From fMRI to fUSI: advancing large-scale brain mapping in rodents

Investigating the principles that underlie human brain function lies at the heart of neuroscience. Among the available tools, functional magnetic resonance imaging (fMRI) has become the reference method for mapping large-scale brain activity in humans. Its non-invasive nature, high spatiotemporal resolution, and ability to capture whole-brain dynamics make it uniquely suited to study neural processes both during task performance and at rest. While fMRI encompasses a broad range of experimental paradigms, resting-state fMRI (rsfMRI) has gained increasing importance as a powerful tool for characterizing the brain’s intrinsic functional organization. During rest, spatially distinct brain regions exhibit temporally correlated fluctuations in the blood-oxygen-level-dependent (BOLD) signal, forming what are known as resting-state networks (RSNs). In this context, the correlated activity between brain regions, referred to as functional connectivity, is often interpreted as a proxy for inter-regional communication. These networks are thought to underpin fundamental aspects of brain function and have been shown to be disrupted across a wide range of neuropsychiatric disorders. Despite its transformative impact on human neuroscience, fMRI remains limited in its ability to directly probe the physiological mechanisms underlying RSNs or their alterations. Addressing these questions requires the integration of invasive approaches that are only feasible in physiologically-accessible model organisms, motivating the development of complementary preclinical imaging methods. Several studies have successfully extended fMRI to animal models, yielding highly promising insights into large-scale brain systems. However, the implementation of fMRI in small animals remains constrained by substantial practical and experimental limitations. The high cost of MRI infrastructure, together with the rigid requirements imposed by the magnetic field and the confined scanner bore, severely limit experimental flexibility and portability, constraining the range of configurations that can be implemented. As a result, there is a growing need for alternative brain-wide imaging methods that preserve cross-species relevance, while offering enhanced versatility in experimental implementation. Functional ultrasound imaging (fUSI) has recently emerged as a powerful neuroimaging method with the potential to overcome many of these limitations. By measuring changes in cerebral blood volume through ultrasound, fUSI can map large-scale brain activity in small laboratory animals with high spatiotemporal resolution, while remaining portable, non-invasive, and relatively low-cost. Its compact and near-silent design enables monitoring of brain activity under experimental conditions that are difficult to access with conventional MRI. These features position fUSI as a promising alternative to fMRI, with the potential to transform both preclinical neuroscience and future clinical applications. Despite this promise, it remains unclear whether fUSI can map distributed RSNs in rodents in a manner comparable to fMRI. Establishing this correspondence is essential to validate fUSI as a tool for systems-level neuroscience. To fill this methodological gap, in the present work I established a robust protocol for mapping large-scale connectivity networks in mice using fUSI under resting-state conditions. The resulting fUSI-derived networks revealed highly consistent large-scale organization, including a Default Mode Network and a Laterocortical Network, reproducing previous fMRI findings. I next examined the relationship between these functional networks and the underlying anatomical structure of the mouse brain, demonstrating that fUSI-derived RSNs are robustly embedded within the structural connectome. Notably, this structure-function relationship could be parsimoniously captured by four dominant axes, which differentially relate functional systems to known anatomical substrates. Beyond this static organization, I further analyzed the spatiotemporal dynamics of RSNs using framewise clustering approaches, showing that fUSI reliably captures hallmark features of rsfMRI dynamics. Specifically, fUSI reveals the presence of recurrent co-activation patterns (CAPs), and a structured transition architecture that converges onto three stable attractor modes. By showing that fUSI reproduces the connectome-scale network organization detected with fMRI, my work establishes fUSI as a robust alternative for mapping brain-wide functional connectivity in rodents. Beyond validating its reliability, this research highlights the potential of fUSI for future investigations of large-scale brain dynamics beyond the constraints of conventional MRI, paving the way for novel approaches in translational neuroscience.