In-vivo detection of oscillatory magnetic field may lead to fundamental development of functional MR imaging, however is still hindered by the sensitivity achievable.¹ Although the recently developed spin-lock oscillatory excitation (SLOE)² method exhibited sub-nanotesla level of sensitivity, its application in vivo is troubled by the main field (B₀) inhomogeneity. In this work, an oscillatory-selective detection (OSD) is proposed to overcome this limitation and hence to improve sensitivity of in-vivo detection of oscillatory magnetic field. OSD was further demonstrated as a viable tool to map the transcranial alternating current stimulation (tACS) induced oscillatory magnetic field in human head.The pulse sequences of SLOE and OSD are shown in Figs. 1a and 1b respectively. In OSD, the spin-lock preparation as used in SLOE is replaced by a series of 180° inverting pulses (with alternating polarity and a phase offset of 90°). The number of 180° pulses needed, given the allowed total preparation time T_OS and oscillatory frequency of magnetic field to be detected f_OS, can be expressed as 2×T_OS×f_OS , where the factor 2 is attributed to the designed mirror symmetry of the first half of the 180° series. In this way, the B₀ and B₁ inhomogeneity can be effectively eliminated. In addition, a phase offset of 45° is used for the last 90° pulse as it maps the magnetization equally on the x and y axis to balance signal and SNR.Phantom and in-vivo experiments were performed using the proposed OSD and SLOE methods on a 3.0 T whole body scanner (GE discovery 750). The phantom was composed of a single copper loop wound around an American College of Radiology (ACR) MRI phantom as shown in Fig. 2a. A function generator was connected to the loop to generate sinusoidal current with an amplitude of 1 nT at the center of the loop according to the Biot-Savart law. An OFF-ON block stimulating pattern with a 60 TR interval was used in experiment. The lengths of preparation modules in SLOE and OSD were the same (100 ms) at different oscillatory magnetic fields (10 Hz, 50 Hz and 100 Hz), so were the imaging acquisition parameters kept identical for comparison. Similar setup and identical imaging strategy were also transferred to an in vivo study, where the copper loop was wrapped around the head. In addition, OSD was applied in mapping tACS area in human brain. Stimulation electrodes were positioned over Cz and Oz (international 10/20 EEG system). A sinusoidal current of 1 mA was applied at 10 Hz using a battery-driven stimulator (DC-Stimulator MR, NeuroConn). An OFF-ON block design with an 80 TR interval was used. In both set of experiments, two sample t-test was performed to determine the activation map after the data was high pass filtered at 0.01 Hz. For better visualization, tACS activation map was registered and overlaid to MNI standard brain using FSL.Detected activation maps in phantom and in vivo studies with varying oscillatory frequencies are shown in Fig. 2 and Fig. 3 respectively. In the phantom experiment, the derived B₀ map (Fig. 2a) shows increased level of field inhomogeneity in the central and right edge, which correspond to the regions where SLOE showed low level of detection. In the in vivo study, the image distortion and non-uniform background level in SLOE image (Fig. 3a) are indicative of the B₀ and B₁ inhomogeneity. As expected, these regions showed much reduced level of detection in SLOE as compared to OSD (Fig. 3b). The derived activation maps in consecutive axial slices of tACS area using OSD are shown in Fig. 4. It can be seen that significant activations were detected from occipital cortex underneath the stimulation electrodes, which agree with previously reports³.OSD utilizes a series of 180° inverting pulses instead of continuous spin-lock in capturing the low oscillatory magnetic field. As a result, OSD is much more immune to field inhomogeneity than SLOE and hence has better level of detection sensitivity, as demonstrated in both phantom and in vivo experiment. OSD has also been demonstrated as a viable tool in tACS experiment, which may confirm the previous speculation that fMRI signal artifact from tACS current exists⁴.