Application of a 1H brain MRS benchmark dataset to deep learning for out-of-voxel artifacts

Neural networks are potentially valuable for many of the challenges associated with MRS data. The purpose of this manuscript is to describe the AGNOSTIC dataset, which contains 259,200 synthetic 1H MRS examples for training and testing neural networks. AGNOSTIC was created using 270 basis sets that were simulated across 18 field strengths and 15 echo times. The synthetic examples were produced to resemble in vivo brain data with combinations of metabolite, macromolecule, residual water signals, and noise. To demonstrate the utility, we apply AGNOSTIC to train two Convolutional Neural Networks (CNNs) to address out-of-voxel (OOV) echoes. A Detection Network was trained to identify the point-wise presence of OOV echoes, providing proof of concept for real-time detection. A Prediction Network was trained to reconstruct OOV echoes, allowing subtraction during post-processing. Complex OOV signals were mixed into 85% of synthetic examples to train two separate CNNs for the detection and predi..., AGNOSTIC was created using 270 basis sets that were simulated across 18 field strengths and 15 echo times. The synthetic examples were produced to resemble in vivo brain data with combinations of metabolite, macromolecule, and residual water signals, and noise. All of the parameters (i.e., amplitudes, relaxation decays, etc.) are included in each of the NumPy zipped archive file., NumPy archive files can be opened using Python and NumPy., # AGNOSTIC: Adaptable Generalized Neural-Network Open-source Spectroscopy Training dataset of Individual Components

#### Published in Imaging Neuroscience: [Application of a 1H brain MRS benchmark dataset to deep learning for out-of-voxel artifacts](https://doi.org/10.1162/imag_a_00025)

---

* **Aaron T. Gudmundson**, Johns Hopkins School of Medicine, Kennedy Krieger Institute, [ORCID: 0000-0001-5104-0959](https://orcid.org/0000-0001-5104-0959)
* **Christopher W. Davies-Jenkins**, Johns Hopkins School of Medicine, Kennedy Krieger Institute, [ORCID: 0000-0002-6015-762X](https://orcid.org/0000-0002-6015-762X)
* **İpek Özdemir**, Johns Hopkins School of Medicine, Kennedy Krieger Institute, [ORCID: 0000-0001-6807-9390](https://orcid.org/0000-0001-6807-9390)
* **Saipavitra Murali-Manohar**, Johns Hopkins School of Medicine, Kennedy Krieger Institute, [ORCID: 0000-0002-4978-0736](https://orcid.org/0000-0002-4978-0736)
* **Helge J. Zöllner**, Johns Hopkins School of Medicine, Kenne...