Abstract – The effect of nonlinear scattering from mi- crobubbles is commonly used for contrast specific imaging. Linear scattering from microbubbles is not contrast specific, yet it contributes more to the echo signal. To optimize flow imaging with contrast agents, it is desirable to exploit linear scattering and nonlinear scattering to improve sensitivity without sacrificing specificity. This can be achieved by incor- porating the amplitude and phase dependencies of the fundamental and harmonic spectral components in the Doppler processing. The separability of the afore- mentioned components is achieved by phase- (or am- plitude-) coding of the transmit pulses. INTRODUCTION The measurement of blood flow velocities is a chal- lenging task, since strong echoes from out-of-plane structures and reverberation superimpose the echoes from actually flowing scatterers. Wall filters solve this problem satisfactorily for high flow velocities. With the introduction of nonlinear imaging tech- niques for contrast agents evaluating the 2 nd harmonic signal component seemed to be promising for imag- ing slow flow velocities [1, 2]. The fundamental sig- nal component, however, contains more energy. Thus, it is desirable to include both signal compo- nents in the flow estimation [3]. This can be done by an independent evaluation of both components. We have worked on a strategy that combines both com- ponents (and echoes form stationary scatterers) in one estimator, thus taking into account the common origin of both signals. Standard pulsed Doppler systems transmit a train of N pulses along the same beam line. Each firing starts at 0 = t , where t denotes the “fast time”. The time interval between subsequent pulse is given by prt prf 1/ T f = , where prf f is referred to as the pulse repetition frequency and defines the sampling rate along the “slow time”. The transmit pulses can be written as () () ( ) 0 cos , 1 , , , i i i i i s t gt t i N =α⋅ ⋅ ω +ϕ = α ϕ∈ … R (1) where () g t is the envelope, 0 ω the carrier fre- quency, and , i i a ϕ consider amplitude and phase cod- ing, respectively. In the following, we will focus on a sequence without amplitude-coding and with phase- coding, where 4 phases are repeated several times: [ ] [ ] 1 2 1 2 3 4 1 4 , , , , , , 4, , i N i N kk ϕ = ϕϕ ϕ = ϕ ϕ ϕ ϕ ϕ ϕ = ∈ α =α … … ` (2) In response to the transmit signals, echoes are re- ceived and quadrature demodulated into complex baseband signals () Bi e t . The mixing frequency and filters are chosen to include fundamental and 2 nd harmonic spectral components. The baseband signal is commonly evaluated at constant depth 0 2/ t zc = in slow time-direction denoted by i , where 0 c is the speed of sound and ( ) s prt 1 t i T = - . Assuming the most basic model for linear scatterers moving at the veloc- ity v with respect to the beam direction, the baseband signal is given by: () DS 0 0 0 j j , 2 / 0 D 0 e e , 2 , , 2 2 i t B B i iz t const z c e t e v v c ω ϕ = = = =β ω β∈ β∝α ω = π = λ ^ (3) We expand the model to account for stationary scatterers, described by the complex constant o and a nonlinear moving scatterer that is described by a fun- damental term (linear term) and a 2 nd harmonic term (quadratic term). The “DC” (broadband) component that is created by the quadratic term will be neglected. CONTRAST-ENHANCED FLOW IMAGING WITH PHASE-CODED PULSE SEQUENCES W. Wilkening 1 , B. Brendel 1 ,V. Uhlendorf 2 , H. Ermert 1 1 Institute of High Frequency Engineering, Faculty of Electrical Engineering, Ruhr-Universitaet Bochum, Germany 2 Research Laboratories of Schering AG, Berlin, Germany