Author: Thomas Madden

Status: Project Concept

Introduction: Holter monitors have been widely circulated throughout healthcare. Most monitors today are used to monitor patients with heart disease and ensure they maintain a steady, low heart rate. However, in most cases, the Holter monitor does not help to identify why the heart rate of a patient rapidly increases. A Holter monitor that compared heart rate increase with possible environmental causes for such an increase would increase the monitor’s effectiveness while maintaining its low cost and ease of use. By connecting the Holter monitor with voice recognition technology, a patient who has a heart condition and must maintain a steady, low heart rate can find out what keywords in conversation make him/her anxious. In this way, the doctor or therapist can help him/her pinpoint what makes him/her anxious and even avoid situations/conversational topics which repeatedly result in increased heart rates for an extended period of time.

Goal: To connect voice recognition software to the traditional Holter monitor so that patients can learn what topics make them anxious, helping them to live comfortable lives and maintain a steady heart rate.

Method: An attachment to the Holter monitor itself will record audio during periods of extreme change in heart rate, storing the data until the monitor can be analyzed in a separate client program. This program will implement voice recognition software from Watson Speech to Text API and analyze data from the Holter monitor with respect to the frequency of word use and the emotional weight of those frequently used words. Words most frequently used during periods of extreme heart rate change are more likely to concern topics that give the patient anxiety or unnecessary stress. However, frequency alone does not determine topical importance to the patient unless those words are given a separate level of importance based on their emotional weight. The client program must also filter out unnecessary words and highlight those words that often carry emotional weight. It is crucial to note that the Holter monitor will only record audio during those times of extreme change in heart rate, and that all other Holter monitor functions will remain the same.

Conclusion: Holter monitors today are tremendously accurate in analyzing heart rate, but will be put to better use when also used to record the words used in conversation and identify possible causes for change in heart rate. This data will be run through a program that analyzes which words are used frequently and matches those frequently said words with an ‘emotional dictionary’ which assigns emotional weight to each word, giving it a value based on that weight. These two variables, frequency of use and emotional weight, can be applied to each word used in conversation during periods of increased heart rate to discover what the cause of that increase is, and in so doing, help a therapist or doctor understand what steps a patient may need to take to release stress, maintain a low heart rate, and ultimately stay healthy.