Identification of myocardial infarction by high-frequency serial ECG measurement
Sandelin, Jonas (2022-04-05)
Identification of myocardial infarction by high-frequency serial ECG measurement
Sandelin, Jonas
(05.04.2022)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
avoin
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe2022041929460
https://urn.fi/URN:NBN:fi-fe2022041929460
Tiivistelmä
The purpose of this study is to attempt to identify acute myocardial infarction with high frequency serial electrocardiogram. High-frequency ECG and serial ECG are both unique ECG analysing techniques. The idea in this study is to combine these two and see if changes between different ECGs from the same person can provide us some information, whether it being in the high-frequency or normal frequency range of the ECG.
To answer the questions, an existing database which contained multiple ECGs for each person with high sampling frequency was used. 5 different machine learning models were trained and tested with this database. The results of the machine learning methods were good, producing the mean accuracy of 91.9%, while the best model was the Extra Trees machine learning model. It produced the accuracy of 97.9% when applying cross-validation to the database.
After these results, high-frequency serial ECG could be stated to be relevant. However, having ECG measured regularly can be expensive and time consuming. Therefore, the possibility of using a wearable ECG device was also studied. With a device called SAFE, developed by the University of Turku, a new high-frequency serial ECG database was gathered. The already existing machine learning model trained with the previous data was applied to this database and produced a mean accuracy of 90%. The quality of the ECGs gathered with the device were also deemed to be viable.
Both high-frequency ECG and serial ECG were found to be relevant methods. A wearable device could be used for AMI detection if the ECG is sufficient enough. Future studies could include increasing the dataset size of the wearable device, investigate other myocardial diseases and exploring the possibilities of high-frequency ECG further.
To answer the questions, an existing database which contained multiple ECGs for each person with high sampling frequency was used. 5 different machine learning models were trained and tested with this database. The results of the machine learning methods were good, producing the mean accuracy of 91.9%, while the best model was the Extra Trees machine learning model. It produced the accuracy of 97.9% when applying cross-validation to the database.
After these results, high-frequency serial ECG could be stated to be relevant. However, having ECG measured regularly can be expensive and time consuming. Therefore, the possibility of using a wearable ECG device was also studied. With a device called SAFE, developed by the University of Turku, a new high-frequency serial ECG database was gathered. The already existing machine learning model trained with the previous data was applied to this database and produced a mean accuracy of 90%. The quality of the ECGs gathered with the device were also deemed to be viable.
Both high-frequency ECG and serial ECG were found to be relevant methods. A wearable device could be used for AMI detection if the ECG is sufficient enough. Future studies could include increasing the dataset size of the wearable device, investigate other myocardial diseases and exploring the possibilities of high-frequency ECG further.