Detection of clinically silent circulating microemboli within cerebral arte
ries by transcranial Doppler ultrasound (US) is now being widely investigat
ed in the hope of identifying patients at increased risk for stroke. Howeve
r, the widespread application of embolus detection is still limited in clin
ical practice because curl ent transcranial Doppler systems have not the re
quired sensitivity and specificity to analyze microembolic signals, particu
larly to distinguish between gaseous, or solid brain emboli and artefacts.
In this work, me proposed to investigate the potential of a new approach fo
r the analysis of microembolic signals via the so-called matching pursuit,
which is closely related to wavelet transform and is not subject to the sam
e limitations as the fast Fourier transform, Our preliminary results clearl
y indicate that matching pursuit is well suited to this task, (C) 2000 Worl
d Federation for Ultrasound in Medicine & Biology.