Brain-Computer Interfaces for Communication: Preferences of Individuals With Locked-in Syndrome

Abstract

Background: Brain-computer interfaces (BCIs) have been proposed as an assistive technology (AT) allowing people with locked-in syndrome (LIS) to use neural signals to communicate. To design a communication BCI (cBCI) that is fully accepted by the users, their opinion should be taken into consideration during the research and development process.

Objective: We assessed the preferences of prospective cBCI users regarding (1) the applications they would like to control with a cBCI, (2) the mental strategies they would prefer to use to control the cBCI, and (3) when during their clinical trajectory they would like to be informed about AT and cBCIs. Furthermore, we investigated if individuals diagnosed with progressive and sudden onset (SO) disorders differ in their opinion.

Methods: We interviewed 28 Dutch individuals with LIS during a 3-hour home visit using multiple-choice, ranking, and open questions. During the interview, participants were informed about BCIs and the possible mental strategies.

Results: Participants rated (in)direct forms of communication, computer use, and environmental control as the most desired cBCI applications. In addition, active cBCI control strategies were preferred over reactive strategies. Furthermore, individuals with progressive and SO disorders preferred to be informed about AT and cBCIs at the moment they would need it.

Conclusions: We show that individuals diagnosed with progressive and SO disorders preferred, in general, the same applications, mental strategies, and time of information. By collecting the opinion of a large sample of individuals with LIS, this study provides valuable information to stakeholders in cBCI and other AT development.

Keywords: brain-computer interface; communication; locked-in syndrome; opinion; user-centered design.

Figure 1.

Representative screenshots of the animation videos used in the questionnaire. A total of 9 animations videos were shown to each participant. A. Section 2 of the questionnaire used an animation video to introduce communication brain-computer interfaces (cBCIs). Three illustrative screenshots of the video explaining what a cBCI is, how to control it, and what can it be used for are shown. In this video, we described the concept of an ideal cBCI that would be 100% accurate and 100% accepted by the users. B. In section 4 of the questionnaire, animation videos were presented to the participants, each describing a different mental strategy. For simplicity, consistency across mental strategies, and to avoid the application biasing the mental strategy, all videos showed a spelling matrix as a control application (left screenshot) and a button press (and subsequent letter selection) as a control output (right screenshot). C. Eight mental strategies were described in the questionnaire: attempted hand movement, attempted body movement (other than hand), attempted speech, counting backward, visual imagery, visual P300, auditory P300, and steady-state visual evoked potentials (SSVEPs). Abbreviation: AT, assistive technology.

Figure 2.

Demographic description of the participants: The demographic information of the participants (n = 28) in this study was extracted from section 1 of the questionnaire. The participants were divided into 2 groups based on the cause of locked-in syndrome—namely, neuromuscular disease (NMD) or sudden onset (SO; A-B) Information (in percentage) about the participants’ gender (male, female), age group (≤50 years old, >50 years old), living situation (at home or in a nursing home), and whether they were naïve to brain-computer interfaces (BCIs) is given for the NMD group (A) and the SO group (B). NMD accounted for 46% (n = 13) of the participants. (C-D) Histogram (in percentage) of the revised Amyotrophic Lateral Sclerosis Functional Rating Scale (ALSFRS-r) score (C) and duration of paralysis in years (D) per group (NMD in green and SO in yellow). Of note, at the time of diagnosis, patients with NMD are often still able to move and speak to a certain extent, hence the exact timing of becoming locked-in (and, therefore, the duration of the locked-in state) is unknown for these participants.

Figure 3.

Current communication channels and residual movement: A. Information (in percentage) about the remaining residual movement of the participants per group (neuromuscular disease [NMD], in green; sudden onset [SO], in yellow) at the time of the questionnaire. B. Histogram (in percentage) of the currently used communication channels per group (NMD in green, and SO in yellow). Abbreviation: LIS, locked-in syndrome.

Figure 4.

Preferred applications and mental strategies: (A-B) Ranking (using the center-of-mass metric, COM) of preferred applications (A) and mental strategies (B) possibly supported by a cBCI, by individuals with neuromuscular disease (NMD; n = 13, in green) and sudden onset disease (SO; n = 15, in yellow). Statistical difference between bars can be evaluated using the Monte Carlo variance indicated on the top right corner: 0.34 in (A) and 0.28 in (B). (C-D) Body parts (other than hand) selected by the participants during the “attempted body movement” strategy for both the NMD (C) and SO (D) groups. Rating scales ranged from 1 (least preferred) to 6 (most preferred) for A, and from 1 (least preferred) to 4 (most preferred) for B. Abbreviations: cBCI, communication brain-computer interface; SSVEP, steady-state visual-evoked potential.

Figure 5.

Time of information: Percentage of participants per group (neuromuscular diseases [NMD], in green; sudden onset [SO], in yellow) who would like to be informed about AT solutions (including cBCIs) in different phases after the onset or during the disease progression. Abbreviations: AT, assistive technology; cBCI, communication brain-computer interface.

Figure 6.

Willingness to consider a communication brain-computer interface (cBCI): Linear regression (dashed lines) between the number of residual movements the participant had at the time of the survey (from 1 to 5; see Figure 3A) and the willingness to consider a cBCI after the end of the questionnaire (from very unlikely to very likely on a 5-point scale). NMD participants are indicated with a green cross and SO participants with a yellow circle. The size of the cross/circle indicates the number of participants per coordinate: that is, the larger the cross/circle, the more the number of participants. The linear regression equation and corresponding r² (per group) are indicated on the top right corner.