Although spiral phase contrast (PC) imaging has several advantages over Cartesian PC, namely its optimum k-space coverage, shorter echo times and robustness against motion artefact, its wide-spread use is often limited by its sensitivity to system imperfections including eddy currents (1). The objective of this work was to analyze and correct for eddy-current related artefacts in spiral PC using a Point RESolved Spectroscopy (PRESS) (2) acquisition technique.Data of a static phantom was acquired from an axial slice using a spiral gradient-echo PC sequence on a 3T MRI system (Philips Healthcare, Best, The Netherlands) with following parameters: FOV: 288x288 mm², slice thickness: 10mm, in-plane resolution: 1x1mm², TE/TR = 2.8/2500 ms, interleaves:10, read-out time: 5 ms, in-plane (AP) velocity encoding, venc=50cm/s. All system-related pre-emphasis corrections were kept unchanged, i.e. turned on. To measure the eddy current related phase evolution following the bipolar velocity encoding gradient used in the PC sequence a pre-scan based on five voxels (10x10x10 mm3) at five locations (isocenter, ±60mm in AP and RL) using a PRESS sequence with and without the bipolar velocity encoding gradient during the FID acquisition was acquired. Second order spherical harmonic basis functions were fitted to the spatially distributed phase evolutions. The effects were also modeled using Bloch simulations. Acquired PC data were corrected by subtracting the 0^th order phase and subsequently correcting the k-space grid. Results were compared to a widely used image-based correction method (3) consisting of a higher order weighted fit of the static background offset .Figure1 shows the PC sequence diagram along the flow encoding direction with the measured phase deviations (b-d) during the read-out. Figure2 shows up-scaled phase difference images in a measured and simulated static water phantom. The magnitude difference (d) shows a geometrical shift between simulated data without and with 0^th & 1^st order eddy currents. Measured uncorrected PC results along with images corrected using the presented method to the 0^th and 1^st spatial order in a structural water phantom are shown in Figure3. The magnitude difference images between the flow and non-flow encoded images are shown as well as up-scaled PC images. A geometrical misalignment similar to the simulation results is observed in (a). The PC root-mean-squared-deviation (RMSD) calculated in measured and simulated data following the proposed correction up to the 1^st order or an image-based correction up to the 2^nd order are shown in Figure4. A decrease from 9.00 % to 0.56 % or to 0.60% of the applied venc is achieved for the proposed or the image-based correction method respectively.The observed exponentially decaying phase offset (approx. time constant: 0.17 ms) and an oscillatory 1^st order term (approx. frequency: 0.6kHz, ascribed to mechanical system vibrations) are in line with previous findings using a magnetic field monitoring device (4). The correction results as well as the simulations consistently show that the main offsets in the reconstructed image can be attributed to the temporally decaying spatial 0^th order term. The presented method was initially applied in static phantoms and eddy-current steady-state effects were avoided using long TRs. The method was performed as a pre-scan and necessitates, similarly to an image-based correction, static tissue in the image. In contrast to an image-based correction however, the SNR or the presence of artefacts in the reconstructed image is irrelevant. Phase offsets through 2^nd order spherical harmonics showed no significant effect in the reconstructed image.A simple, hardware-independent, eddy-current mapping technique is presented and applied to spiral PC. It is shown that spatially constant (0^th order) eddy-currents can lead to spatially quadratic phase offsets and to geometrical distortions. In contrast to the presented method, a quadratic phase fit is therefore necessary to achieve similar correction results. The presented method successfully corrects distortions and velocity offsets are reduced to a residual of under 1% of the used venc.