AI-Driven Disability Support

The development of AI technologies has led to the generation of new ideas in healthcare, such as disability support. What kind of dreams will be envisioned next? Tatsuo Matsunaga, Director of the Division of Hearing and Balance Research at Tokyo Medical Center, and three consultants from Fujitsu Research Institute participated in an interview on AI-driven disability support. (Interview date: May 11, 2018)

Speakers (from the left)
Yoshiyuki Okihara (moderator): Principal Consultant, Business Science Group, Fujitsu Research Institute
Ryo Tamura: Consultant, Business Science Group, Fujitsu Research Institute
Tatsuo Matsunaga: Director, Division of Hearing and Balance Research, Clinical Research Center, Tokyo Medical Center
Chizuko Kamenoi: Senior Managing Consultant, Business Science Group, Fujitsu Research Institute

* This article was published in Chisounomori 2018, volume 4 on July 27, 2018. Chisounomori is an information magazine published by Fujitsu Research Institute (FRI).

* The department and position of the author and the content are from the time of publication.

The Importance of Early Detection and Treatment

―― According to a research firm, the AI market’s size in 2015 was approximately 3.7 trillion yen, but it is predicted to balloon to 87 trillion yen by 2030. Also, the 2016 Cabinet Office Basic Statistics reported that in Japan there are 3.93 million individuals with physical disabilities and 9 million individuals with intellectual or psychiatric disabilities. Dr. Matsunaga, your research specialty is hearing and balance disorders. How do you feel about disability support from the perspectives of medical practice and the front lines of research?

Matsunaga: Rehabilitation for individuals with disabilities is rapidly becoming sophisticated. I think this is due to advances in science, such as engineering techniques--I refer to human engineering based on human body mechanisms. In reality, however, most disabilities cannot be cured fundamentally. As a doctor, I hope to mitigate disabilities even a little or to stop them from worsening even though they cannot be cured completely, but this has rarely been done. Still, early detection and treatment can reduce the impact of disabilities that remain for a long time. Through early detection and intervention, although disabilities cannot be completely cured, their severity can be reduced at least slightly. This is what I am working on now in a practical sense.

―― Are early detection and treatment easy to do at medical facilities today?

Matsunaga: I think it is easier now compared to 20 years ago because the social system has improved. From zero back then until now, great advancements have been made. The reality, however, is that we still have a long way to go.

Tatsuo Matsunaga
Director, Medical Genetics Center (Department of Otorhinolaryngology)
Director, Division of Hearing and Balance Research,
Clinical Research Center, Tokyo Medical Center

―― Given such circumstances, are there expectations for AI-driven support at medical facilities?

Matsunaga: Yes, expectations are high. Disorders in the human body have very complex causes that greatly exceed doctors’ understandings and memories, and these disorders go through large changes. While I believe we should make the most of our strength, namely generating new ideas and concepts, our results will be much more effective if AI can help us analyze the huge volume of data, select the appropriate pieces, and put them together. For example, when we decide if hearing loss has been caused by genetic changes, we examine a wide variety of data and read papers while following internationally standardized rules. For this, I think we can use AI.

―― Mr. Tamura, you used AI to develop a finger braille machine* to help individuals with visual and hearing impairments. Would you tell us about the background to that development and describe the machine’s strengths and its degree of acceptance?

Tamura: The deaf-blind communication support device that I developed was named the Finger Braille Converter** by Dr. Matsunaga (Figure 1). I was originally a patient at his hospital. I created the device after hearing multiple doctors say, "There is an app for the hearing impaired that uses voice recognition capabilities. Small changes to this app could also make it available to individuals with visual and hearing impairments." When I showed the device to Dr. Matsunaga, he said, "This is fascinating. Let's improve it together." He was particularly interested in using AI to realize disability support at low cost. If you hire a voice to finger braille translator, it costs several thousand yen per hour. This machine makes it possible for individuals with visual and hearing impairments to achieve independence, and that is what I think is this finger braille machine’s significance.

Figure 1 Characteristics of the Finger Braille Converter (support by AI)

―― How did people with visual and hearing impairments react?

Tamura: I used the smartphone-based prototype finger braille machine to see whether I could communicate with them without helpers. When I asked, "Do you understand what I am saying?" they replied, "Yes." In the future, these people may be able to enjoy news programs or TV and radio dramas.

―― In addition to hearing impairments, there are various other types of impairments. What kind of needs do you perceive?

