Influence of Error Rate on Frustration of BCI Users

Abstract

International audienceBrain-Computer Interfaces (BCIs) are still much less reliable than other input devices. The error rates of BCIs range from 5% up to 60%. In this paper, we assess the subjective frustration, motivation, and fatigue of BCI users, when confronted to different levels of error rate. We conducted a BCI experiment in which the error rate was artificially controlled. Our results first show that a prolonged use of BCI significantly increases the perceived fatigue, and induces a drop in motivation. We also found that user frustration increases with the error rate of the system but this increase does not seem critical for small differences of error rate. Thus, for future BCIs, we would advise to favor user comfort over accuracy when the potential gain of accuracy remains small

Usability Testing on Motionmouse: A Prototype to Control Android Tablet Using Emotiv Epoc+

“…In addition, to know how well (or not) each participant performed for every application, a total number of clicks per application are also recorded in Table 3. The above results show the overall accuracy of the prototype is 63.82% over the whole operating system, which leaves an average error rate of 36.18%, which is considered acceptable, as some researches indicate that the average error rate in BCI starts from 5% up to 60% [21]. But if we looked at the accuracy based on applications that the participants used in the experiment, the entertainment application (bubble shooter) gets the highest accuracy percentage with the lowest error rate, and the communication application (notepad) get the lowest accuracy percentage with highest error rate.…”

“…(iii) satisfaction via a set of post-questionnaires. Previous studies used post-questionnaire and direct questions to evaluate BCI applications like in [17,21,22]. The questions were designed as to address the metrics as per outlined by the ISO 9241-11:2018 standards [18] and in other works such as [21,22].…”

“…Previous studies used post-questionnaire and direct questions to evaluate BCI applications like in [17,21,22]. The questions were designed as to address the metrics as per outlined by the ISO 9241-11:2018 standards [18] and in other works such as [21,22]. The idea of the experiment is to let the participants move a mouse cursor in an Android device with their head movement and perform clicking by winking an eye.…”

Who Willed It?

“…The remaining 25% constitute false negative cases, and in some cases outcomes occur without input attempt (false positives). Studies have referred to this level of control as actual level of control [36,41], input attempt recognition [26], or decoding accuracy [12].…”

“…All the studies were surrogate BCI studies -they simulated BCI input such that they could manipulate the level of control. Figure 2 (left) synthesizes normalized levels of frustration from two such surrogate BCI studies [36,41]. Both studies aimed at finding actual control rates that minimized frustration and agreed on a necessary rate of 90%.…”