BCI Performance Impacted by the Removal of Inter-Stimulus Interval

Brain-computer interfaces (BCI) technology can provide an alternative form of communication for individuals who lose their ability to communicate due to stroke, brain injury, or from amyotrophic lateral sclerosis (ALS). One of the most studied BCI\u27s is known as the P300 Speller. The P300 Speller presents a matrix of letters, numbers, and computer commands (8x9 in the current study). Each item in the matrix flashes at a rapid rate and the participants\u27 task is to count the number of times the desired letter flashes. When the attended letter flashes a P300 event-related potential will occur. The BCI can detect which stimulus produced a P300 and provide the stimulus as feedback to the participant. A primary goal of BCI research is to improve the speed and accuracy of the system by manipulating stimulus presentation parameters. One such parameter is the inter-stimulus-interval (ISI), which is the time between the offset of one stimulus (i.e., character in a grid of letters and numbers) and the onset of another stimulus. Previous studies have examined the effect of varying the ISI; however, to date, no study has included a condition that has no ISI between stimuli. This study compares BCI speed, accuracy, and bit rate from a condition that has no ISI to a condition that has an ISI of 62.5ms. In addition to evaluating performance, the current study also evaluates which of the two conditions is preferred by the participant. Preliminary data have been collected from seven able bodied participants. Each participant completed No-ISI and a 62.5ms ISI spelling task (counter-balanced). Before each spelling task a calibration period including eighteen characters (three six letter words) was conducted. After calibration, the participant was presented with an additional 18 characters and the computer provided online feedback indicating whether or not the BCI selected the correct character. Performance data were analyzed using t-tests. The No-ISI condition produced significantly higher bit rate than the 62.5ms ISI condition (p=0.039). No differences in accuracy (i.e., number of correct selections) or selections per minute (i.e., letters typed per minute) were observed, p=0.106 and p=0.088, respectively. In regard to the preference data, no significant difference was observed between the No-ISI and 62.5ms ISI condition (p=0.356). These data indicate that a No-ISI P300 BCI is a viable alternative to paradigms that included an ISI between sequential flashes of the matrix items. In addition, the results yielded higher bit rates and no difference in preference was observed. Based on these preliminary data, we suggest that the No-ISI paradigm should be tested in participants who have severe communication disorders