Skip to main content

From oculomotor action to perception

Periodic Reporting for period 1 - ProactionPerception (From oculomotor action to perception)

Reporting period: 2015-04-01 to 2017-03-31

What is the problem/issue being addressed?
The proposed scientific project addressed a gap in our current knowledge in neuroscience linking sensory physiology and motor behaviour to active perception: animals, including humans, disambiguate between motion in the environment and self-movement using corollary discharge, proprioceptive signals and reafferent sensory input. These signals can also be used to influence information acquisition to optimize sensory processing. We explored the links between oculomotor action and perception to understand how eye movements reshape processing in the early visual system.

Why is it important for society?
The present project seeks to establish links between sensory neural circuits and perceptual functions, on the one hand, and infer possible central neural mechanisms mediating the impact of goal directed eye movements on low-level visual perception, on the other hand. Perception linked to action in this manner constitutes ‘active vision’ which, because humans are largely ‘visuals animals’, is of primary importance to our behaviour in our everyday lives. A better knowledge of this phenomenon is relevant to improve our understanding of the visual system and to the study of all sensory systems in general. It will also be relevant for a deeper understanding of information processing algorithms, including the value of internal predictive paradigms, used throughout the brain.

What are the overall objectives?
In the present project, we explored the links between oculomotor action and perception, focusing on the underlying neurophysiological mechanisms because the gap in our current knowledge in this regard was preventing the advancement of the field.
Our specific aims are:
AIM #1: Study the impact of reafference (retinal signals) in the response of early visual neurons to visual input.
AIM #2: Understand the role horizontal connections in early visual cortex play in processing global motion and flow.
AIM #3: Confirm our hypotheses about the impact of oculomotor signals in the early visual cortex and the role of horizontal connections by studying the perceptual consequences associated with the underlying neurophysiological mechanisms.
Our overall long-term goal is to understand how motor control may shape sensory input to the brain by acting as a dynamic perceptual filter.
Work performed
During the initial 6 months of the grant, based on preliminary data collection, we refined the visual stimulation paradigms and collected a first batch of experimental data. In-depth analyses of these data at the single neuron level allowed us to design further controls (to test the feature-specific role of the apparent motion induced flow) that were implemented during the second part of the data collection (months 6 through 18). During the last 6 months we further analysed the data, including population analyses, and started writing-up the manuscript for publication.

Summary of main results
Our results show that centripetal iso-oriented flow at saccadic speeds enhances the response of early visual neurons, while centrifugal, cross-oriented, or random flows do not. This indicates that the underlying network may only be recruited when the visual input carries a sufficient level of spatial and temporal coherence, as we hypothesized. Saccadic speeds, matching the conduction speeds of horizontal connectivity are also needed to obtain the boosting effect, which becomes weaker and eventually vanishes at lower speeds. We also found that, in a subset of neurons, sequences restricted to the silent surround (and excluding the centre of the receptive field) elicit a significant response; these cells can be interpreted as ""able to generate filling-in responses"".

Exploitation and dissemination of results
These results were presented at the American Society for Neuroscience Meeting (SfN 2015), the European Conference on Visual Perception (ECVP 2016), the Göttingen Meeting of the German Neuroscience Society (2017) and the NeuroFrance Meeting of the French Society for Neuroscience (2017). We are currently writing up the final manuscript to be submitted to a high impact factor journal.

We strongly believe that as scientists our job is not only to make discoveries but also to promote communication and transfer of knowledge to society. The fellow supported by this Marie Skłodowska-Curie Action has been involved in numerous outreach and science communication activities, often in collaboration with the DG Education and Culture:
- Conferences & demos for the general public
- Visits to high school classes
- EU scientific strategy and policy meetings (where scientists interact with political leaders, policy makers, stake-holders and industry representatives).
She was one of the 12 Marie Curie Alumni invited to give a TED-like talk at the “20 Years of Marie Skłodowska-Curie Actions” celebration in Brussels in November 2016.
Our project has improved our understanding of the relationship between oculomotor action and perception, as well as the neural mechanisms in charge of perceptual grouping and motion perception and the role played by horizontal connections in primary visual cortex. This has wide implications in the fields of visual and systems neuroscience, eye movement control, and psychology. A better understanding of this particular neural system has a good chance of revealing basic concepts of brain function that will prove generalizable in a much greater general context.
In addition, our results have the potential of impacting **medically oriented areas** of human health and visual prosthetics. A thorough understanding of the cortical circuits in the first stages of the visual system may make it possible in the future to repair visual areas damaged by trauma, vascular accidents or disease related pathologies. Another potential application will be in contribution to the design and development of retinal and cortical prostheses that one day may enable blind people to see.
In the fields of **artificial vision and robotics**, a better understanding of perceptual grouping has applications leading to more effective algorithms for object recognition and many other perceptual capabilities. Understanding how the brain uses internal predictions and motor action for perceptual filtering may allow for a significant speed up of current artificial perception systems, very relevant for example in development of active perception for autonomous navigation systems.
To fulfill the translational potential in the fields of medicine, computation and robotics, the host laboratory (CNRS-UNIC) contributes to the already established partnerships in European Commission sponsored projects such as ‘BrainScaleS’ (FP7 ICT-FET Integrated project) and the Human Brain Project (FP7 and Horizon 2020 Flagship project) which have a strong focus on technological advance (such as neuromorphic computing) and on translational impact to the medical field.