Application de l’imagerie fonctionnelle par ultrasons à la découverte d’empreintes pharmacologiques de la connectivité fonctionnelle chez la souris éveillée

Functional ultrasound imaging (fUSI) is a promising neuroimaging technique for studying brain function in preclinical models. fUSI offers advantages in cost, spatiotemporal resolution, and the ability to image awake, behaving animals. However, critical methodological challenges have limited its widespread adoption for awake paradigms, particularly the presence of motion artifacts that systematically vary between experimental conditions. This thesis addresses this limitation through comprehensive methodological innovations and validates fUSI for neuropharmacological research. Through systematic evaluation of nearly 800 preprocessing configurations across multiple datasets, we characterized the spatial and temporal properties of motion artifacts in awake fUSI and developed optimized denoising paradigms. Our findings demonstrate that adaptive clutter filtering outperforms traditional static methods, high pass filtering exceeds conventional band pass approaches, and anatomical CompCor confounds improve upon global signal regression for functional connectivity analysis. These advances improve the robustness of group level comparisons previously compromised by motion related confounds. We then validated fUSI as a tool for neuropharmacological investigations through systematic studies of three major drug classes—opioids, cannabinoids, and serotonergic psychedelics—in awake mice. Each drugclass produced distinct functional connectivity signatures that could serve as pharmacological “fingerprints”. The opioid study revealed cortico subcortical dysconnectivity correlating with analgesic efficacy while showing minimal association with adverse effects. The cannabinoid study demonstrated biphasic connectivity alterations featuring rapid cortical thalamic effects followed by slower hippocampal alterations. Finally, the psilocybin study showed dose dependent alterations in hippocampo thalamic networks partially sensitive to 5 HT2A receptor blockade. Pharmacological relevance was validated through evaluation of dose dependence, temporal dynamics, receptor specificity, and tolerance development. To conduct these studies, we developed research infrastructure including fUSI BIDS, a community driven data standardization framework extending established neuroimaging standardsto accommodate fUSI specific requirements, and PyfUS, a comprehensive Python analysis package that bridges academic research needs with commercial software development. PyfUS served as a testing platform for the methodological innovations developed in this work, before implementation in the Iconeus software suite. These contributions demonstrate fUSI’s potential as a cost effective neuroimaging tool for discriminating between compounds with distinct mechanisms of action, suggesting applications in pharmacological screening and drug development.