The US National Institutes of Health will provide $14 million to support the use of machine learning in diagnostics. It is assumed that artificial intelligence will be able to analyze the voices of patients to diagnose and study diseases, reports The Register.
Twelve research institutions led by the University of South Florida will receive the money over the next four years. They will collect a training database of people’s voices with privacy in mind. This database will be used to train programs to recognize common features in the voices of patients diagnosed with certain diseases. Eventually, doctors will be able to use this to detect diseases and neurological disorders by studying human speech.
The Voice as a Biomarker of Health project will focus on software that can detect five types of diseases:
“Our team chose the five categories of diseases based on existing work in voice AI that has been published over the last 20 years,” said the project leader and assistant professor at USF’s Department of Otolaryngology Yael Bensoussan.
Recent advances in machine learning algorithms for analyzing voice and speech data have shown how the technology can be used to assess physical and mental health. For example, a study conducted by researchers from the Massachusetts Institute of Technology linked nervousness and tremors in speech with depression and anxiety.
Scientists believe that the results of such plans are promising, so listening and processing the sound of speech or breathing with the help of artificial intelligence can become an inexpensive method of detecting diseases and disorders in the early stages.
“Voice is one of the cheapest biomarkers to study,” Bensoussan said. “When you think of biomarkers such as genetic testing or imaging like MRIs or scans, they are all quite resource-intensive and can be invasive in a sense. CT scans cause radiation for patients, for example. Voice is the easiest biomarker to collect, does not cause any physical risk for patients, and can be collected in very low resource settings especially with modern technology.”
In its first year, the US National Institutes of Health will provide $3.8 million to the Voice as a Biomarker of Health Initiative for participants to build a large, diverse voice database. It can be evaluated in conjunction with other data collected through medical imaging and genomics. Language data from patients will be recorded in a clinical setting during a pilot study during the first year.
To make sure sensitive data is kept safe, the models will be trained using federated training with support from Owkin, a startup focused on helping biomedical research with machine learning software.
Additional funds for this initiative, in addition to the planned $14 million, can be provided with the approval of Congress. And this project is part of a broader effort by the US National Institutes of Health to accelerate the introduction of artificial intelligence into research and development, with the hope that new technologies will update health care in the USA. The organization has pledged to invest $130 million over four years in multiple projects aimed at creating flagship biomedical datasets, universal software tools and resources for training AI health researchers.