BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Healthcare Outcomes Improve When Technology Imitates Nature

This article is more than 9 years old.

Computers are increasingly integrated into almost every aspect of our health and well-being.  But computers usually just do what you tell them to do.  Their programming dictates exactly how they behave, and if you want to have a computer act in an intelligent way, you have to anticipate and design every contingency into their programming.  For decades, people have dreamed of making computers that can “learn” from experience and eventually become more than the rules that were contained in their original programming.  This concept is called machine learning.

Recently, advances in computer algorithms have made it possible to come closer to achieving this lofty goal of machine learning.  A key element of this success has come by taking inspiration from nature to incorporate biological processes into the software.  One of these inspirations relies on simulating natural evolution, a method known as evolutionary computing.

As most of us know, the theory of natural selection explains, in part, how species evolve over time to adapt to their environment.  The idea is that individuals have varying traits and these variations may make an individual better able to survive in its environment (for example, better natural camouflage to hide from predators).  These unique individuals produce more offspring, further spreading their variation to future generations.

When applied to computers, evolutionary computing applies the same principles found in nature to an engineering or mathematical problem.  Potential solutions to a complex problem are evaluated against a specific “fitness” function to determine how good each solution is.  Better solutions are allowed to remain and poorer solutions are discarded from the population.

Dr. Lawrence Fogel, a pioneer in artificial intelligence, wrote the first book on evolutionary computing in 1966.  He founded the company Natural Selection, Inc. (NSI) in 1993.  With Larry passing away in 2007, his two sons Drs. David and Gary Fogel, have carried on the family tradition of applying these novel computing techniques across a variety of domains.  Dr. David Fogel, President of NSI, offered some valuable insights as to how evolutionary computing has impacted, and will continue to impact, several aspects of the healthcare industry.  A few of the more interesting examples include:

Exploiting population-based data for improved drug development:  Evolutionary computing can help doctors determine what is the right drug for the right patient at the right time, an application being pursued by companies such as Theragence, Inc.  It can do this by accurately and efficiently gleaning insights from large volumes of data from diverse sources, including personal information about patients that include demographics and their genetic profile.  These techniques can help determine which patients from Phase II clinical data are most likely to benefit from a particular drug, and therefore help pharma companies better craft their Phase III trials for success.

Personalized medical treatments:  Personalized medical treatment has been generating more and more attention in recent years.  Evolutionary computing approaches can help tailor treatments for individuals by receiving and analyzing feedback about what is working and what is not.  For example, people who suffer from tinnitus often use other sounds or noises to mask the ringing in their ears.  Under funding from the National Science Foundation, NSI has used these techniques to adapt “sound masks” to help alleviate that ringing both while the patient listens to the mask and after it is turned off.

Improved understanding of emotional states and mental health:  The words that we use impact others’ emotions and also reflect our own emotional state.  Sentiment analysis software, such as EffectCheck®, can analyze the word choices we make and provide a reflection of a person’s emotional state across six emotions - anxiety, hostility, depression, confidence, compassion, and happiness.  Evolutionary computing techniques can help recognize patterns in those emotions over time.  This information may be able to help people quickly identify mental health concerns or even improve business communications and marketing.

Improved stem cell therapies:  An interesting area where machine learning techniques have been applied is in the field of stem cell research.  The goal is to extract the mechanics of stem cell differentiation, leading to more beneficial stem cell therapies.  Stem cells start as sort of a “blank slate,” but then specialize as they develop.  Specific ribonucleic acids called microRNAs help determine this differentiation.  Machine learning can identify which factors are truly influential in this developmental process, allowing researchers to quickly understand and regulate the system.

Predicting how viruses will change over time within specific individuals:  Evolutionary computing techniques can help clinicians and researchers understand how HIV evolves in patients infected with the virus.  HIV changes inside people who are infected and these changes are even specific to different types of tissues.  Funding from the National Institutes of Health and the National Science Foundation through companies like BioInfoExperts have shown that these techniques can help recognize and anticipate how the virus will change over time, which may lead to more effective treatments.

In many ways, improving outcomes in healthcare may not require new or revolutionary approaches.  It may just require a more informed look at the way Mother Nature has already been working for billions of years.

Follow me on Twitter or LinkedInCheck out my website