AI In HealthCare: Leaps Into Heart Attack Detection & Disease Diagnosis

Artificial intelligence could extract features from voice and train itself through machine learning to detect various conditions and help in real-time diagnosis and saving lives.

AI In HealthCare

Artificial intelligence (AI) has made considerable leaps over the past few years and implemented in variety of applications across various sectors. Scientists, experts, and executives of top tech firms have expressed their concerns about the advancements in AI. Some of them believe it could cause destruction, however, some of them casted it as a blessing to human beings. Companies have invested enormously into the AI research by setting up research labs and recruiting experts to design and implement the technology for various applications such as self-driving cars or virtual assistance. Now, AI has been emerging in the healthcare sector by playing a crucial part in detection of heart attack and diagnosis of diseases.

AI could lead doctors to monitor and detect health issues remotely with the help of smartphones and wearable devices. But speech could also be significant factor in detection of conditions. When someone gets cardiac arrest outside hospital, time is significant. The most critical step is the first step, which is to determine if it is a cardia arrest. Dispatchers on the phone need to rely on panicked friend or relative to determine the conditions. But their task will be eased up with the help of AI. In Copenhagen, Denmark, dispatchers will have an AI assistant Corti, which would use speech recognition software to transcribe the interaction and machine learning to analyze speech and other clues to determine heart attack. Dispatcher will get a real-time update from Corti, which trains itself in each phone call and improves its model.

Scientists have been researching on detection of mental health disorders such as post-traumatic stress disorder (PTSD), depression, and traumatic brain injury (TBI) with the help of AI. They decided to use voice to extract features such as tone, rhythm, pitch, and volume and use machine learning analysis in detection and diagnosis of such diseases. They have been collecting voice samples from war veterans and people with these conditions. The algorithms will compare features in their voices with features in voices of healthy people to diagnose these diseases.

Charles Marmar, chairman of the department of psychiatry at New York University’s Langone Medical Center collaborated with researchers at SRI International, a nonprofit research institute in Northern California. The research team picked 30 characteristics in voice of people with PTSD and TBI out of 40,000 features that were extracted through machine learning. They conducted a study in 2015 in which they studied 39 men. The voice test gave 77 percent of accuracy in differentiating between patients with PSTD and healthy patients. After the study, they collected more samples till now and their research advanced up to differentiating patients between PTSD and TBI. Marmar outlined that accuracy in detection of diseases increases with vast amount of data that can be analyzed through AI.

Beside mental health conditions, the Mayo Clinic has been developing vocal biomarkers to enhance remote health monitoring for heart disease. It collaborated with Israeli firm Beyond Verbal to examine voices patterns of patients with coronary artery disease, which is the most common type of heart disease. They opined that chest pain during the disease affects voice production. Researchers collected voice samples from patients at the risk of coronary heart disease and analyzed with the help of machine learning. They found one feature related to the frequency of voice that could increase likelihood of coronary heart disease by 19 times. This feature was not discernible by human ear, but the machine algorithm could detect it efficiently. They found that specific patterns in voice could predict the amount of blockage detected by angiography.

The advancement of AI in healthcare sector could offer a low-cost solution to some health conditions. Moreover, it could prove vital in saving lives during critical situations.


Please enter your comment!
Please enter your name here