Magnetic Resonance Fingerprinting (MRF) [1] is a new technique that promises multi-parametric, quantitative MRI. A series of RF pulses generates a complex signal response in non-steady-state. For each voxel, the measured complex signal evolution is then matched against a series of simulated signal evolutions. This so called “dictionary” contains all possible signal evolutions that are simulated for a combination of T1 and T2. The dictionary entry being most similar to the simulated signal evolution determines the voxel’s T1 and T2. Spatial variation of the RF transmit field (B1+) is a commonly observed effect, in particular due to dielectric properties of the sample at higher field strengths. Such changes may lead to systematic errors in MRF results [2, 3], since the B1+ field impacts the applied flip angles and thus not only scales the acquired signal evolution but also changes its characteristic behavior (for an example see Figure 1). A straight-forward approach would incorporate the magnitude of the B1+ field as a separate dimension into the dictionary. But also a dictionary with a B1+ dimension is vulnerable for mismatches in case of the additional dimension. In order to improve the reliable assignment of quantitative parameters, we propose a novel encoding scheme for MRF which introduces an additional B1+ dependent component into the complex signal evolution.We use the MRF method from [5] that is based on a fast imaging with a steady-state precession (FISP) sequence. Instead of a two-dimensional spatial encoding using a spiral readout trajectory, a one-dimensional Cartesian encoding was applied which measures a projection of the object perpendicularly to the readout direction. This prototype projection sequence enables the investigation of the MRF signal in short scan times, without any undersampling artifacts and at high SNR. Usually, the MRF sequence employs conventional slice-selective RF excitation pulses with identical phase. We propose to replace each conventional RF pulse by a composite variant comprising two pulses with a relative phase shift of 90 degree. For simplicity, in the current implementation both pulses are played out successively, rewinding the slice selective gradient in between. This results in a phase evolution of the signal depending mainly on B1+, a property that has been used for the purpose of mapping the B1+ field before [4].The impact of an artificial B1+ variation on the phase evolutions was measured, simulated and compared. All scans were performed using a quantitative phantom (EuroSpin “Test Object T05", Sonar Diagnostics Ltd, Livingston/UK) on a clinical 3T MR scanner (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany). For the measurements, selected tubes of the phantom with defined T1 and T2 values were located side by side such that their signal contributions could be clearly separated despite the one-dimensional spatial encoding.Simulated and measured signal evolutions show the same B1+ dependent phase evolution behavior. Figure 2 exemplarily displays three measured phase evolutions for one tube with specific T1 and T2 values at three different B1+ field settings. Figure 3 displays the corresponding simulated phase evolutions. Comparing these results to the conventional MRF sequence reveals the advantage of the new technique. Figure 4 exemplarily shows two measured phase evolutions for one tube with different B1+ field settings. Figure 5 shows the corresponding simulated phase evolutions. Simulated and measured curves do not contain phase information depending on B1+ that can be used for the pattern matching.We demonstrate a novel method to better discriminate entries of an MRF dictionary including the B1+ field as a separate dimension. Results show that this can be achieved using dedicated composite RF pulses. Future work will investigate more sophisticated, frequency-modulated, single-pulse designs. The new degree of freedom can be used in different ways, for example, only for a fraction of the RF pulse series or with varying relative phase shift.