Project supported by the National Natural Science Foundation of China (Grant Nos. 11474166 and 11604156), the Natural Science Foundation of Jiangsu Province, China (Grant No. BK20161013), the Postdoctoral Science Foundation of China (Grant No. 2016M591874), the Postgraduate Research & Practice Innovation Program of Jiangsu Province, China (Grant No. KYCX17 1083), and the Priority Academic Program Development of Jiangsu Provincial Higher Education Institutions, China.
Project supported by the National Natural Science Foundation of China (Grant Nos. 11474166 and 11604156), the Natural Science Foundation of Jiangsu Province, China (Grant No. BK20161013), the Postdoctoral Science Foundation of China (Grant No. 2016M591874), the Postgraduate Research & Practice Innovation Program of Jiangsu Province, China (Grant No. KYCX17 1083), and the Priority Academic Program Development of Jiangsu Provincial Higher Education Institutions, China.
† Corresponding author. E-mail:
Project supported by the National Natural Science Foundation of China (Grant Nos. 11474166 and 11604156), the Natural Science Foundation of Jiangsu Province, China (Grant No. BK20161013), the Postdoctoral Science Foundation of China (Grant No. 2016M591874), the Postgraduate Research & Practice Innovation Program of Jiangsu Province, China (Grant No. KYCX17 1083), and the Priority Academic Program Development of Jiangsu Provincial Higher Education Institutions, China.
By combining magnetics, acoustics and electrics, the magneto-acoustic-electrical tomography (MAET) proves to possess the capability of differentiating electrical impedance variation and thus improving the spatial resolution. However, the signal-to-noise ratio (SNR) of the collected MAET signal is still unsatisfactory for biological tissues with low-level electrical conductivity. In this study, the formula of MAET measurement with sinusoid-Barker coded excitation is derived and simplified for a planar piston transducer. Numerical simulations are conducted for a four-layered gel phantom with the 13-bit sinusoid-Barker coded excitation, and the performances of wave packet recovery with side-lobe suppression are improved by using the mismatched compression filter, which is also demonstrated by experimentally measuring a three-layered gel phantom. It is demonstrated that comparing with the single-cycle sinusoidal excitation, the amplitude of the driving signal can be reduced greatly with an SNR enhancement of 10 dB using the 13-bit sinusoid-Barker coded excitation. The amplitude and polarity of the wave packet filtered from the collected MAET signal can be used to achieve the conductivity derivative at the tissue boundary. In this study, we apply the sinusoid-Barker coded modulation method and the mismatched suppression scheme to MAET measurement to ensure the safety for biological tissues with improved SNR and spatial resolution, and suggest the potential applications in biomedical imaging.
Significant differences in electrical impedance among different tissues or tissues in different pathological states were observed[1,2] in previous studies. It was reported that the conductivity of muscle was about ten times that of the fat tissue,[3] and the conductivity of breast cancer was much higher than that of the normal one.[4] Because of the wide variation range of electrical conductivity, images with improved contrast were achieved using electrical impedance tomography (EIT).[5,6] However, the spatial resolution and imaging depth of EIT are still limited by the shielding effect and the high-level current injection.[7] For the limited number of electrodes, the practical application of EIT is also hindered by the ill-posed inverse problem solution.[8] By combining the interaction among acoustic, magnetic and electrical fields, magneto-acousto-electric tomography (MAET),[9] also called Lorentz force impedance tomography[10,11] was proposed as an improved modality for mapping the distribution of electrical conductivity inside tissues. With the excitation of ultrasound and the detection of electrical signals in a static magnetic field, the merits of high spatial resolution and image contrast[12] of MAET have been demonstrated by theoretical and experimental studies in the past decades.
