Text entry rate of access interfaces used by people with physical disabilities: A systematic review

Simulation of other-initiated repair using AAC

This simulation study assessed the ability of Speech-Output Technologies (SOTs) to keep in-time during conversational repair. Fifty-eight Other Initiated Repair (OIR) initiators were collected from transcripts of repair interaction sequences collected from past research. A range of selection latencies were then used to calculate simulated utterance composition delays for the OIR initiators using two popular SOT software apps, with and without the use of word prediction. To determine whether OIR utterances could be produced within a socially sensitive temporal gap, composition delay was compared to a conservative temporal limit obtained for oral communicators (Kendrick, 2015). Even at the fastest 0.5 s selection latency level, utterance-level composition delays for both SOTs were substantially greater than the OIR limit set for this study. Next, AAC production rate data spanning a variety of technologies, access methods, tasks and user profiles was obtained from the literature. Communication performance for these groups was then evaluated against the identified temporal OIR limit. None of the user groups were found to be capable of producing full OIR utterances within the temporal limits of oral-speech conversation, with most unable to type even a single selection within these bounds. Because of the frequency and importance of repair in conversation, these results have important implications for designing devices to enable their users to successfully engage in such important conversational activities.

A click-based electrocorticographic brain-computer interface enables long-term high-performance switch-scan spelling

Background

Brain-computer interfaces (BCIs) can restore communication in movement- and/or speech-impaired individuals by enabling neural control of computer typing applications. Single command "click" decoders provide a basic yet highly functional capability.

Methods

We sought to test the performance and long-term stability of click-decoding using a chronically implanted high density electrocorticographic (ECoG) BCI with coverage of the sensorimotor cortex in a human clinical trial participant (ClinicalTrials.gov, NCT03567213) with amyotrophic lateral sclerosis (ALS). We trained the participant's click decoder using a small amount of training data (< 44 minutes across four days) collected up to 21 days prior to BCI use, and then tested it over a period of 90 days without any retraining or updating.

Results

Using this click decoder to navigate a switch-scanning spelling interface, the study participant was able to maintain a median spelling rate of 10.2 characters per min. Though a transient reduction in signal power modulation interrupted testing with this fixed model, a new click decoder achieved comparable performance despite being trained with even less data (< 15 min, within one day).

Conclusion

These results demonstrate that a click decoder can be trained with a small ECoG dataset while retaining robust performance for extended periods, providing functional text-based communication to BCI users.

Cutting-edge communication and learning assistive technologies for disabled children: An artificial intelligence perspective

In this study we provide an in-depth review and analysis of the impact of artificial intelligence (AI) components and solutions that support the development of cutting-edge assistive technologies for children with special needs. Various disabilities are addressed and the most recent assistive technologies that enhance communication and education of disabled children, as well as the AI technologies that have enabled their development, are presented. The paper summarizes with an AI perspective on future assistive technologies and ethical concerns arising from the use of such cutting-edge communication and learning technologies for children with disabilities.

Toward a Competency-based Approach to Co-designing Technologies with People with Intellectual Disability

Ability-based design is a useful framework that centralizes the abilities (all that users can do) of people with disabilities in approaching the design of assistive technologies. However, although this framework aspires to support designing with people with all kinds of disabilities, it is mainly effective in supporting those whose abilities can be clearly defined and measured, in particular, physical and sensory attributes of ability. As a result, the ability-based design framework only provides limited guidance to design with users with intellectual disability, whose cognitive, physical, sensory, and practical abilities vary along a spectrum. In this article, we reflect on a long-term co-design study where we leveraged what we termed “competencies,” i.e., the representative practical skills people develop from their participation in life activities, in particular, mainstream technologies, such as social media and the Internet. Our reflection is based on our experience in designing SkillsTube, a web application we co-designed with young adults with intellectual disability to support them to learn life skills through videos. The app's design, which explored and leveraged their social media participation competencies, supported the fundamental participation of all participants and their peers. Their familiarity with the app's social media-inspired design features fostered confidence in their participation, usability, and engagement. Drawing on the findings and design process of the app, we discuss a Competency-based approach to designing with people with disabilities that extends upon ability-based design, by grounding it in user competencies.

Reliability and validity of the usability scale for assistive technology for computer access: a preliminary study using video-based evaluation

Objective: The focus of this research was to establish the interrater reliability and content validity of the Usability Scale for Assistive Technology-Computer Access (USAT-CA), an evaluation tool to facilitate selection of computer access AT devices for individuals with disabilities. Methods: AT providers used the USAT-CA to evaluate two clients with physical disabilities interacting with their computer. Initially, an expert cohort of six providers used a draft version of the USAT-CA and provided feedback. Subsequently, 13 AT providers used a revised version to evaluate the clients. The providers filled out a content validation questionnaire and recommended computer access solutions for each client. Results: Overall intraclass correlation coefficient (ICC) scores were 0.8 and 0.9 for the AT provider cohorts, respectively. In each domain, ICC scores ranged from 0.7 to 0.9. There was agreement among the providers on the tool's comprehensiveness, clarity, time needed, and utility. For Client 1, 16 out of 19 AT providers recommended speech recognition software in combination with selected direct selection interfaces. For Client 2, 14 out of 19 providers recommended ergonomic and workspace modifications. Conclusion: The USAT-CA was found to have good to excellent inter-rater reliability, good content validity, and consistency in recommendations to optimize client-computer interaction.

