Online cardiac arrhythmia classification by means of circle maps analysis implemented on an intelligent miniaturized sensor

The intermittent occurrence of cardiac arrhythmias like e.g. atrial fibrillation hampers their diagnosis and hence the treatment. Since persons suffering from atrial fibrillation are known to have a remarkable increased risk of stroke the diagnosis of atrial fibrillation is a matter of great importance. Easy and comfortable to use long term ECG recording systems capable of online arrhythmia classification might help to solve this problem. We developed an intelligent, miniaturized, and wireless networking sensor which allows lossless local data recordings up to 4 GB. With its outer dimensions of 20mm per rim and less than 15g of weight including the Lithium-Ion battery our modular designed sensor node is thoroughly capable of up to eight channel recordings with 8 kHz sample rate each and provides sufficient computational power for online digital signal processing. For online arrhythmia classification we will record one ECG channel and 3-axis accelerometer data with 512 Hz each, the later being used for activity classification based artifact identification. We adapted our recently developed circle maps analysis of short term heart rate variation to run on this miniaturized intelligent sensor powered by the Texas Instruments MSP430 microcontroller derivate F1611. With this configuration we started to evaluate the cardiac arrhythmia classification in long term ECG recordings.