BCI is still relatively poorly understood and how to optimally train this ability is currently under investigation. Despite their promises and achievements, traditional training programs have been shown to be sub-optimal and could be further improved. In order to optimize user training and improve BCI performance, human factors should be taken into account. An interdisciplinary approach should be adopted to provide learners with appropriate and/or adaptive training. In this article, we provide an overview of existing methods for MT-BCI user training-notably in terms of environment, instructions, feedback and exercises. We present a categorization and taxonomy of these training approaches, provide guidelines on how to choose the best methods and identify open challenges and perspectives to further improve MT-BCI user training.

show abstract

Turning negative into positives! Exploiting ‘negative’ results in Brain–Machine Interface (BMI) research

Cited by 13 publications

References 59 publications

Motor Imagery EEG neurofeedback skill acquisition in the context of declarative interference and sleep

Motor Imagery EEG neurofeedback skill acquisition in the context of declarative interference and sleep

Modern Machine-Learning Algorithms: For Classifying Cognitive and Affective States From Electroencephalography Signals

A review of user training methods in brain computer interfaces based on mental tasks

Contact Info

Product

Resources

About