Medicine is a field that largely depends on an adequate dialogue between physicians and patients. A faithful understanding of a patient's problem is essential for an accurate diagnosis. The main tools used to confirm the presence of problems are body fluid tests and imaging such as MRI and X-ray.
Read more: Does the diagnosis of depression guarantee the right to any benefit from the INSS?
see more
Secret of youth? Researchers reveal how to reverse…
The "powers" of porridge: check out the benefits of oats in…
However, there is a potential new pathway that could be useful in diagnosing many diseases. Researchers are working to develop artificial intelligence systems that can use timbre as a diagnostic tool to pinpoint disease.
How can the voice be a diagnostic tool?
Illnesses can affect organs such as the heart, lungs, brain, muscles or vocal cords, which can alter a person's voice. Voice analysis through artificial intelligence opens up new possibilities for the health system through the use of audio biomarkers for the purpose of preventive risk diagnosis and remote monitoring of clinical outcomes and symptoms. With that in mind, there may be several possible uses of the voice for health-related purposes.
There is enormous potential in this environment from both a patient and clinic point of view.
The voice, which is a complex set of timbres that comes from our vocal cords, contains a large amount of information and plays an important role in social interactions. It allows us to share insights about our feelings, fears, emotions and excitement.
You can adjust volume or tone. Virtual and voice assistants on smartphones or smart home devices like connected speakers are now popular and have paved the way for this new reality. Advances in speech technology, speech signal analysis, natural language processing, and comprehension techniques have opened up for many potential speech applications, such as identifying speech biomarkers for diagnosis, classification or monitoring remote.
Artificial intelligence will use the voice as a diagnostic tool
Researchers are now developing AI-based tools that could eventually diagnose serious illnesses. They target everything i.e. from Alzheimer's to cancer.
The project funded by the National Institutes of Health, announced last Tuesday, aims to to transmute the human voice into something that can be used as a disease biomarker, such as blood or temperature.
According to the National Institutes of Health website, they will invest $130 million over four years, depending on the availability of funds, to accelerate the widespread use of artificial intelligence (ANDAVA) by the biomedical and research communities behavioral.
The Joint Fund of the National Institutes of Health’s Bridge to Artificial Intelligence (Bridge2AI) program brings together team members from diverse disciplines and backgrounds to create tools, resources and rich data that respond to methods of AI. At the same time, the program will ensure that its tools and data do not perpetuate inequalities or ethical issues that may arise during data collection and analysis.
According to Olivier Element, a professor at the Institute for Computational Biomedicine at Weill Cornell Medicine and one of the leading project researchers, the great thing about voice data is that it is one of the most powerful and inexpensive types of data that can be collected of people. It is also very affordable and easy to take for any patient. This is useful when creating large databases.
Yaël Bensoussan, an otolaryngologist at USF Health and another principal investigator on the project said that while there have been similar efforts in the past, most of them were too small to be effective. The lack of an adequate database was also an important factor. Because this is a comparatively new area of study, researchers have not yet discovered the best information gathering practices for this system. Ongoing projects will establish data collection standards for this.
The team starts by creating an app that collects voice data from participants with vocal cord paralysis, Alzheimer's disease, Parkinson's, depression, pneumonia and autism. A physician will be assigned to monitor all collections. “For example, someone who has Parkinson's disease will find that their voice can be lower and that their speech is also slower,” says Bensoussan.
The application requires you to record an audio, read a sentence and read an entire text. According to some doctors, you can tell that patients have brain metastases by the way they talk.
How is diagnostic speech data protected?
The poll team is collaborating with artificial intelligence company Owkin to build and train the artificial intelligence models in the project. Under the Owkin framework, collected patient data remains at the center where it was collected while the AI model travels between institutions. The model is trained separately on each dataset, and then the results of those trainings are returned to a central location where they are mixed.
This provides an additional layer of privacy for voice data. A team of bioethicists is working on the project to examine the ethical and legal implications of a language database and language-based diagnostics. This means you might wonder if your voice is protected under the Health Insurance Portability and Accountability Act (HIPAA), if your patients have voice data, etc.
Is it possible to use voice data for diagnosis?
Speech data is already useful for diagnosing and treating voice disorders. All of this appears to have tremendous potential for treating mental health disorders such as depression, anxiety, post traumatic stress disorder and many others also in this area.
Collecting voice samples from veterans and analyzing vocal signals such as pitch, rhythm, frequency and volume can be useful for look for signs of invisible injuries such as post-traumatic stress disorder (PTSD), traumatic brain injury (TBI) and depression. By using machine learning to explore functions in voice, algorithms are also being worked on that select voice patterns in people with these disorders and compare them with voice samples from people healthy.