The timely diagnosis of stroke at the initial examination is extremely important given the disease morbidity and narrow time window for intervention.
A recent population-based study across 9 states indicated that the incident rate of missed stroke diagnoses in the Emergency Department (ED) is ?13%.1At the same time, several studies have demonstrated that the prevalence of stroke mimics among patients presenting to the ED with acute stroke symptoms can be as high as 30%.2 There are currently several validated stroke and stroke mimic recognition tools for prehospital and hospital settings. Face Arm Speech Time, Los Angeles Prehospital Stroke Screen,3 Recognition of Stroke in the Emergency Room,4 and FABS scoring system (facial droop, atrial fibrillation, age, blood pressure, history ofseizures, and sensory)5 are some of the examples. However, none of these tools can rapidly and concurrently consider various stroke-related parameters (eg, medical history, family history, or laboratory studies) in a user-independent fashion.
Supervised machine-learning methods from artificial intelligence can learn complex structures using a training data set and use that knowledge to predict the outcome of an unobserved situation. Artificial neural network (ANN) is one example of supervised machine-learning methods inspired by biologicalneural systems. That includes a computational approach that is based on a large collection of neural units modeling the way the brain solves problems. Each neural unit is connected with many others, and links can be enforcing or inhibitory in their effect on the activation state of connected neural units.
Neural networks typically consist of multiple designs, and the signal path traverses from front to back. ANN-based models are effective at capturing nonlinear relationships that make them ideal candidates for complex and multifactorial disease classification.