Development of precise off-diagonal magnetoimpedance gradiometer for magnetocardiography Open Access

Magnetocardiography (MCG) is a noninvasive, noncontact tool for the diagnosis of heart diseases. MCG measurements do not affect the electrophysiological properties of the heart. However, existing experimental and clinical data show that MCG is more sensitive (in comparison with methods based on the measurement of electric potentials) to local currents that are generated at the interfaces of myocardium fragments with different electrophysiological properties.¹ 

Superconducting interferential devices (SQUIDs) have previously been used to measure biomagnetic signals from humans. Recently, atomic vapor laser magnetometers, which have the ability to measure very clear MCG signals, have started to be used for mapping the MCG signal above the chest. The possibility of MCG measurement using a room-temperature flux gate sensor has also been reported.² A flux gate sensor with a probe a few centimeters in length can achieve a picotesla per square-root hertz noise level.

Other developments include a linear magnetic sensor based on the off-diagonal giant magnetoimpedance GMI effect.^3,4 Furthermore, arrays of flexible GMI sensors have been developed to measure magnetic brain signals.⁵ Compared with other sensors used for MCG measurements, the MI sensor has the advantage of being compact and easy to handle. Using a highly sensitive off-diagonal MI sensor, we have successfully measured biomagnetic fields in living cell-tissue preparations as small 5 mm² in size.⁶ In this study, we developed a precise first-order off-diagonal MI gradiometer for the purpose of future medical use as an MCG device.

The noise level of the MI gradiometer, which can operate in an unshielded environment and at room temperature, is 200 pT root-mean-square (rms) in a 100 Hz bandwidth. The achieved noise value corresponds approximately with the peak value of the MCG signal in the literature. Each sensor probe in the head pair of the gradiometer was 10 mm in length. The peak magnetic signal intensity of 100 pT (a value which corresponds to a conventional MCG) was observed when the front edge of the sensor head was set 10 mm apart from the chest surface. Under these conditions, the signal needed to be averaged over more than 50 cycles to identify the peak magnetic signal.

Figure 1 shows the MI characteristics of the amorphous wire element of length 1 cm and diameter 30 μm used in this study. When the high-frequency ac current passes through the amorphous wire MI sensor element, the skin effect is induced, and the impedance of the amorphous wire depends on the skin depth. In general, the skin depth is related to the circumferential permeability of the soft magnetic material, and therefore, the impedance depends on the magnetic field through the change in permeability.⁷ The change in impedance with applied magnetic field Z(H_ex)/Z(H_ex = 0) is more than 100% when the ac current is higher than 40 MHz. In this study, we used the MI element in combination with a pick-up coil for the magnetic-field detector, which utilizes the off-diagonal MI effect. The sensitivity of the off-diagonal MI effect, which can achieve a linear response for an external magnetic-field intensity that is less than the anisotropy field of the element, is higher than that of the original GMI effect under the condition of no magnetic-field bias.

It is necessary to develop a precise gradiometer to detect very weak magnetic fields such as biomagnetic fields in an unshielded environment. The basic circuit of the pulse-excitation-type complementary metal–oxide–semiconductor (CMOS) MI gradiometer is shown in Figure 2. The sharp pulse current having a short rising time t_r (10 ns or less) is supplied by the CMOS inverter to the amorphous wire of the MI element. The equivalent ac frequency for the skin effect is roughly estimated by f_eq = 1/2t_r = 50 MHz. The pick-up coil is used in combination with the amorphous wire 30 μm in diameter). The signal detection circuit consists of an analog switch, hold condenser, and timing circuit to control the analog switch. The output voltage of the MI sensor (E_out) is therefore proportional to the applied external magnetic field (H_ex). The linearity and hysteresis of the sensor is less than 2% over a ±10 μT full-scale range. To cancel out the effect of uniform magnetic-field noise, the amorphous wire has two coils: a sensing coil and a reference coil. The distance between the two pick-up coils is set to 3 cm. The difference in the sensitivities of the sensing and reference elements can be adjusted to within 1.5% by the use of a variable resistance.

Figure 3 shows the magnetic-field noise spectral density of the gradiometer in an unshielded environment compared with the environmental magnetic-field spectral density measured by commercial flux gate sensor. The environmental magnetic noise is 1 nT/Hz^1/2 at 1 Hz, while the gradiometer noise is 20 pT/Hz^1/2 at 1 Hz. The cancelling ratio of the environmental magnetic field is calculated to be 34 dB. The rms noise of 200 pT in a 100 Hz bandwidth (1–100 Hz) can be estimated by operating the gradiometer in an unshielded environment.

MCG measurements using the MI gradiometer were carried out on a male subject (aged 54) in the sitting position. We measured the MCG at a point on the chest surface 25 mm to the left of the pit of the stomach. Figure 4(a) illustrates the experimental setup. The MI gradiometer was placed on a wooden table, and we measured the magnetic-field component (B_z) perpendicular to the chest surface. The distance D is that from the chest surface to the front edge of the MI gradiometer. Figure 4(b) and (c) represent the simultaneous measurement of the electrocardiography (ECG) and MCG signals when D is less than 3 mm. In spite of superposing noise with an amplitude 300 pT, we can identify a sharp magnetic peak that corresponds to the R wave of the ECG. However, a negative magnetic peak is observed following the T wave of the ECG. The magnetic signal averaged over ten cycles is shown in Figure 5(b). The peak magnetic field is about 400 pT, which is much higher than the maximum MCG value (100 pT) reported so far. The signal seems to be affected by the movement of the chest surface caused by the heartbeat. To avoid this effect, the sensor head was placed 10 mm apart from the chest surface. The peak magnetic signal following the ECG T wave is not observed, but in this case, we were able to identify a peak magnetic signal intensity of 100 pT after averaging over 50 cycles as shown in Figure 5(c).

We have developed a precise first-order off-diagonal MI gradiometer for MCG measurements in an unshielded environment and at room temperature. We achieved an rms noise level of 200 pT in a 100 Hz bandwidth in an unshielded environment. We successfully measured the MCG signal perpendicular to the chest surface at a point 25 mm to the left of the pit of the subject’s stomach. The peak magnetic signal of 100 pT (a value which corresponds to a conventional MCG) was observed when the front edge of the sensor head was set 10 mm apart from the chest surface to eliminate the effect of the chest surface movement. Under these conditions, the signal needed to be averaged over more than 50 cycles to identify the peak magnetic signal.