The origin of MAET is based on the principle of Hall effects imaging (HEI),[13,14] which was proposed by Han et al. With the excitation of a high-voltage pulse on an ultrasound transducer, images of the internal structure of plastic phantoms and tissue samples with apparent conductivity variations were restored. Then, the reciprocal method called magnetoacoustic tomography with magnetic induction (MAT-MI) was proposed by He et al.[15] and several algorithms were established for conductivity reconstruction both in the isotropic and anisotropic cases.[16–18] Based on the reciprocity theorem between electronics and acoustics, the MAET was also developed by Haider et al.[19] and the distributions of the induced current density inside the phantom were reconstructed. By applying the modified Wiener inverse filter and Hilbert transform to the collected MAET signal, the amplitude and polarity of conductivity variation were retrieved by Zhou et al.[20] and the one-dimensional distribution of conductivity along the acoustic transmission path was restored. In addition, some mathematical models and numerical frameworks were established for accurate conductivity reconstruction with the requirement for a higher signal-to-noise ratio (SNR) of collected signals.[21–23] The practical application of MAET was still hindered by the obvious drawbacks of low-level amplitude of the induced current and poor SNR of the collected MAET signal. In order to improve the performance of image reconstruction, a focused ultrasound beam was employed in practical measurements,[24] and the sectional impedance image for a beef sample was obtained with a high contrast. Although the difference frequency magneto-acousto-electrical tomography[25] was reconstructed by Renzhiglova et al. with a focused ultrasound excitation at the carrier frequency of 2.25 MHz and the modulated frequency of 2 kHz, the acoustic intensity in the focal zone was estimated at 37 W/cm2, which is still far above the safety limit.
In the past decades, pulse compression technology originating from radar applications has been introduced into ultrasound imaging[26,27] to enhance the imaging depth and SNR with a longer time duration and a lower amplitude of acoustic excitation. The matched and mismatched filters were used to compress the bursts into a short interval pulse and suppress side lobes of received signals. Due to the improvement of SNR in ultrasound imaging, the linearly frequency modulated pulse[28,29] with an amplitude of 60 V was introduced into MAET by Sun et al. The spatial peak temporal average intensity of 17.4 mW/cm2 was produced in the region of the sample and the discontinuous conductivity distribution with a spatial resolution of 1 mm was also obtained by the spectrum of the intermediate frequency signal. However, the polarity (direction) of conductivity variation was lost in Fourier transformation.[30] Except for the amplitude or frequency modulated technologies, several bi-phase coded excitation modalities[31] were also used to realize the effective pulse compression in ultrasonic imaging, such as the Barker code,[32] the Golay code,[33] etc. Compared with other schemes, the Barker code proved to have the lowest autocorrelation range of side-lobe level and the highest sensitivity for phase among the bi-phase codes with the same length.[34]
To enhance the energy of the incident acoustic signal in this study, the longest 13-bit sinusoid-Barker coded excitation is introduced into MAET and the mismatched compression filter is employed to suppress the noise-level of side lobes. The formula of the collected MAET signal with 13-bit sinusoid-Barker coded excitation is derived and simplified for the transducer with a strong directivity. Numerical simulations are conducted for a four-layered gel phantom with the 13-bit sinusoid-Barker coded excitation and the improved performance of wave packet recovery with the suppressed side-lobe level is obtained by using the mismatched compression filter, which is also demonstrated by the experimental measurements for a three-layered gel phantom. It is proved that with the 13-bit sinusoid-Barker coded excitation, the amplitude of the driving signal can be reduced to half that of the conventional single-cycle sinusoidal excitation with an SNR improvement of about 10 dB. The value and polarity of the conductivity derivative at the tissue boundary can be achieved by the amplitude and direction of the decompressed wave packet. In this study, we provide a comprehensive method of the 13-bit sinusoid-Barker coded modulation and the mismatched suppression scheme to MAET measurement for biological tissues with improved SNR and spatial resolution, and we also suggest the potential applications in biomedical imaging.
The schematic diagram of the MAET measurement system using sinusoid-Barker coded excitation is shown in Fig.
In an in-viscous homogeneous medium, the particle velocity v satisfies the motion equation ρ0 d v/dt = − ∇p, where p is the acoustic pressure and ρ0 is the density of the medium. For the planar piston transducer with the center frequency f and radius a located at z = 0, the velocity on transducer surface is u = ua exp (jωt), where ua and ω are the amplitude and the angular frequency of the acoustic wave. Thus, the acoustic pressure at (r, φ, z) in cylindrical coordinates can be described by
Then, the particle velocity of a charged particle can be calculated from
A matched filter is employed to restore the output from the collected MAET signal, which is defined as the autocorrelation function of the sinusoid-Barker coded signal and expressed as d(t) = T(t) ⊗ T(−t). As shown in Fig.
