Comparison of Four Control Methods for a Five-Choice Assistive Technology

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Severe motor impairments significantly impact communication abilities.
• Visual capabilities critically influence the selection of augmentative and alternative communication (AAC) systems.
• Understanding user initial impressions of AAC is crucial for effective assistive technology development.

Purpose of the Study:

• To compare user performance and initial impressions of visual (P300 BCI, eye-tracker) and non-visual (auditory P300 BCI, tactile P300 BCI) AAC systems.
• To evaluate the accuracy and information-transfer rates (ITRs) of different AAC control methods.
• To assess the correlation in performance between visual and non-visual AAC systems.

Main Methods:

• Eleven healthy participants completed a five-choice task using four AAC systems: visual P300 BCI (face stimuli), eye-tracker, auditory P300 BCI (Japanese Hiragana), and tactile P300 BCI (finger stimulation).
• Performance metrics including accuracy and ITRs were calculated based on selection time and accuracy above 70%.
• Correlations in performance were analyzed between visual and non-visual system pairs.

Main Results:

• The eye-tracker achieved 100% accuracy (28.2 bits/min ITR), and the visual P300 BCI achieved 88% accuracy (20.9 bits/min ITR).
• Non-visual systems showed lower accuracies: auditory BCI (70%, 3.3 bits/min) and tactile BCI (71%, 3.4 bits/min).
• Strong performance correlation was observed between the eye-tracker and visual P300 BCI, while tactile and auditory BCIs showed lower correlation.

Conclusions:

• Visual AAC systems, particularly eye-trackers and visual BCIs, demonstrate high performance and strong inter-system correlation.
• Non-visual AAC systems exhibit lower performance correlation, suggesting that individual user needs are paramount for system selection.
• The findings highlight the importance of personalized AAC system choice, especially for non-visual interfaces, to accommodate diverse user requirements.