version française


Inference in Visual Behaviours

Vision is a major sensory input for guiding our actions, perceiving our environment and con-ducting cognitive tasks. Yet the visual inputs that reach the brain represent a computational challenge:
they are ambiguous, dynamical and segmented into a myriad of piecewise cues. Imagine, for example, that you have to visually track a moving object in rainy conditions and that this target disappears behind a large object for a while. Such an apparently easy task requires segregating the various motion signals in order to integrate the trajectory of the target and continuously track it with the gaze.

It is obvious that, to overcome this problem, our visual system must combine sensory inputs with a priori knowledge at multiple spatial and temporal scales. Our research projects aim at elucidating the neural computations that meet these challenges from two different theoretical standpoints that are detailed below.
To do so, the InVIBE team develops multiple expertises involving behavioral studies in both humans and monkeys, electrophysiological and real-time optical imaging studies in behaving monkeys and Bayesian modeling approaches.

Our first aim is to understand how information is integrated, neurally represented and mod-ulated for guiding the visual exploration and the buildup of percepts.
Our originality is to ex-plore the underlying dynamical computations that run in parallel at different spatio-temporal scales, within different brain regions (i.e. V1, V4, MT, Superior Colliculus, Cerebellum) and different levels (from neurons to behavior).
To link our studies across these dimensions, we have developed standardized paradigms (such as motion integration) using both classical and newly-designed, high-dimensional stimuli.
In parallel, we investigate how these context-dependent dynamical representations are guiding, or conversely influenced by, goal-directed oculomotor behaviors.

The brain uses inference to construct an optimal interpretation of noisy and ambiguous inputs. Our second goal is to understand how the visual inputs interact with dynamical and hie-rarchical inferences embodied in the early cortical stages to control visual motion integration.
First, we record how solutions of inference computation are diffused within a cortical area to reconstruct global information, but also between areas, to use internal knowledge from higher stage.
Second, instead of using an ad hoc definition of prior knowledge, we have begun to directly probe its dynamics using eye movements.
A long term perspective here is to elucidate how prior and sensory information are combined by comparing on-going and evoked activity at different scales, in the context of well-defined tasks such as integration within and across apertures.

Over the last 6 years, we also started an applied research program to investigate the cortical dynamics when a massive re-mapping of the inputs is caused by retinal lesions.
On the one hand, we want to image changes in the dynamics of cortical activity following artificial scoto-ma.
On the other hand, we investigate new therapeutic perspectives by gauging the functional impact of retinal neuroprostheses (in collaboration with CEA-Leti, Institut de la Vision).
Our objective is to understand and refine patterns of retinal stimulations to generate natural-like cortical dynamics as well as behaviours. These two complementary approaches are con-ducted in collaboration with the Department of Ophthalmology in Marseille.

Finally, our team also develops methodological research projects, to better understand (i) the neuronal underpinnings of the voltage sensitive dye fluorescent signal, (ii) the neurovascular coupling in the healthy and epileptic brain (in collaboration with INSERM U751) and (iii) optical imaging in biological, highly diffusive, media (in collaboration with the Fresnel Institute in Marseille).

The research project of our team aims at merging behavioral, physiological and theoretical approaches.
The detailed projects presented above are rooted on these efforts and upon our determination, consolidated over the years, to combine our various expertises on a common paradigm.
They are powered by the technical expertise of the team members, ranging from behavioral studies in both humans and monkeys to electrophysiological and real-time optical imaging studies in behaving monkeys, to modeling approaches

Team manager

CHAVANE Frederic

Chef d'équipe-Team Leader

MASSON Guillaume

Chef d'équipe-Team leader

Team member

Selected publications

  • Taouali W., Benvenuti G., Wallisch P., Chavane F., and Perrinet L.U. (2016). Testing the odds of inherent vs. observed overdispersion in neural spike counts. Journal of Neurophysiology, 115: 434-444.

  • Matonti F., Roux S., Denis D., Picaud S., and Chavane F. (2015). Cécité et réhabilitation visuelle. Journal Français d'Ophtalmologie, 38: 93-102.

  • Quinet J. and Goffart L. (2015). Does the Brain Extrapolate the Position of a Transient Moving Target? Journal of Neuroscience, 35: 11780-11790.

  • Quinet J. and Goffart L. (2015). Cerebellar control of saccade dynamics: contribution of the fastigial oculomotor region. Journal of Neurophysiology, 113: 3323-3336.

  • Meso A.I. and Masson G.S. (2014). Dynamic resolution of ambiguity during tri-stable motion perception. Vision Research, 107C: 113-123.

  • Muller L., Reynaud A., Chavane F., and Destexhe A. (2014). The stimulus-evoked population response in visual cortex of awake monkey is a propagating wave. Nature Communications, 5.

  • Vidal M. and Barrès V. (2014). Hearing (rivaling) lips and seeing voices: how audiovisual interactions modulate perceptual stabilization in binocular rivalry. Frontiers in Human Neuroscience, 8.

  • Da Silva A., Deumié C., and Vanzetta I. (2012). Elliptically polarized light for depth resolved optical imaging. Biomedical optics express, 3: 2907-2915.

  • Leon P.S., Vanzetta I., Masson G.S., and Perrinet L.U. (2012). Motion clouds: model-based stimulus synthesis of natural-like random textures for the study of motion perception. Journal of Neurophysiology, 107: 3217-3226.

  • Simoncini C., Perrinet L.U., Montagnini A., Mamassian P., and Masson G.S. (2012). More is not always better: adaptive gain control explains dissociation between perception and action. Nature Neuroscience, 15: 1596-1603.
  • Publications

    CNRS logo université Aix Marseille logo | Sitemap | Terms & Conditions | contact | admin | intranet | intcloud |