DS0407 -

Neural bases of Causal Inference in Multisensory perception – NeuroCIM

Submission summary

Our perception of our environment is fundamentally multisensory, based on the combination of extrapersonal (visual and auditory) and self-related sensory information (tactile and vestibular). Psychophysical data show that multisensory information improves perception by reducing ambiguities. This translates into increased capability for detection and discrimination of multisensory stimuli. Remarkably, overall, a large portion of the cortex is involved in multisensory processing. However, the neural bases of a functional specialization (w/ location, identity whether physical, emotional or social, valence, salience etc.) within this cortico-cortical multisensory network is actually little explored to date. The exact functional contribution of multisensory modulatory influences in early cortical sensory cortices is also unclear. Additionally, the contribution of thalamic inputs to these cortical signals is also unexplored. Last, not all sensory inputs require to be combined, and to adequately reflect the world, the brain needs to integrate multisensory information that is associated to a same source while segregating multisensory information from different sources. Behaviorally, this is well accounted for by a Bayesian Causal Inference framework. The precise neural bases of this causal inference are yet unknown. Here, we combine fMRI and dense neuronal recordings, in monkeys and in humans, to fill three main knowledge gaps: (1) provide a long missing multisensory mapping of the thalamus and of its functional connectivity to the cortex; (2) identify the local neuronal and network cortical and subcortical mechanisms by which source fusion or segregation is achieved and dynamically updated in a changing world; and (3) understand how local and long-range mechanisms through which priors are selected and dynamically updated at the upper level of the cortical multisensory hierarchy bias causal inference processes and downstream multisensory integration processes. We will conduct a multi-scale approach in humans and non-human primates combining in functional imaging, neurophysiology and modeling. This multidisciplinary approach will allow addressing the brain network as well as the specific neural computations and long-range modulations underpinning multisensory interaction and multisensory causal inference processes. A key aspect of the present project is to take a comparative perspective running exactly the same experiments on non-human primates and on humans, a unique opportunity to provide a functional link between EEG and fMRI human studies and single cell recording studies in the non-human primate. In both species, (1) global networks of interest and their fine topography and functional organization will be identified using fMRI and resting state-MRI while (2) the directional flow of information and its dynamical updating will be investigated using high temporal resolution methods: EEG in humans and multielectrode neuronal recordings in non-human primates. Only a systematic and cross-sectional study may provide a better understanding of multisensory interactions. A direct comparison of the neural basis of multisensory inference processes in both species using as close as possible methodologies will allow to precisely define their degree of functional homology. As a result, the NeuroCIM project is an ambitious project at the crossing of sensory physiology, cognitive systems and sensorimotor interactions. The translational specificity of this project is designed to understand the mechanisms underlying multisensory perception with straightforward applications in the field of cortical plasticity and sensory compensation for rehabilitation perspectives. As a result, this challenging basic science project is expected to have important clinical implications. The present project has strong links with other public health issues (psychiatric disorders, sensory deficits and their rehabilitation).

Project coordination

Suliann BEN HAMED (Centre de Neuroscience Cognitives)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.


CNC - CNRS Centre de Neuroscience Cognitives
CerCo Centre de Recherche Cerveau & Cognition

Help of the ANR 614,109 euros
Beginning and duration of the scientific project: December 2016 - 48 Months

Useful links

Explorez notre base de projets financés



ANR makes available its datasets on funded projects, click here to find more.

Sign up for the latest news:
Subscribe to our newsletter