Artificial intelligence can spot an irregular heartbeat from a scan which may appear normal to doctors, scientists claim.
The test was able to identify abnormalities in just 10 seconds, compared to current tests which can take weeks or years to interpret.
Doctors said the AI system, which uses deep learning, can find signals in heart tests that may be 'invisible to the human eye'.
Researchers at the The Mayo Clinic in Rochester, Minnesota, trained an AI model to detect the signature of atrial fibrillation rhythm, and then put it to test.
The test was able to accurately spot the unusual heart rhythm 83 per cent of the time, according to a study.
Atrial fibrillation, which raises the risk of stroke and heart failure, often goes undetected because patients' hearts go in and out of the abnormal rhythm.
AI can spot an irregular heartbeat from a ten second scan compared to current tests that can take years to diagnose atrial fibrillation (stock image)
The study, published in The Lancet, involved data from almost 181,000 patients who were already being investigated for having an abnormal heart rhythm.
Around 650,000 ECG scans were taken from the period between 1993 and 2017 in the study.
During an ECG, small stickers called electrodes are attached to the arms, legs and chest, and connected by wires to an ECG machine.
Every time the heart beats, it produces tiny electrical signals. An ECG machine traces these signals onto paper, which is read by a doctor.
The data was divided into patients who had tested either positive or negative for AF.
When the AI was tested on the first ECG from each patient, it accurately spotted the presence, history or impending AF 79 per cent of the time.
When using multiple ECGs for the same patient, the accuracy improved to 83 per cent.
Dr Paul Friedman, chair of the department of cardiovascular medicine at The Mayo Clinic, said: 'Applying an AI model to the ECG permits detection of atrial fibrillation even if not present at the time the ECG is recorded.
'It is like looking at the ocean now and being able to tell that there were big waves yesterday.'
Writing in the journal, they said: 'We used an AI model to find signals in the ECG that might be invisible to the human eye but contain important information about the presence of atrial fibrillation.'
The study is the first to use deep learning, which is a form of AI in which the system is able to improve each time it repeats the same task by learning from its past mistakes.
Since the AI is only as good as the data it is trained against, in this case using ECGs from people already being investigated, there could be mistakes in the interpretation when the test is used on the general population.
It also did not test on people with unexplained stroke.
'However, the ability to test quickly and inexpensively with a non-invasive and widely available test might one day help identify undiagnosed atrial fibrillation and guide important treatment, preventing stroke and other serious illness,' Dr Friedmen said.
After an unexplained stroke, it is important