Airborne ultrasonic tactile display contactless brain-computer interface paradigm

• ¹ University of Tsukuba, Life Science Center of TARA, Japan
• ² RIKEN Brain Science Institute, Japan
• ³ The University of Tokyo, Japan

We present results of a study in which contactless and airborne ultrasonic tactile display (AUTD) stimuli delivered to the palms of a user serve as a platform for a brain computer interface (autd- BCI) paradigm. The reported research project received The Annual BCI Research Award 2014 for ”A fascinating new idea never explored before...,” as said by the research award committee chairman Prof. Gernot R. Mueller-Putz from Institute for Knowledge Discovery, Graz University of Technology, Austria. In the reported autdBCI paradigm the six palm positions are used to evoke somatosensory brain responses, in order to define a novel contactless tactile autdBCI. A comparison is made with classical attached vibrotactile transducers (vtBCI) revealing differences in ERP responses at 130 ms latencies with a significant negative deflection in case of the novel proposed autdBCI somatosensory modality. Thirteen male volunteer BCI users participated in the reported ex- periments. The AUTD stimulus generator produced vibrotactile contactless stimulation of the human skin via the air using focused ultrasound. The effect was achieved by generating an ultra- sonic radiation static force produced by intense sound pressure amplitude (a nonlinear acoustic phenomenon). The EEG signals in the autdBCI and vtBCI experiments were captured with an EEG amplifier system g.USBamp by g.tec, Austria, using 16 active electrodes. The BCI experiments based on six digits spelling (chance level of 16.6%) resulted with 15 trial averaged ERPs mean accuracies of 63.8% and 69.4% for autdBCI and vtBCI, respectively. The best results were 78.3% and 84.6% respectively. The differences were not significant, supporting the concept of autdBCI paradigm. However, a single trial classification offline analysis of the collected responses resulted in mean accuracies of 83.0% for autdBCI and 53.8% for vtBCI, leading to a possible ITRs of 19.2 bit/min and 7.9 bit/min, respectively. In the case of the autdBCI, only a single user’s results were bordering on the level of chance, and four subjects attained 100% (from 10 trials averaging). On average, lower accuracies were obtained with the classical vtBCI, with which three users bordered on the level of chance, and only one user scored 100% accuracy level in classified averaged responses. This case study demonstrates results obtained with a novel autdBCI paradigm which addi- tionally have shown an interesting stimulus depended ERP pattern negative deflections in the at 130 ms latencies. We present results of a study in which contactless and airborne ultrasonic tactile display (AUTD) stimuli delivered to the palms of a user serve as a platform for a brain computer interface (autd- BCI) paradigm. The reported research project received The Annual BCI Research Award 2014 for ”A fascinating new idea never explored before...,” as said by the research award committee chairman Prof. Gernot R. Mueller-Putz from Institute for Knowledge Discovery, Graz University of Technology, Austria. In the reported autdBCI paradigm the six palm positions are used to evoke somatosensory brain responses, in order to define a novel contactless tactile autdBCI. A comparison is made with classical attached vibrotactile transducers (vtBCI) revealing differences in ERP responses at 130 ms latencies with a significant negative deflection in case of the novel proposed autdBCI somatosensory modality. Thirteen male volunteer BCI users participated in the reported ex- periments. The AUTD stimulus generator produced vibrotactile contactless stimulation of the human skin via the air using focused ultrasound. The effect was achieved by generating an ultra- sonic radiation static force produced by intense sound pressure amplitude (a nonlinear acoustic phenomenon). The EEG signals in the autdBCI and vtBCI experiments were captured with an EEG amplifier system g.USBamp by g.tec, Austria, using 16 active electrodes. The BCI experiments based on six digits spelling (chance level of 16.6%) resulted with 15 trial averaged ERPs mean accuracies of 63.8% and 69.4% for autdBCI and vtBCI, respectively. The best results were 78.3% and 84.6% respectively. The differences were not significant, supporting the concept of autdBCI paradigm. However, a single trial classification offline analysis of the collected responses resulted in mean accuracies of 83.0% for autdBCI and 53.8% for vtBCI, leading to a possible ITRs of 19.2 bit/min and 7.9 bit/min, respectively. In the case of the autdBCI, only a single user’s results were bordering on the level of chance, and four subjects attained 100% (from 10 trials averaging). On average, lower accuracies were obtained with the classical vtBCI, with which three users bordered on the level of chance, and only one user scored 100% accuracy level in classified averaged responses. This case study demonstrates results obtained with a novel autdBCI paradigm which additionally have shown an interesting stimulus depended ERP pattern negative deflections in the at 130 ms latencies.