Predicting risk of suicide using resting state heart rate

Abstract

This study investigates the potential of using heart rate-related measurements to aid clinicians in predicting suicide risk. For this purpose, heart rate was recorded during 10 minutes resting state from 15 patients with suicide ideation and 15 healthy subjects using an affordable and wearable sensor. Our results showed statistically significant differences (p<;0.05) in two time-domain features measuring overall heart rate variability and short term heart rate variations. KNN and SVM classifiers were implemented on the features obtained. Our results showed that using heart rate-related features the risk of suicide could be predicted by an average accuracy of 80%.