The Berlin Brain-Computer Interface: Progress Beyond Communication and Control

Blankertz, Benjamin; Acqualagna, Laura; Dähne, Sven; Haufe, Stefan; Schultze-Kraft, Matthias; Sturm, Irene; Uscumlic, Marija; Wenzel, Markus; Curio, Gabriel; Müller, Klaus Robert

doi:10.3389/fnins.2016.00530

Cited by 181 publications

(135 citation statements)

References 157 publications

Supporting

Mentioning

131

Contrasting

Order By: Relevance

“…A common problem is that machine learning based classifiers exploit any type of task-related information in the signals, including movement artifacts and systemic physiological changes of non-neuronal origin. This can lead to improved discriminability within the designed study but also to a greatly reduced performance when applied outside of the exact same experiment, and is a known pitfall in EEG-based BCI Blankertz et al, 2016).…”

Section: Preprocessing In Fnirs-based Bci: An Overview and Perspectivementioning

confidence: 99%

Using the General Linear Model to Improve Performance in fNIRS Single Trial Analysis and Classification: A Perspective

Lühmann

Ortega-Martínez

Boas

et al. 2020

Front. Hum. Neurosci.

View full text Add to dashboard Cite

Within a decade, single trial analysis of functional Near Infrared Spectroscopy (fNIRS) signals has gained significant momentum, and fNIRS joined the set of modalities frequently used for active and passive Brain Computer Interfaces (BCI). A great variety of methods for feature extraction and classification have been explored using state-of-the-art Machine Learning methods. In contrast, signal preprocessing and cleaning pipelines for fNIRS often follow simple recipes and so far rarely incorporate the available state-of-the-art in adjacent fields. In neuroscience, where fMRI and fNIRS are established neuroimaging tools, evoked hemodynamic brain activity is typically estimated across multiple trials using a General Linear Model (GLM). With the help of the GLM, subject, channel, and task specific evoked hemodynamic responses are estimated, and the evoked brain activity is more robustly separated from systemic physiological interference using independent measures of nuisance regressors, such as short-separation fNIRS measurements. When correctly applied in single trial analysis, e.g., in BCI, this approach can significantly enhance contrast to noise ratio of the brain signal, improve feature separability and ultimately lead to better classification accuracy. In this manuscript, we provide a brief introduction into the GLM and show how to incorporate it into a typical BCI preprocessing pipeline and cross-validation. Using a resting state fNIRS data set augmented with synthetic hemodynamic responses that provide ground truth brain activity, we compare the quality of commonly used fNIRS features for BCI that are extracted from (1) conventionally preprocessed signals, and (2) signals preprocessed with the GLM and physiological nuisance regressors. We show that the GLM-based approach can provide better single trial estimates of brain activity as well as a new feature type, i.e., the weight of the individual and channel-specific hemodynamic response function (HRF) regressor. The improved estimates yield features with higher separability, that significantly enhance accuracy in a binary classification task when compared to conventional preprocessing-on average +7.4% across subjects and feature types. We propose to adapt this well-established approach from neuroscience to the domain of single-trial analysis and preprocessing wherever the classification of evoked brain activity is of concern, for instance in BCI.

show abstract

Section: Preprocessing In Fnirs-based Bci: An Overview and Perspectivementioning

confidence: 99%

Using the General Linear Model to Improve Performance in fNIRS Single Trial Analysis and Classification: A Perspective

Lühmann

Ortega-Martínez

Boas

et al. 2020

Front. Hum. Neurosci.

View full text Add to dashboard Cite

show abstract

“…EMG and EEG processing. Processing of EMG and EEG signals was carried out using custom software and the BBCI toolbox 70 . EMG data were first decimated (data were low-pass filtered at 200 Hz prior to downsampling) to 500 Hz to match EEG sampling frequencies and subsequently high-pass filtered at 10 Hz, motivated by the fact that the power density function of surface EMG signals has insignificant contributions at frequencies <10 Hz 71 .…”

Section: Scientific Reports |mentioning

confidence: 99%

“…Additionally, EEG data were bandpass filtered at a passband of 0.5-100 Hz. Subsequently, all EEG signals were cleaned to remove eye-blink and scalp EMG artifacts by means of independent component analysis (ICA), using FastICA algorithms 72,73 implemented in the BBCI toolbox 70 . On average, 2.5 ± 1.2 components were excluded.…”

Section: Scientific Reports |mentioning

confidence: 99%

Corticomuscular interactions during different movement periods in a multi-joint compound movement