Tamura: I conducted a barrier-free design survey based on the thought that individuals who have limited abilities to obtain information, including those with disabilities, may go out and use the passenger facilities of public transportation systems. There are information barriers. For example, not only individuals with disabilities but also elderly people and those from overseas who visit Japan cannot read the signs or understand announcements; thus, they cannot obtain information. Meanwhile, there are mobility barriers which make it difficult for people to descend staircases or to move around due to differences in level. They may also feel that the space between the train and platform is dangerous. To resolve all these issues with a single measure, I thought I could use an AI-based app as a means to provide information to meet various needs.

Ryo Tamura
Consultant, Business Science Group
Fujitsu Research Institute

Kamenoi: Among the five human senses, AI usage has made significant advancements in terms of the visual and auditory senses. Deaf-blind communication support devices like the one developed by Mr. Tamura also make the most of AI to employ voice recognition technology, which can now recognize different people’s voices more accurately. I think it is important to clarify needs to decide how far such basic technology should evolve in order to fulfill the intended purpose.

From Rare to Common Diseases: The Benefits of Disability Support

―― Dr. Matsunaga, you study rare diseases***. Will your research results apply to common diseases such as glaucoma, cataracts, or hearing loss in elderly people?

Matsunaga: I have high expectations. Studying individuals with common diseases is the initial approach. Note that in common diseases, the inability to see or hear, for example, can be caused by various factors. On the other hand, with rare diseases there is a one-on-one relationship between the cause and the disease. Therefore, the disease mechanism is clear. Thus, studying such diseases provides clear answers about each mechanism, such as why this occurs or which treatment is effective for that. If you see everything clustered together, you cannot understand anything. So in rare disease research, we break down clusters into elements and examine them. Studying each element may lead us to come up with a better approach to the entire cluster.

―― I understand that rare disease research broadens researchers' perspectives and leads to the treatment of diseases that affect the elderly. Are there any examples of a single research outcome being applicable to other areas, expanding the scope of technological application?

Kamenoi: When the megaquake occurred, an automobile manufacturer displayed the drive performance data that it collected on a map. Information such as "We can drive on this road up to that point" and "This road is blocked" helped relief efforts in the affected areas. Why not use the same idea to collect and analyze wheelchair travel data? Doing so will enable us to show safer, more wheelchair-accessible routes to elevators. If we take the same approach to wheelchairs as rare diseases, and take the same approach to walkers and elderly people as common diseases, I believe we can significantly widen the range of technological applications. Additionally, if we can use image recognition technology to make electric wheelchairs automatically detect red lights and stop, users can move and feel safer. Also, we can ease the lives of pedestrians with a white cane by providing information about their surroundings.

A Vision of Social Inclusion: A Society with No Boundaries between People with and without Disabilities

―― Thus far, our discussion has assumed that people with disabilities are different than those without, but there is also the concept of social inclusion****. This refers to creating a society in which everyone is equal and there are no boundaries between those with and without disabilities. Prior to aiming to realize this concept of social inclusion, we should consider what kind of measures we can implement using AI technology. My guess is that AI can promote social participation and employment of individuals with disabilities. How should we view the relationship between AI and employment?

Kamenoi: There have been sensational articles stating that AI will steal jobs, but I think it is actually a positive for disability employment. For example, disability employment will increase if individuals with disabilities can use voice-to-text technology and obtain information in real time without an interpreter and create a work environment in which they can work independently. Companies will more readily accept them, too. Today’s speech recognition technology is making advances in multiple languages. If a non-Japanese language is recognized, another technology can translate it. This will create work infrastructure by which these individuals can work alongside someone who is not a native Japanese speaker.

Chizuko Kamenoi
Senior Managing Consultant, Business Science Group
Fujitsu Research Institute

Matsunaga: When individuals with disabilities play an active role in society, it benefits not only them but society as a whole. Therefore, I hope that those without disabilities are aware that this matters for them too, instead of thinking that it is not a concern of theirs.

―― Before, Dr. Matsunaga told me that, though people troubled by rare diseases are found throughout Japan, those who live far from Tokyo feel lonely and isolated because information is concentrated in Tokyo and they do not know who to talk to or how to talk to them. Will AI help to increase the inclusion of these individuals who cannot access information sufficiently?

Matsunaga: People away from Tokyo previously were inconvenienced because they were far away, there were no people around them, and they had little information. However, gathering and analyzing information using AI and ICT, which used to be difficult, are now highly feasible.

