Bipolar disorder affects at least six million Americans each year, or roughly 4 percent of U.S. adults. Treatment typically involves some combination of mood-stabilizing or anti-psychotic medications and psychotherapy. However, recent discoveries in the field of artificial intelligence may prove revolutionary for individuals suffering from this condition.
Artificial intelligence, commonly known as “AI,” refers to intelligence exhibited by computerized machines. The idea revolves around a perceived need to create machines that can perform tasks previously limited to humans, such as those that involve decision-making. Essentially, the more technologically advanced we become, the more important it is to ensure that we can minimize human error.
One area where AI has shown great promise is in military applications. In air-to-air combat simulations, it has been known to outmaneuver even highly-skilled Air Force pilots. This same technology has now demonstrated a similar ability to predict how patients with bipolar disorder might respond to lithium treatment. Accurate predictions are difficult because of frequent fluctuations between periods of mania and depression. These fluctuations necessitate changes in treatment approaches, as they occur.
A new study conducted by the University of Cincinnati College of Medicine has shown that using AI can dramatically improve treatment effectiveness and efficiency. Various models have been employed by UC to predict how a particular patient might respond to lithium. The best of these models was known to be accurate 75 percent of the time. By comparison, the model employing AI was found to be accurate 100 percent of the time. Furthermore, it was 92 percent accurate in predicting when a patient’s manic symptoms were likely to decrease.
Artificial intelligence programs arrive at their conclusions based on generalizations, rather than specific definitions. They continuously refine their answers in a manner analogous to Darwin’s natural selection. Essentially, AI provides a way for doctors to use “fuzzy logic” to help them battle a notoriously difficult medical problem.
While it might seem as though air combat and medicine have little in common, they both involve an orderly process to arrive at the best possible decisions. While these new algorithms are clearly not sentient beings, as might be imagined in the world of science fiction, they are valuable tools that can be adapted to suit a large number of different applications.