Kenville¹,

Maudrich²,

Vidaurre³

et al. 2020

Sci Rep

View full text Add to dashboard Cite

While much is known about motor control during simple movements, corticomuscular communication profiles during compound movement control remain largely unexplored. Here, we aimed at examining frequency band related interactions between brain and muscles during different movement periods of a bipedal squat (BpS) task utilizing regression corticomuscular coherence (rCMC), as well as partial directed coherence (PDC) analyses. Participants performed 40 squats, divided into three successive movement periods (Eccentric (ECC), Isometric (ISO) and Concentric (CON)) in a standardized manner. EEG was recorded from 32 channels specifically-tailored to cover bilateral sensorimotor areas while bilateral EMG was recorded from four main muscles of BpS. We found both significant CMC and PDC (in beta and gamma bands) during BpS execution, where CMC was significantly elevated during ECC and CON when compared to ISO. Further, the dominant direction of information flow (DIF) was most prominent in EEG-EMG direction for CON and EMG-EEG direction for ECC. Collectively, we provide novel evidence that motor control during BpS is potentially achieved through central motor commands driven by a combination of directed inputs spanning across multiple frequency bands. These results serve as an important step toward a better understanding of brain-muscle relationships during multi joint compound movements.Many muscles are involved in the execution and control of a bipedal squat (BpS) 1-3 , with Solomonow, et al. 4 estimating over 200 muscles to be recruited. In contrast to simple movements, BpS, as well as compound everyday life activities, i.e. walking stairs, picking up loads or carrying loads across distance, require extensive intra-and interlimb coordination 5 , which is why BpS is an ideal, naturalistic model for compound movement control. In contrast to simple movements, voluntary control of each muscle during BpS seems unlikely, especially when considering varying requirements on acting muscles due to changes in muscle function throughout different movement periods. Movement periods, i.e. dynamic (eccentric (ECC) and concentric (CON)) and static (isometric (ISO)) contraction periods require all muscles involved to dynamically change their function throughout the movement. This is evident for example from elevated proprioception, reflected by increased muscle spindle activities during eccentric contractions compared to both isometric and concentric contractions 6,7 . Thus, BpS demands continuous, extensive central-nervous information integration while placing high physical stress on the body 1-3 , signifying that motor commands have to be flexibly deployed and adapted throughout this movement in order to enable successful execution. Frequency band related neural synchrony between brain regions and muscles, detectable through corticomuscular coherence (CMC) measurements 8,9 , potentially provides an efficient solution to the challenge of enabling dynamic information processing on a whole body level during compound movement control. Indeed...

show abstract

“…"Dataset IV" is a NIRS dataset used in the study of Shin et al (2018b). All data processing was performed using MATLAB R2018b (Mathworks, MA, United States) and the BBCI toolbox 1 (Blankertz et al, 2016). A brief summary of the datasets I-IV is given in Table 1.…”

Section: Methodsmentioning

confidence: 99%

Performance Improvement of Near-Infrared Spectroscopy-Based Brain-Computer Interface Using Regularized Linear Discriminant Analysis Ensemble Classifier Based on Bootstrap Aggregating

Shin

2020

Front. Neurosci.

View full text Add to dashboard Cite

Ensemble classifiers have been proven to result in better classification accuracy than that of a single strong learner in many machine learning studies. Although many studies on electroencephalography-brain-computer interface (BCI) used ensemble classifiers to enhance the BCI performance, ensemble classifiers have hardly been employed for near-infrared spectroscopy (NIRS)-BCIs. In addition, since there has not been any systematic and comparative study, the efficacy of ensemble classifiers for NIRS-BCIs remains unknown. In this study, four NIRS-BCI datasets were employed to evaluate the efficacy of linear discriminant analysis ensemble classifiers based on the bootstrap aggregating. From the analysis results, significant (or marginally significant) increases in the bitrate as well as the classification accuracy were found for all four NIRS-BCI datasets employed in this study. Moreover, significant bitrate improvements were found in two of the four datasets.

show abstract

The Berlin Brain-Computer Interface: Progress Beyond Communication and Control

Cited by 181 publications

References 157 publications

Using the General Linear Model to Improve Performance in fNIRS Single Trial Analysis and Classification: A Perspective

Using the General Linear Model to Improve Performance in fNIRS Single Trial Analysis and Classification: A Perspective

Corticomuscular interactions during different movement periods in a multi-joint compound movement

Performance Improvement of Near-Infrared Spectroscopy-Based Brain-Computer Interface Using Regularized Linear Discriminant Analysis Ensemble Classifier Based on Bootstrap Aggregating

Contact Info

Product

Resources

About