Diagnostics in the chat
Daily conversations with AI will help you recognize the early signs of Alzheimer's disease
VictoriaSeredina, IBM blog, Habr
There is no cure for Alzheimer's disease. But the earlier it is diagnosed, the more likely it is to slow down the progression.
A group of researchers from IBM and Tsukuba University (Japan) has developed an AI model that helps to detect the onset of mild cognitive disorder (LCR, a transitional stage between the natural aging process and dementia) by asking elderly people everyday questions. An article published in the journal Frontiers in Digital Health (Yamada et al., Tablet-Based Automatic Assessment for Early Detection of Alzheimer's Disease Using Speech Responses to Daily Life Questions) provides the first empirical evidence of the effectiveness of automated assessment of the condition of patients based on speech analysis using tablets.
Unlike previous studies, this AI-based model analyzes speech responses to everyday questions using an app on a smartphone or tablet. For example, patients may be asked questions about their mood, plans for the day, well-being or last night's dinner. Earlier studies mainly analyzed speech responses during cognitive tests. For example, patients were asked to count down from 925 in increments of three or describe the image shown to them in as much detail as possible.
The researchers found that the accuracy of detecting early signs of LCR in tests with answers to such simple questions was almost 90%. This is comparable to the results of more complex cognitive tests. Thus, to monitor the state of health, this AI algorithm can be integrated into intelligent speakers or home devices so that during their operation it is possible to monitor the signs of LCR.
The results of this study seem promising, since cognitive tests are not an easy test for patients. They are forced to follow complex instructions and are subjected to high mental loads. Therefore, such assessments are not carried out so often, which may prevent the detection of early signs of Alzheimer's disease. The analysis of everyday speech data allows the assessment to be carried out much more often, while it requires less resources.
For the purposes of the study, speech responses of 76 elderly Japanese, including those with LCR, were collected. Then the scientists analyzed the features of their speech: the pitch of their voice, the frequency of pauses during a conversation, etc.
The group of researchers knew that it is not easy to identify subtle cognitive differences in everyday speech with a low level of cognitive load. Under such conditions, the differences in the speech of healthy people and people with LCR are less noticeable than when conducting cognitive tests.
IBM and Tsukuba University scientists have coped with this complexity by combining the use of answers to multiple questions designed to identify changes in the ability to remember information and perform purposeful actions, as well as changes, in fact, in speech function, indicating LCR and dementia. For example, an AI application asks the question: "What did you eat for dinner yesterday?" An elderly person with LCR will answer: "Japanese noodles with tempura – tempura of shrimp, radish and mushrooms."
At first glance, there is nothing wrong with the answer. But AI reveals differences in voice pitch, pauses in speech and other acoustic characteristics. The researchers found that, compared with cognitive tests, questions from everyday life help to identify less obvious, but statistically significant differences in speech features, indicating mild LCR. The accuracy of detecting mild cognitive disorder using this AI model was 86.4%, which is comparable to the accuracy of cognitive tests.
Portal "Eternal youth" http://vechnayamolodost.ru