Diagnosing abnormal heart rhythm with web camera
Atrial fibrillation (AF) is one of the commonest abnormal heart rhythms. In AF, the normal regular electrical impulses generated by heart's native pacemaker (SA node) are overwhelmed by disorganized electrical impulses from elsewhere. As a result of irregular conduction of impulses, the atria (upper heart chambers) and ventricles (lower heart chambers) beat out of sync. This abnormal rhythm is associated with a variety of medical conditions. The incidence and prevalence of atrial fibrillation increases with increasing age. .
Individuals with atrial fibrillation can be asymptomatic or may have symptoms like palpitations, chest pain, generalised weakness etc. They are at a higher risk of developing stroke as the irregular asynchronous rhythm promotes blood clot formation. These blood clots (thrombi) can detach and flow into blood vessels supplying the brain and thereby occluding it, resulting in stroke.
Atrial fibrillation can be suspected by palpation of the arterial pulse during physical examination of the patient. ECG gives definite diagnosis. But many individuals can have intermittent atrial fibrillations few or several times a day. In these individuals prolonged ECG recording is necessary to record the abnormal rhythm which may occur infrequently.
Researchers from University of Rochester and Xerox Corporation have proposed a new method of detecting atrial fibrillation using a web camera. As explained earlier, atrial fibrillation has irregular rhythm and one consequence of this is uneven blood flow in the blood vessels. This uneven blood flow causes subtle changes in the skin color, quite different from the normal individuals. Since the skin over the face is thinner, this area is considered as suitable for detecting these subtle changes. Changes in the skin color that are imperceptible to the naked eye are detected by the camera and a software algorithm devised by Xerox corporation analyses the patterns and infers the presence of abnormal heart rhythm.
The current error rate is relatively high at 20%. But with the improvement in the technology, this is expected to come down as per Jean-Philippe Couderc, the lead author of the study. The common use of web cameras by significant number of people implies that this technology can be potentially used in web based health care diagnostics.