AAC and Artificial Intelligence (AI)

Artificially intelligent tools have given us the capability to use technology to address ever more complex challenges. What are the capabilities, challenges, and hazards of incorporating and developing this technology for augmentative and alternative communication (AAC)? Artificial Intelligence can be defined as the capability of a machine to imitate human intelligence. The goal of artificial intelligence is to create machines that can use characteristics of human intelligence to solve problems and adapt to a changing environment. Harnessing the capabilities of AI tools has the potential to accelerate progress in serving individuals with complex communication needs. In this article, we discuss components of AI, including: (a) knowledge representation, (b) reasoning, (c) natural language processing, (d) machine learning, (e) computer vision, and (f) robotics. For each AI component, we delve into the implications, promise, and precautions of that component for AAC.

Influence of training protocols on text input speed on a computer in individuals with cervical spinal cord injury: a randomised controlled trial

STUDY DESIGN

Randomised controlled trial.

OBJECTIVES

To evaluate the impact of two different word prediction software (WPS) training protocols on text input speed (TIS) in people with tetraplegia and to determine which was the most effective.

SETTING

Rehabilitation department, Garches, France.

METHODS

Participants with neurological levels between C6 and C8 were allocated to one of three different groups. The REHAB group underwent training with an occupational therapist. The SELF group carried out a standardised home self-training with a written training guide. The CONTROL group had no training. Participants were assessed at day 15 (D15) and day 30 (D30). The primary outcome was a copying task with and without WPS (WITH and WITHOUT).

RESULTS

Forty-two participants (mean age ± SD of 39.8 ± 12) were included and 38 completed the study. At D30, the mean (95% confidence interval) difference in TIS between the CONTROL and SELF groups was 3.8 [-1.7 to 9.4] characters per minute (cpm) (p = 0.23), between the REHAB and SELF groups was 12.9 [7.4 to 18.4] cpm (p < 0.001), and between the REHAB and CONTROL groups was 9.1 [3.5 to 14.6] cpm (p < 0.001).

CONCLUSIONS

The results of this study showed that occupational therapist-supervised training improved TIS but word prediction software did not increase TIS. These results suggest that supervised training should be provided to all individuals who are prescribed with devices and systems to facilitate computer access in order to increase their TIS.

Effectiveness and usability of Scanning Wizard software: a tool for enhancing switch scanning

Scanning Wizard software helps scanning users improve the setup of their switch and scanning system. This study evaluated Scanning Wizard's effectiveness and usability. Ten people who use switch scanning and ten practitioners used Scanning Wizard in the initial session. Usability was high, based on survey responses averaging over 4.5 out of 5, and qualitative feedback was very positive. Five switch users were able to complete the multi-week protocol, using settings on their own scanning system that were recommended from the Scanning Wizard session. Using these revised settings, text entry rates improved by an average of 71%, ranging from 29% to 172% improvement. Results suggest that Scanning Wizard is a useful tool for improving the configuration of scanning systems for people who use switch scanning to communicate. Implications for Rehabilitation Some individuals with severe physical impairments use switch scanning for spoken and written communication. Scanning Wizard software helps scanning users improve the setup of their switch and scanning system. This study demonstrated high usability of Scanning Wizard (with 10 switch userpractitioner teams) and increased text entry rate by an average of 71% (for five switch users). Results suggest that Scanning Wizard is a useful tool for improving the configuration of scanning systems for people who use switch scanning to communicate.

Effect of diagnosis, body site and experience on text entry rate of individuals with physical disabilities: a systematic review

OBJECTIVE

This study systematically reviewed the research on computer text entry by people with physical disabilities, and conducted a quantitative synthesis of text entry rates associated with individuals' diagnosis, body site used with the interface and their level of experience.

METHOD

We searched 10 databases and included studies in which: typing speed was reported; the access interface was available for public use; and individuals with physical impairments were in the study population. For quantitative synthesis, we used only the text entry rates (TER) reported for individuals with physical impairments; studies also had to report the sample size, and the average and standard deviation for the text entry rates.

RESULTS

Thirty-nine studies involving 248 subjects met the criteria for quantitative synthesis. Cerebral palsy was associated with significantly slower TER, at 5.5 wpm, than muscular dystrophy (12.5 wpm), spina bifida (10.4 wpm), SCI high cervical (10.1 wpm) and SCI low cervical (13.3 wpm). Among the 19 body sites represented, the Fingers bilateral category had the highest average, at 17.72 wpm. Head (2.92 wpm) and Hand (non-typing) (3.95 wpm) were each associated with significantly slower TER than Hands unspecified, Fingers bilateral, Hand with control enhancer, Voice and Mouth. The three levels of experience examined, LowPlus, Medium and High, provided very similar TER.

CONCLUSION

This study contributes external evidence for service providers who provide computer access interventions for individuals with disabilities. The analyzed text entry performances serve as benchmarks to be used as guidelines during interface selection and training. Implications for Rehabilitation The median text entry rate (TER) across the literature for individuals with physical disabilities is 7.0 wpm. TER is affected by a user's diagnosis and the body site used for typing. These TER data can serve as diagnostic norms and benchmarks to be used as guidelines during interface selection and training. We recommend that practitioners measure text entry rate with clients, to place their performance in the context of these results and provide a baseline against which to measure effectiveness of an intervention.