Toward more intuitive brain–computer interfacing: classification of binary covert intentions using functional near-infrared spectroscopy

Abstract

In traditional brain-computer interface (BCI) studies, binary communication systems have generally been implemented using two mental tasks arbitrarily assigned to “yes” or “no” intentions (e.g., mental arithmetic calculation for “yes”). A recent pilot study performed with one paralyzed patient showed the possibility of a more intuitive paradigm for binary BCI communications, in which the patient’s internal yes/no intentions were directly decoded from functional near-infrared spectroscopy (fNIRS). We investigated whether such an “fNIRS-based direct intention decoding” paradigm can be reliably used for practical BCI communications. Eight healthy subjects participated in this study, and each participant was administered 70 disjunctive questions. Brain hemodynamic responses were recorded using a multichannel fNIRS device, while the participants were internally expressing “yes” or “no” intentions to each question. Different feature types, feature numbers, and time window sizes were tested to investigate optimal conditions for classifying the internal binary intentions. About 75% of the answers were correctly classified when the individual best feature set was employed (75.89% ± 1.39 and 74.08% ± 2.87 for oxygenated and deoxygenated hemoglobin responses, respectively), which was significantly higher than a random chance level (68.57% for p < 0.001). The kurtosis feature showed the highest mean classification accuracy among all feature types. The grand-averaged hemodynamic responses showed that wide brain regions are associated with the processing of binary implicit intentions. Our experimental results demonstrated that direct decoding of internal binary intention has the potential to be used for implementing more intuitive and user-friendly communication systems for patients with motor disabilities.

Brain-Based Binary Communication Using Spatiotemporal Features of fNIRS Responses

“…In line with this, it has been demonstrated that HbO signal is more robust than HbR for motor imagery specific activation (Mihara et al, 2012). Likewise, Rezazadeh Sereshkeh et al (2018) reported that HbO signals yielded the highest accuracies in their 3-class BCI using imagined speech, and Hwang et al (2016) reported that HbO features yield more discriminative information than HbR features in 2-class communication.…”

Toward a compact hybrid brain-computer interface (BCI): Performance evaluation of multi-class hybrid EEG-fNIRS BCIs with limited number of channels

“…EEG and functional near-infrared spectroscopy (fNIRS) are the two major neuroimaging modalities most frequently employed for implementing noninvasive BCIs. These methods have several advantages over invasive methods, such as being safer and offering high accessibility, a more affordable cost, scalability, and portability [5][6][7][8][9]. With the rapid development of BCI technology, increasing interest has been drawn toward hybrid BCIs, which are combinations of two or more BCI systems.…”

Reduction of Onset Delay in Functional Near-Infrared Spectroscopy: Prediction of HbO/HbR Signals

“…Thus far, the features frequently used from these windows include signal mean, signal slope, signal peak, skewness, kurtosis, variance, standard deviation, number and sum of peaks, root mean square, median, etc. (Hwang et al, 2016;Naseer et al, 2016;Liu and Hong, 2017;Hong et al, 2018b;Wibowo et al, 2018).…”