Tamura: With the help of medical care and AI, I can work as a consultant. I cannot hear anything around me in my natural state. I am impressed by the fact that receiving a cochlear implant that uses AI-like technology to adjust the stimulus to the auditory nerves in response to the surrounding sound has enabled me to get a job that was once considered difficult for the hearing impaired. I think this has made my parents happier than anyone else. Not only those who contract a disease but also those around them may sometimes feel deep sadness, fear, or anxiety. I feel that the advancement of AI and medicine will lead to an era in which such people can break away from these kinds of emotions. Social inclusion aims not only to achieve the social participation of individuals with disabilities like myself but also the creation of an integrated society that overcomes distinctions based on race, religion, faith, gender, social status, disability, and sexual orientation. Therefore, it is desirable for society to always enable all elderly people, individuals with serious illnesses, expecting mothers, and people who are raising children to play relevant roles. Before we can realize such a society, many problems must be addressed. As a consultant with disabilities, I will tackle them one by one while harnessing the power of AI and medicine.

―― Our consulting services usually focus on numbers; we explore ways to improve operational efficiency, to increase sales, and to reduce costs for customers. I think it is important to note that AI technology can evolve to help improve our lives.

Starting from Hearing Impairments and Medical Research: Possibilities for Further Breakthroughs

―― In the wider medical field, without limiting the conversation to hearing impairments, which issues remain to be solved or for which issues may ICT possibly bring about a breakthrough?

Matsunaga: Now that computers have become this powerful, I wonder whether they can simulate biology, meaning the various types of proteins or gene behavior in cells. We construct a hypothesis, conduct an experiment, clarify the mechanism, and then, for example, develop a drug or treatment method. Actually, hypotheses are based on our hunches. If we can rely on more scientifically based hunches instead of pure hunches, experiments will be more effective. In such experiments, if we input into computers information on how each protein or gene behaves in advance, the computers can reproduce that behavior without us actually having to keep and conduct experiments on animals. If computers can reproduce the primitive and fundamental components of human bodies, science may advance at a different pace on a different level than before, leading us to meaningful discoveries.

Tamura: If you can perfectly simulate by computer how an organ behaves at the cellular level, what kind of breakthroughs would you predict?

Matsunaga: We could know the effects of drugs without animal testing. This would also save animal lives.

Kamenoi: While animal testing may produce outcomes after a month or a year, computer simulation may shorten that cycle. In this sense, it may be possible to reduce the development lead time for treatment methods and drugs.

Matsunaga: We can expect greater cost effectiveness because we could predict effects without using expensive drugs.

Tamura: How will diagnoses and preventive medicine change?

Matsunaga: We make diagnoses by weighting and grading various elements. This takes time and effort to collect each piece of information. If a computer can do everything and we only have to check the workflow, we can make diagnoses much faster and more accurately.

Kamenoi: Advancement in that area will lead to early detection and treatment. As in the case of rare diseases, we can determine the cause and disease relationship.

Matsunaga: That will give us the ability to predict "this person will later have that kind of disease." This will allow for preventive measures such as lifestyle improvements in advance or drug administration in order to delay the onset of symptoms.

Tamura: Dr. Matsunaga previously conducted gene testing on me. What will medical care look like when gene testing and AI are used together?

Matsunaga: Humans have about 25,000 genes. Hundreds of changes in each gene are related to diseases. Although we cannot check them all, if AI can check the meaning of every single change and examine the relationship between such changes and the actual symptoms of the individual examined, the diagnostic yield will become much higher. Possible treatment methods will be found, and the individual can receive the most appropriate treatment and preventive measures. That said, the number 25,000 seems surprisingly low to us.

Tamura: I am from Yamagata prefecture, and there were no large hospitals near where I lived. Thus, I could not receive high-level medical care until I moved to Tokyo. I think that in the future, ICT and AI can be utilized so that people on remote islands and deep in the mountains can still receive the best care. There will come an era in which people can see the best doctor like Dr. Matsunaga even at clinics in remote areas.

[Moderator]
Yoshiyuki Okihara
Principal Consultant, Business Science Group
Fujitsu Research Institute

―― The possibility that two people can meet one another like Mr. Tamura met Dr. Matsunaga gives us great hope. Thank you very much for your time today.

*: Finger braille: A communication method for individuals with visual or hearing impairments. The keys of a braille typewriter are allocated to the fingertips, and two people communicate with their hands. One person uses the index, middle, and ring fingers of both hands as six dots to create a braille character and reproduce it by making the appropriate combinations on the fingers of the other person by tapping.

**: Finger Braille Converter: A prototype finger braille machine named after the inventor of braille, Louis Braille.

***: Rare diseases: A collective term for diseases that have a small number of patients. In Japanese, they are also referred to as "rare incurable diseases." It is an antonym of "common diseases," which include many types of lifestyle diseases and infectious diseases. While the prevalence of common diseases is obtained from the general population, that of rare diseases is obtained from a population of 100,000.

****: Social inclusion: The concept of protecting all citizens, including the socially disadvantaged, from exclusion, friction, loneliness, and isolation; including them as members of society; and supporting them.