Micro-embolic signatures detection through GARCH parameters

International audienceDetection of micro-emboli is of great clinical importance to prevent cerebro-vascular events. Standard detection techniques implemented in the most commonly used systems are generally based on the comparison between a decision information (energy) and a threshold. The value of this threshold can be set just above the statistically highest detected energy of a blood Doppler signal. This choice of threshold consequently prevents all detection of micro-emboli events whose energy might be lower than the systolic energy. In this study three detectors were tested and compared: i) a standard energy detector, ii) a detector based on AR parameters and iii) a detector based on GARCH parameters. From in vivo signals and for a detection rate set to 100%, the false alarm rates were of 33%, 5% and 3% for the standard energy detector, for the detector based on a AR(2) model and for the proposed detector based on the GARCH(1,1) model, respectively. This study demonstrates that GARCH technique detects micro-emboli which did not identified by classical methods. Large micro-emboli are all detected, but small micro-emboli are only detected with parametric techniques. The new detector opens up new prospects to detect small emboli, despite the need for further studies to incorporate technique "on line"