Artificial intelligence, or AI is something we hear a lot about today. In this interview with Life
Extension’s Michael A. Smith, MD, Kristen Willeumier, PhD, provides some insight into AI technology and its relationship with psychiatry which, along with neurology, studies and treats diseases of the brain. Dr. Smith predicts that AI will soon be an important part of how we understand and treat disease. According to Dr. Willeumier, some of that technology is now “ready for prime time.” Download this Live Foreverish podcast episode for FREE on iTunes!
What is artificial intelligence?
Artificial intelligence is, simply, the intelligence of machines as opposed to human or animal intelligence. According to the New World Encyclopedia™, “Artificial intelligence (AI) is a branch of computer science and engineering that deals with intelligent behavior, learning, and adaptation in machines. John McCarthy coined the term to mean ‘the science and engineering of making intelligent machines.’”1
How can artificial intelligence be used in psychiatry?
Dr. Willeumier served as director of research for the Amen Clinics, a group of mental and physical health clinics that treat mood and behavior disorders, from 2009 to 2016. The clinics utilize single photon emission computed tomography (SPECT) imaging of the brain to aid in the diagnosis and treatment of patients.2
“For those of us who are in the field of psychiatry, we understand that the field as a whole has a lack of effective biomarkers, such as neuroimaging biomarkers and genetic biomarkers that help us figure out how to treat psychiatric disorders,” Dr. Willeumier lamented. “Many of the disorders are very complex and highly comorbid, meaning that anxiety and depression will go together, or you might have ADHD (attention deficit hyperactivity disorder) and anxiety. As clinicians, we’re faced with multiple treatment options. Because a lot of times the symptoms can cluster together, it can be very challenging to differentiate what type of disease someone has.”
Dr. Willeumier provided the example of traumatic brain injury and posttraumatic stress disorder, which can have overlapping symptoms yet very different treatments.3 Artificial intelligence may become a useful tool to provide diagnoses in cases such as these.
What psychiatric conditions can AI be used for?
In people with depression, specific neurobiologic patterns in the brain are associated with different types of the disease. Different patterns necessitate different targeted treatments to get the best outcome. Machine learning can rapidly diagnosis the type of depression a patient may have and determine the treatments that are most effective for an individual. Neuroimaging data or clinical variables from patient reports provide data that can be analyzed by a machine that learns what patterns in the brain correlate with specific clinical symptoms and which medications are the best targeted treatments, which can eliminate the need for depressed patients to go through the process of trial and error to find out which treatment is best for them.
“Machine learning is designed to look for very specific patterns,” Dr. Willeumier explained. “When we use neurobiology and brain imaging to target treatment for patients with depression, we saw that there were various types. For example, if we saw increased activity in the anterior cingulate gyrus, we knew that those patients are probably going to respond best to SSRIs [selective serotonin reuptake inhibitors]. Alternatively, if we see low functioning in the prefrontal cortex and they have symptoms of depression, they might respond more positively to an SNRI [norepinephrine reuptake inhibitor].”
Alzheimer’s disease is another brain disorder that has benefited from AI. This progressive, devastating illness still can’t be diagnosed with certainty in living people and can only be verified during examination of the brain at autopsy. However, researchers are using brain imaging data to train machines to detect whether a living individual has Alzheimer’s disease and how it could progress.4 Data concerning sleep and movement patterns that are used to predict the risk of falls are now being analyzed for changes in movement that can predict Alzheimer’s disease, which needs to be detected early if available therapies that can slow its progression are to have the best chance of being effective.
Autism spectrum disorder is a condition that can benefit from machine learning due to its diverse clinical picture. Machine learning was able to detect autism with 81% accuracy in a comparison of neuroimaging results from 928 individuals diagnosed with autism and those of 100 healthy control subjects.5 This tool will enable brain scanning to be used in the provision of accurate autism diagnoses, which can sometimes take a year or more for a patient to receive. Early diagnosis can improve the quality of life for individuals with autism.6
Coming soon to a clinic near you
When asked “where we’re at” with this futuristic technology, Dr. Willeumier unhesitatingly replied, “This is ready for prime time.”
“AI is being used in multiple platforms across multiple diseases, and I think there’s a lot to be excited about,” she enthused.
Dr. Smith and Dr. Willeumier agreed that technology is all about improved outcomes. “I think we can harness the power of technology to help us live healthier lives,” Dr. Willeumier concluded.
About Live Foreverish: Join Dr. Mike as he sits down with some of today’s leading medical, health, and wellness experts to discuss a variety of health-related topics. From whole-body health to anti-aging and disease prevention, you’ll get the latest information and helpful advice to help you live your life to the fullest. See the full list of Live Foreverish Podcast episodes, available on demand.
- “Artificial intelligence.” New World Encyclopedia. 1 Aug 2017, 15:17 UTC. 3 Jan 2019, 00:14 http://www.newworldencyclopedia.org/p/index.php?title=Artificial_intelligence&oldid=1006061
- Amen DG et al. J Psychoactive Drugs. 2012 Apr-Jun;44(2):96-106.
- Raji CA et al. Brain Imaging Behav. 2015 Sep;9(3):527-34.
- Mirzaei G et al. Rev Neurosci. 2016 Dec 1;27(8):857-870.
- Amen DG et al. J Syst Integr Neurosci. 2017 Apr 10;3(3):1-9.
- Elder JH et al. Psychol Res Behav Manag. 2017 Aug 24;10:283-292