ANR ReViS-MD

Macular degeneration is the main cause of visual impairment in Western countries. It is manifested by the gradual appearance of a scotoma in the macula which causes central vision loss and considerably handicaps patients in their everyday life.

ReVis-MD is an interdisciplinary and multi-centric project (Toulouse / Grenoble) which aims at better understanding functional reorganization in patients following the onset of the scotoma. By combining ophthalmological, psychophysical, neuronal (fMRI) and neuro-computational (artificial neural networks) measurements, the project pursues three main objectives.

The first objective is to characterize the cortical reorganizations which spontaneously occur in patients and how they modify their visuo-cognitive skills, in comparison with a control group of age-matched participants whose central vision is masked by an artificial scotoma in order to reproduce the same visual stimulation conditions as in patients. Our hypothesis is that central vision loss in patients could be partially compensated by an improvement of their visual-cognitive skills in peripheral vision which remains preserved. This hypothesis will be tested in patients with age-related macular degeneration (AMD) as well as in patients with the juvenile form of the disease (Stargardt syndrome). In particular, we will test their ability to detect movements and recognize visual scenes.

The second objective of the project is to test whether these cortical reorganizations can be reinforced by perceptual learning approaches based on intensive training of the patient visuo-cognitive skills in peripheral vision. Patients (AMD and Stargardt Syndrome), as well as age-matched controls whose central vision will be masked by an artificial scotoma, will have to carry out psychophysical tasks based on motion disctrimination and visual scene categorization over several weeks. We will characterize the evolution of performances for each of the group, as well as the potential changes in their fMRI activations after training. This will allow us to quantify learning effects in each patient with respect to his/her control. If successful, the project could pave the way for future rehabilitation strategies for patients based on perceptual learning.

The third objective of the project is to model the spontaneous and training-induced reorganizations from approaches in computational neurosciences. For this, we will use neural networks which will be trained with images or videos masked by an artificial scotoma. The properties of the network after learning will permit to better understand the mechanisms involved in reorganizations (for example, do they occur following changes in cortico-cortical connections?). In order to validate the model, the network responses will be compared with those measured in patients (fMRI and psychophysics). The network will also be used to predict the effects of other types of learning (e.g., using stimuli such as faces).

The project has multiple implications at the scientific, clinical and societal level. In particular, it could improve rehabilitation strategies for patients suffering from macular degeneration, but also from other visual diseases (glaucoma, retinitis pigmentosa) or more generally from sensory deficits following a pathology or stroke. The data collected during the project will be made available to the scientific community on servers and the results promoted both academically and to the general public.

See the publications in the portail HAL-ANR Portal