Retino-cortical computational model to study suppressives waves in fast motions
Résumé
Oculars saccades are very fast eyes movements, occurring about 3 times per second and with a speed of 300 to 1500 mm/s. This process is essential to explore a visual scene and capture its information. It also carries our facial recognition and reading abilities. During saccades, the representation of the visual scene moves at the same speed than the saccade. However, we don't perceive this motion. On the contrary, the visual scene stays stable, clear and precise. This masking effect is called "saccadic suppression". Current explanation of this phenomenon is based on a signal called "corollary discharge" (Robert H. Wurtz, 2018). On one hand, this copy of the ocular motor command is transmitted to the frontal eye field where it allows a remapping of the visual field. On the other hand, the suppressive effect of this corollary discharge is sent to the middle temporal resulting in a cancellation and in the absence of perception of the saccade. However, an alternative explanation is possible. Recents psychophysical experiments of Mark Wexler lab (M.Wexler & P. Cavanagh, 2019) have managed to reproduce saccadic suppression without any eyes movements in human subject. In this experiments, ocular saccades movements was reproduced by using a stimulus moving at the speed of the saccade and displayed with a device with a frame rates of 1440Hz. Traditionnaly, stimuli in psychophysics experiments are displayed with a 60Hz videoprojectors. With such a refresh rate the subjects observe an apparent motion. This illusory motion is perceived more blurry and with a smear around the trajectory of the stimulus. In contrast, using a video-projector with a refresh rate of 1440Hz allows to remove this apparent motion perception. In the experiment, two types of stimuli have been used : a simple white bar in motion or the same bar in motion preceded and followed by static phases. Interestingly, only motion with static phases led to saccadic suppression with a slower and clear motion. Indeed, stimulus with motion only gave a faster and blurred motion with a smear behide the bar. It results from these experiments that an alternative hypothesis to corollary discharge could explains saccadic suppression by activity waves propagating into the cortex and the retina. This hypothesis is supported by experiments in Monkey primary visual cortex made by Frederic Chavane (Institut des Neurosciences de la Timone) exhibiting suppressive waves in the opposite direction of a motion. Indeed, these activity waves propagate to previous activated area by the motion and suppress their residual activity. This suppressive effect from last position of the stimulus on the older one could explains saccadic suppression and his ability to return a clear stimulus during motion.
This work holds in the context of an ANR collaboration between Mark Wexler, Frédéric Chavane, Olivier Marre, Alain Destexhe and Bruno Cessac aiming at testing this hypothesis. Our goal is to reproduce these activity waves in a retino-cortical model and simulate saccades in it, so as to observe the experimental evidences made by M. Wexler and F. Chavane. Our model emulates a retinal entry (developped in collaboration with O. Marre) feeding a V1 model (developped in collaboration with A. Destexhe, M. DiVolo and F. Chavane). Using this model we want first to explore the effects of high and low frame rates video on the retino-cortical model response. Our retino-cortical model produces an activity corresponding to pixel intensity in voltage sensitive dye imaging (VSDI). It also reproduces activity waves in cortex in agreement to experiments. Both models are merged into a retino-cortical model implemented in a software made at INRIA and called Macular. This modular platform allows to create network of object as biological cells or cortical columns, set connections between them and put a spatio-temporal input coming from a video. In this setting we want to show how an apparent motion (low refresh rate) influences the response activity profile in contrast to a fluid motion (high refresh rate). Thus, we could explain why the use of 1440Hz instead of 60Hz is mandatory in motion experiments. Secondly, still with 60 versus 1440Hz stimuli to mime saccades, we will the use of Wexler's stimuli with simple motion or motion surrounding of static phases. We want to understand which impact have the static phase on cortical activity and how the static phase potentially triggers suppressive waves. Finally, we also want to understand how cortical and retinal waves can interact together in this simulated saccades so as to better understand the respective role of the retina and of V1 in this context. In this poster we will present ongoing results on this topic.
Origine | Fichiers produits par l'(les) auteur(s) |
---|