Researchers at a Japanese university have devised an artificial intelligence (AI) based test to help detect early signs of dementia in patients.
It is hoped that the new clinical system will lead to better treatments of the disease earlier, as it doesn’t involve invasive medical imaging.
Currently, scans are used in conjunction with cognitive function tests, such as the Mini-Mental State Examination (MMSE) – a combination that can be expensive, time-consuming, and distressing for some patients. However, an ageing population means that more people are developing the disease, hence the need to pursue faster, smarter, and easier-to-use tests.
Helping patients to identify problems sooner may also help them stave off the effects of some conditions, and manage them more effectively.
A group of researchers from Osaka University and Nara Institute of Science and Technology have now demonstrated that it is possible to detect dementia simply from conversations between humans and AI agents.
Avatars
The technique uses machine learning to study the way that elderly people speak when answering simple questions from a computer avatar.
The model deploys the interactive characters to create a dialogue with elderly patients, recording their speech patterns, language, and faces. The AI system then evaluates a subject’s gaze, response times, and intonation, and analyses which nouns and verbs they are using.
According to tests of the machine learning system, it was able to distinguish those suffering from dementia from healthy people in a control group. Researchers claim that the success rate was around 90 per cent, from asking just six questions.
The questions are based on internationally recognised criteria contained in the Diagnostic and Statistical Manual of Mental Disorders.
Senior research author Takashi Kudo said, “If this technology is further developed, it will become possible to know whether or not an elderly individual is in the early stages of dementia through conversations with computer avatars at home on a daily basis. It will encourage them to seek medical help, leading to early diagnosis.”
Plus: Flexible ‘postage stamp’ can take blood pressure
In related health technology news, researchers have developed a flexible, adhesive patch that can monitor a patient’s blood pressure, according to the MIT Technology Review. Made of silicon elastomer, the postage-stamp sized wearable works by sending ultrasonic waves into the skin, which reflect off the wearer’s bodily tissues and blood.
In theory, the patch could be used to monitor patients at home, with the data collected over time and analysed on a laptop. As well as avoiding the need for multiple appointments, uncomfortable pressure tests, and invasive procedures, the wearable may help cut costs and reduce the risk of infection.
The system is being developed at the University of California, San Diego.
Additional reporting: Chris Middleton.
Internet of Business says
The combination of wearable devices and AI has been a major development hotspot in 2018, as these recent Internet of Business reports explain:-
- Read more: AiServe develops A.I. wearable for the blind and partially sighted
- Read more: Health IoT: Wearable can predict older adults’ risk of falling
- Read more: Healthtech: Wearable helps injured athletes recover faster
- Read more: Mind-reader: MIT’s AlterEgo wearable knows what you’e going to say
- Read more: Health IoT: New wearable can diagnose stomach problems
- Read more: Health IoT: Scientists develop diet wearable – for your teeth
- Read more: Consumer wearables can detect major heart problem
Our Internet of Health conference is taking place in Amsterdam on 25-26 September. Click the logo, below, for more details.
