|Artificial intelligence is increasingly becoming more sophisticated at doing tasks humans normally do more quickly, more efficiently, and at a lower cost at that / Photo by: taa22 via 123RF|
Artificial intelligence is increasingly becoming more sophisticated at doing tasks humans normally do more quickly, more efficiently, and at a lower cost at that. One of the industries that AI continues to influence is healthcare. The technology is being used to detect diseases, take a more comprehensive approach for disease management, streamline drug discovery and drug repurposing, and many others.
PricewaterhouseCoopers, a global network of firms delivering world-class assurance, tax, and consulting services for businesses, reported that 39% of people are willing to engage with AI/robotics for healthcare. Of all the age groups, those who are 18 to 24 are the most willing at 55%. The willingness decreases by age group with just 33% for those over 50 years old. It was also reported that the countries that are far more willing to engage with AI for healthcare in the UK are Wales and Scotland at 52% and 47%, respectively.
PwC’s “What Doctor? Why AI and Robotics Will Define New Health” report showed that 55% of the respondents stated that they are willing to use advanced computer technology or robots with AI that perform tests, answer health questions, make a diagnosis, and recommend treatment. AI is now also being trained to predict diseases like epilepsy.
People with epilepsy have lived a life fearing they might black out anytime. It could happen while they are traveling, taking a bath, working out, or even sleeping. Reports showed that about 50 million people across the world have epilepsy. Neurologist Mark Cook stated that epilepsy is an “electrical problem” in the brain. Normally, the human brain constantly sends out electrical signals, but the brains of people suffering from epilepsy emit extra bursts of electricity. This causes a sort of electrical storm and manifests in seizures.
There are several different kinds of seizures, including losing consciousness, convulsing, blinking or twitching, experiencing difficulty speaking, and more. The main problem with epilepsy is that it’s extremely hard to predict. What if sufferers could receive a “seizure forecast”—a real-time prediction that would let them prepare for a seizure?
Predicting Seizures Through a Novel AI System
Epilepsy is a debilitating disorder that’s hard to predict. According to reports, 30 to 40% of people with epilepsy have an intractable form of the disease that doesn’t respond to medication. This means that they could have a seizure at any time, at any place, and often without warning. But even people with controlled epilepsy can also suffer from spontaneous seizures despite their symptoms being controlled by medicine most of the time.
This disorder can put the lives of people in danger as they are 15 to 19 times more likely to drown while bathing or swimming than the rest of the population. At the same time, they tend to suffer from anxiety and depression because of the aftereffects of seizures, side effects of medication, and the strain of dealing with the condition. These problems can now be addressed by AI.
Recently, two AI researchers from the University of Louisiana at Lafayette developed a new AI system that can detect epileptic seizures with 99.6 percent accuracy up to an hour before they occur. According to Interesting Engineering, a cutting edge, leading community designed for all lovers of engineering, technology, and science, the researchers combined EEG (electroencephalogram) technology and predictive modeling. Both of these techniques have been used in previous studies.
|Epilepsy is a debilitating disorder that’s hard to predict. According to reports, 30 to 40% of people with epilepsy have an intractable form of the disease that doesn’t respond to medication / Photo by: rawpixel via 123RF|
The study published in IEEE Transactions on Biomedical Circuits and Systems is lauded as “a major leap forward from existing prediction methods.” Researchers Hisham Daoud and Magdy Bayoumi were able to get earlier and more accurate seizure predictions by bringing together extraction and classification processes into a single automated system. They also incorporated another classification approach to further boost the accuracy of the AI system.
According to Unite.AI, an online site that offers detailed analysis and news on the latest robotics and AI breakthroughs, the researchers used a deep learning algorithm to extract and analyze the spatial-temporal features of the patient’s brain activity from different electrode locations. They also identified the most appropriate predictive channels of electrical activity through an additional algorithm. However, the AI system needs to be trained on each patient.
“In order to achieve this high accuracy with early prediction time, we need to train the model on each patient. This recording could be [done] off-clinic, through commercially available EEG wearable electrodes,” Daoud said.
As of now, the researchers are developing a customized computer chip to process the algorithms. “We are currently working on the design of efficient hardware [device] that deploys this algorithm, considering many issues like system size, power consumption, and latency to be suitable for practical application in a comfortable way to the patient,” Daoud added.
Machine Learning and AI-Powered Algorithm in Predicting Seizures
A 2018 study was conducted by researchers from the University of Sydney aimed at developing a portable, affordable, and non-surgical device that will give a reliable prediction of seizures for people living with treatment-resistant epilepsy. They used advanced AI and machine learning that can alert epilepsy sufferers within 30 minutes of the likelihood of a seizure.
Science Daily, an American website that aggregates press releases and publishes lightly edited press releases about science, reported that the team had proposed a generalized, patient-specific, seizure-prediction method to predict when seizures will strike that will not require surgical implants. The study published in Neural Networks used three data sets from Europe and the United States. The researchers built a predictive algorithm using the data with a sensitivity of up to 81.4% and a false prediction rate as low as 0.06 an hour.
"While this still leaves some uncertainty, we expect that as our access to seizure data increases, our sensitivity rates will improve," said Dr. Omid Kavehei from the Faculty of Engineering and IT and the University of Sydney Nano Institute.
AI has truly done wonders in the healthcare industry. This not only helps save lives and address the current issues with epilepsy but also opens the doors for future researchers to conduct further studies that will maximize the benefits of using the technology to cure illnesses.