Thorough, quantitative analysis of behaviour is essential for advancing neuroscience and neurological disorder research and has the potential to significantly improve the characterization of phenotypes in genetic, pharmacological and toxicological screens commonly employed in the pharmaceutical and biotech industries. Achieving high-throughput screening, without compromising the richness of phenotypic data requires scalable, standardized and cost-effective solutions for behavioral recording that also allow seamless integration across labs. To address this need, we developed megabouts.ai (Jouary et al., Biorxiv, 2024) an AI-based pipeline for behavioral analysis capable of reliably inferring high-speed behavioral kinematics from diverse and low-resolution data sources. Complementing this, we have designed a low-cost, modular behavioral testing system optimized to leverage the strengths of the megabouts.ai pipeline, and enable sophisticated behavioral experiments compatible with high-throughput screening approaches. In this proposal, we will develop and validate these tools across a range of behavioral screening paradigms, exploring their potential applications in both basic and translational research. Our goal is to disseminate these innovations, promoting their effective use in academic and industry settings, unlocking new experimental possibilities and enhancing the capabilities of existing commercial and custom-built solutions.

Everyday behaviours, like mastering dance moves or hiking through uneven terrains, require motor learning. Although essential, cellular and circuit mechanisms for this type of learning are poorly understood. In the context of locomotion, learning can be induced on a split-belt treadmill, where each side of the body moves at different speeds; this is currently used as a rehabilitative therapy in patients presenting asymmetric gaits following brain damage. The receiving laboratory recently showed that mouse locomotor learning is remarkably similar to humans. Regaining gait symmetry is achieved through specific adjustments in interlimb coordination. Moreover, the laboratory demonstrated that this form of motor learning depends critically on intermediate cerebellum. Strikingly, spatial and temporal components of learning (i.e. “where” and “when” paws land) proceeded independently, suggesting that they might be dissociable at the neural circuit level. However, how cerebellar outputs act to calibrate spatiotemporal coordination between limbs is unknown. IP2adapt aims to decipher the functional connectivity of cerebellar output circuits conveying corrective calibration signals to modify locomotor movements in space and time. Specifically, we will: 1) Use advanced viral tools to map output circuits downstream of the cerebellum and identify candidates for spatial and temporal calibration signals; 2) Use high-resolution quantitative behavioural analysis and cutting-edge circuit manipulation to test specific functional roles of projection-defined subsets of cerebellar outputs in locomotor learning; 3) Use state-of-the-art recordings of these functionally-defined neurons during learning to understand how the cerebellum transforms locomotor errors into spatial and temporal calibration of limb movement. This work will reveal unprecedented insights into neural circuits that ensure the adaptability of motor commands, with direct relevance for rehabilitative therapy.

cDcIVA aims to reveal a gut-bone marrow axis mediated by retinoic acid (RA) as a fate-instructive factor for the development of conventional dendritic cell (cDC) precursors in the bone marrow. Communication between the intestine and other organs, such as the bone marrow is mediated by several nutrients, including Vitamin A. This proposal focuses on examining how the Vitamin A derivative, RA, reaches the bone marrow and boosts the output of a specific RA-responsive cDC subset with implications on predisposing certain effector functions of the adaptive immune system. By integrating in vitro differentiation assays, spatial transcriptomics, and in vivo RA manipulations (genetically and pharmacologically), this proposal aims to: 1) establish the role of Vitamin A, as a mediator of the gut-bone marrow axis, in regulating cDC development, 2) demonstrate that retinoic acid is the sole Vitamin A metabolite that can have cell intrinsic effects on cDC progenitors and 3) reveal the cells that metabolise and supply RA to cDC progenitors and form part of an instructive bone marrow niche. cDcIVA will advance our understanding of how the interplay between nutrition and immunity can regulate the development of the cells that govern the effector functions of the adaptive immune system, offering valuable insights for future research in vaccination and cancer immunotherapy.

Functional-Magnetic Resonance Imaging (fMRI) has transformed our understanding of brain function due to its ability to noninvasively tag ‘active’ brain regions. Nevertheless, fMRI only detects neural activity indirectly, by relying on slow hemodynamic couplings whose relationships with underlying neural activity are not fully known. We have recently pioneered two unique MR approaches: Non-Uniform Oscillating-Gradient Spin-Echo (NOGSE) MRI and Relaxation Enhanced MR Spectroscopy (RE MRS). NOGSE-MRI is an exquisite microstructural probe, sensing cell sizes (l) with an unprecedented l^6 sensitivity (compared to l^2 in conventional approaches); RE MRS is a new spectral technique capable of recording metabolic signals with extraordinary fidelity at ultrahigh fields. This proposal aims to harness these novel concepts for mapping neural activity directly, without relying on hemodynamics. The specific objectives of this proposal are: (1) Mapping neural activity via sensing cell swellings upon activity (μfMRI): we hypothesize that NOGSE can robustly sense subtle changes in cellular microstructure upon neural firings and hence convey neural activity directly. (2) Probing the nature of elicited activity via detection of neurotransmitter release: we posit that RE MRS is sufficiently sensitive to robustly detect changes in Glutamate and GABA signals upon activation. (3) Network mapping in optogenetically-stimulated, behaving mice: we propose to couple our novel approaches with optogenetics to resolve neural correlates of behavior in awake, behaving mice. Simulations for μfMRI predict >4% signal changes upon subtle cell swellings; further, our in vivo RE MRS experiments have detected metabolites with SNR>50 in only 6 seconds. Hence, these two complementary –and importantly, hemodynamics-independent– approaches will represent a true paradigm shift: from indirect detection of neurovasculature couplings towards direct and noninvasive mapping of neural activity in vivo.