Chiasmal malformations dataset: a unique neuroimaging testbed

Abstract

The human optic chiasm is formed normally by almost equal populations of crossing and non-crossing optic nerve fibers. This proportion can be affected by rare disorders, such as albinism or achiasma (Hoffmann and Dumoulin, 2015), causing, respectively, over- or under-representation of crossing fibers. In the light of recent studies revealing those differences from anatomical measures (Puzniak et al., 2019), the optic chiasm appears as a valuable model for researchers interested in the impacts of deficits of vision on brain white matter or tractography methods developments. In order to provide the research community with an access to this unique model, we make the MRI dataset, thoroughly describing chiasmal malformations, available. We recruited patients with albinism (n=9), achiasma (n=2), and matching controls (n=8). For the patients we collected ophthalmological records including visual acuity, fixation stability estimates and results of visually evoked potentials measurements. Further, we collected high quality structural and diffusion brain MRI data of all participants. In addition, in a subset of participants with albinism (n=6) we acquired retinotopic maps using functional MRI (Ahmadi et al., 2019). The MRI data was clinically examined, technically validated, preprocessed using state-of-the-art pipeline and made publicly available on the brainlife.io platform (Avesani et al., 2019). This includes structural, diffusion and functional MRI data (both raw and preprocessed versions) and their derivatives (such as manually curated white matter masks). Through sharing a MRI dataset on chiasmal malformations, the wide research community can incorporate this unique condition in their research approach.