PAC is committed to lead clinical and computational operations around cognitive neuroscience. With that objective in mind, we are developing various centers listed below. We also provide access to non-MD practitioners for a broad range of screening, testing and treatment modalities traditionally reserved for medical doctors.  

Within that business model, we use artificial intelligence for a number of our applications through which we provide those services. In this sense, AI is really embedded across the entire company ecosystem.

We started out several years ago offering only blood testing, then we stood up a team of applied researchers with the goal of looking at all possible advancements in the entire field of machine learning and figuring out what of that realm was at all relevant to medicine. Then after we narrowed it down from all those possibilities, we had to decide which tools apply directly to our data. But we couldn’t stop there.  Our next goal is to actually generalize that data, build it into some of our clinical platforms to both make it available to our clinicians company-wide, but ultimately worldwide.

At the moment, we have a very strong interest in tabular machine learning. In this particularly medical domain of deep learning, many of our data problems are formulated within this realm of tabular data, which by the way, has a significant presence in the broader industry. Banks use it and most businesses run on it.

Another example of our deployment of AI is anomaly detection, important because we come across novel biomarkers for existing diseases all the time.  Because we serve conventional medicine, as well as practitioners of functional, chiropractic, naturopathic and Traditional Chinese Medicine, we have access to a greater pool of data, thus increasing our anomaly detection capabilities.  We might use supervised machine learning where we build these massive models to examine all of the data and determine the probability that a certain biomarker correlates with a certain disease; But in addition to that, we realize that there is a broad range of anomaly detection algorithms that would allow us to capture emerging biomarker trends for yet undiscovered diseases and conditions that might be escaping what supervised models have seen in the past. Supervised models can only generalize based on what they know, so if it is a new syndrome of biomarkers, in other words a new fact pattern, it will escape detection by supervised models that are not trained to find it.

Now beyond those areas, which are the ones that we are pressing and focused on at the moment, there are also areas which are more future-looking game-changers. Often when we are looking at those areas, we partner with academic institutions, such as the Harvard Dataverse Repository, which houses a free collection of scientific literature open to all researchers.  This repository helps us flesh out the ideas, perform some of the more experimental research, publish papers and engage with the research community.  Communities like this have some really nice benefits, one is that if it is an area that’s underexplored in the research community, by funding, then publishing it, we catalyze more interest, leading to more research, publications, and ultimately more funding.

For example, Transcranial Magnetic Stimulation is a treatment for anxiety and major depressive disorder, with a greater efficacy, with a lower side-effect profile, proven in a head-to-head clinical trial against four antidepressants is still underutilized today despite being developed by Dr. Yi Jen, MD, PhD and approved by the FDA in 2008. To get the word out, we have partnered with Dr. Jen who has granted us first right of refusal to publish findings resulting from the clinical treatment of patients in our new scientific journal Mind the Gap: Journal of Bioelectronic Brain Medicine.