By taking the inverse Fourier transform of Eq. (
In order to verify the performance of MAET measurement using sinusoid-Barker coded excitation, numerical studies were performed for a four-layered gel model. As illustrated in Fig.
With the excitation of a 13-bit sinusoid-Barker coded signal, the MAET signal collected by the electrodes was simulated as plotted in Fig.
The sketch map of the experimental MAET measurement system using the 13-bit sinusoid-Barker coded excitation is illustrated in Fig.
A three-layered cubic gel phantom with a side length of 50 mm was prepared by sol-gel method in an acrylic mold,[20] and the electrical conductivity of each layer was controlled by adjusting the concentration of NaCl solution, which was measured by the impedance analyzer (Agilent 4294A, Agilent Technologies, USA). In order to analyze the accuracy of MAET measurement, a three-layered gel phantom was prepared layer by layer as shown in Fig.
The collected MAET signal with the 13-bit sinusoid-Barker coded excitation at a peak amplitude of 500 mV is present in Fig.
To compare the performance of the collected MAET signal with sinusoid-Barker coded excitation, an experimental measurement using single-cycle sinusoidal excitation was conducted for the same gel phantom The amplitude of the sinusoidal signal was set to be 1 Vpp, which was twice that of the 13-bit sinusoid-Barker coded signal as mentioned above. The waveform collected by the electrodes is plotted in Fig.
By adjusting the output amplitude of signals, the excitation amplitude dependence of SNR for the collected MAET signals with single-cycle sinusoidal and sinusoid-Barker coded excitations are measured as plotted in Fig.
Like the other ultrasound imaging methods, the SNR of the MAET signal is dependent on the acoustic energy incident to the medium,[38] and the spatial resolution is determined by the frequency of the sound wave and the radiation pattern of the transducer. With a given pressure amplitude, the energy emission is proportional to the time duration, which is beneficial to enhancing the penetration depth in the safety limit. In this study, the longest 13-bit sinusoid-Barker code excitation is employed to improve the SNR of the MAET signal with a relatively low acoustic pressure, while the spatial resolution of conductivity boundary differentiation can be reduced to a certain extent. The waveform of the simulated ultrasound is the convolution of the driving signal and the impulse response of the transducer with a limited bandwidth, which makes the wave packet wider and reduces the spatial resolution of the experimental system. The modified Weiner filter and Hilbert transformation are employed to reduce the wave packet influence of the impulse response[20] to achieve an improved performance. Although the transducer with a centre frequency of 500 kHz is used in this experiment, the 13-bit sinusoid-Barker code excitation can still be extended to a high-frequency system to improve the spatial resolution. In addition, by combining the linear frequency modulation with the Barker coded sequence, a higher SNR with good spatial resolution is obtained in B-mode sonography,[39] which might be used to improve the performance of MAET measurement for biological tissues in further studies.
Based on the principles of MAET and pulse compression technology, theoretical and experimental studies on MAET measurement using sinusoid-Barker coded excitation are conducted for a multi-layered gel phantom. The explicit formula of the MAET signal is derived and simplified for a strong directional transducer, and the corresponding decoded algorithm with side-lobe suppression is also employed to enhance the SNR and spatial resolution. Based on the simulation results of the MAET signal and the decoded waveforms, the performance of wave packet recovery with suppressed side-lobe level is improved by using the mismatched compression filter, which is also demonstrated by the experimental measurements for a three-layered gel phantom. With the sinusoid-Barker coded excitation, the amplitude of the driving signal can be reduced greatly with an SNR improvement of about 10 dB. The value and direction of conductivity variation (conductivity derivative) at tissue boundaries can be achieved by the corresponding amplitude and polarity of the decoded wave packet. The favorable results provide a coded excitation method and a compression algorithm to improve the SNR and spatial resolution for MAET measurement, and suggest the potential applications in biomedical imaging.
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] | |
[22] | |
[23] | |
[24] | |
[25] | |
[26] | |
[27] | |
[28] | |
[29] | |
[30] | |
[31] | |
[32] | |
[33] | |
[34] | |
[35] | |
[36] | |
[37] | |
[38] | |
[39] |