"What is a scientist after all? It is a curious man looking through a keyhole, the keyhole of nature, trying to know what's going on."
Jacques Cousteau, oceanographer and inventor
In the last decade, functional ultrasound imaging, invented in the Tanter lab in Paris, opened new vistas of the brain. Scroll down for significant technological and scientific advances enabled by fUS technology.
Multimodal insight. Perform unbiased whole-brain imaging of brain circuits in awake, active animals, study cerebral blood volume variation or neurovascular coupling in genetic models, integrate easily with electrophysiology, optogenetics, guided injections or else, as you envision.
Functional ultrasound neuro imaging has been used to map the brain response to external stimuli. We realized very early in our seminal study on anesthetized rats (Mace et al, Nature Methods, 2011) , where we used functional ultrasound to map the response of the brain to whisker stimulation (including single whisker stimulation) that functional ultrasound imaging is a very sensitive neuroimaging modality with excellent spatial resolution.
Functional ultrasound (fUS) imaging evoked in the left (e) or right (f) somatosensory cortex, hindlimb part (S1HL) using electrical stimulation (5 Hz, 0.2 mA, 100 μs width for 10 s, separated by 20 s) of the right (e) or left (f) sciatic nerve applied either on the right (e) or left (f) side, respectively. (g) Time course changes in the evoked fUS response in the left S1HL following stimulation of the contralateral sciatic nerve. From: Osmanski et al, Nature Communications, 2014 - Open Access - Courtesy B. Osmanski
Functional activations can be used as a first control to validate probe positioning and animal response before more complex functional experiments. For instance, researchers using Iconeus technology are now quantifying functional response to standard stimuli to study genetically- or pharmacologically-induced alterations in the neurovascular coupling and pericyte functions (unpublished data).
The high spatial resolution of the fUS technique has allowed to create detailed retinotopic maps of the rat, mouse and even awake primates performing complex visual task.
In the recent awake mice study of the Roska lab, Emilie Macé, alumnus of the Tanter lab, has shown that functional ultrasound could reveal 87 brain structures involved in the optokinetic reflex alone (Mace et al., Neuron, 2018).
Researchers have also began to study audition in awake ferrets, an eminent model for audition research. By mapping the response to auditory stimuli with different frequency, they have been able to construct highly-detailed sonotopic maps and to demonstrate that neighboring pixels at 100 micrometers distance show independant auditory response curves, further demonstrating the high resolution of functional ultrasound (Bimbard et al, eLife, 2018).
fUS imaging reveals the tonotopic organization of cortical, sub-cortical, and intracortical auditory structures in the awake ferret. (a) Left: UFD-T of the left and right craniotomies, superimposed on an MRI scan of a ferret brain. Right: magnification of the blue bounding box (left). Auditory structures: auditory cortices (AC), medial geniculate body (MGB), inferior colliculus (IC). Other structures: hippocampus (Hip), visual cortex (VC). (b) Structural view of a tilted parasagittal slice (~30° from D-V axis) of the visual and auditory cortices (represented as a blue plane on the 3D brain). Lining delineates the cortex. (c) Upper left: Tonotopic organization of the slice described in (b). Lower left: tuning curve (mean ± sem) and average responses in %CBV (see Materials and methods) for the voxel located in the upper panel (black cross). Upper right: combination of 16 similar slices over the surface of the AC, arrow depicts slice of (b). AEG/MEG/PEG: anterior/middle/posterior ectosylvian gyrus. Lower right: 3D reconstruction of the whole AC’s functional organization. Bimbard et al, eLife, 2018, CC-BY 4.0
In behaving non-human primates, researchers have shown how functional ultrasound can be used to map the regions involved in complex task and rule handling in the Supplementary Eye Field (SEF), by following the regions response trial-by-trial with high sensitivity. fUS is able to assess local changes in cerebral blood volume during cognitive tasks, with sufficient temporal resolution to measure the directional propagation of signals. In two macaques, they observed an abrupt transient change in supplementary eye field (SEF) activity when animals were required to modify their behaviour associated with a change of saccade tasks. Notably, SEF activation could be observed in a single trial, without averaging (Dizeux et al., Nature Communications, 2019).
Functional ultrasound neuro imaging is now also being investigated in non-human primates for its high potential for Brain Computer Interfaces (BCI).
Finally, the Charpak lab provided recently an extensive comparison between calcium imaging, bold-fMRI at very high field (17.2 Tesla) and functional ultrasound neuro imaging in the same animal. By imaging functional responses in the mice olfactory bulb during odor presentation, researchers conclude that functional ultrasound “is a very efficient technique for measuring mesoscopic vascular responses. Its high SNR and temporal resolution allows the generation of voxel-based correlation maps even at low-odour concentration.” whereas the very same experiment “ was not ideal for BOLD-fMRI mapping, due to the lower SNR and temporal resolution of our BOLD-fMRI protocol.” (Boido et al, Nature Communication, 2019).
Imaging intrinsic brain connectivity, often called resting-state connectivity in clinical fMRI studies, is a powerful non-invasive approach to map dynamic brain networks through detecting correlated fluctuations of spontaneous blood flow. A seminal fUS study, co-authored by a Paris research team, including three future Iconeus founders, has shown in 2014 that fUS allows to detect highly-resoved and highly-contrasted intrinsic brain connectivity patterns in the rat brain (Osmanski et al. Nature Communications, 2014).
Correlation matrices of the functional connectivity in the rat. (a,c) Matrices at Bregma +0.84 mm and (b,d) matrices at Bregma −2.16 mm. (c,d) The correlation matrices obtained show a strong bilateral correlation of signals in the cingular, retrosplenial granular, motor and S1 cortices, as well as between the motor and S1 cortices. Other brain areas, such as the caudate putamen, septum, hippocampus and thalamus showed bilateral correlations and weak-to-no correlation with the somatomotor areas. Scale bar, 1.1mm. Osmanski et al, Nature Communications, 2014 - Open Access - Courtesy B. Osmanski
Global spatial modes in the rat. (a) Global spatial modes (GSMs), representing reproducible connectivity patterns (N=6), at the three investigated coronal levels. The first GSMs show synchronized haemodynamic fluctuations in the cortical ribbon. Subsequent GSMs delineate highly contrasting cortical connectivity patterns, where the midline anterior cingulate and retrosplenial cortices (red and yellow on GSM 2 and 3), prominent hubs of the putative default-mode network are temporally anticorrelated with the task-dependent lateral sensorimotor network (blue and light blue on GSM 2 and 3, respectively). These two main cortical networks are also correlated with subcortical regions such as the caudate, putamen and septum. Scale bar, 1 mm. Osmanski et al, Nature Communications, 2014 - Open Access - Courtesy B. Osmanski
Similary to fMRI, connectivity matrices, modes or seed-based maps can be used as an efficient functional readout. Functional ultrasound connectivity has been used as such to study the role of ocytocin in rat pups (Mairesseet al. Glia, 2019), but also to study pain sensitivity in anesthetized rats or pharmacologically-induced connectivity alterations in rats and mice (unpublished data).
Importatly, recent development shows that thanks to the light ultrasonic probes, robust connectivity matrices can be obtained in awake freely-moving or head-fixed mice (unpublished data), eliminating the bias of anesthetics. This technological advance may be particularly relevant to analyse genetically modified mice models.
Contrary to magnetic fields, ultrasound has little to no impact to the surroundings, it is thus straightforward to combine functional ultrasound neuro imaging with other modalities such as electro-encephalography (eeg), multi electrodes array or optical imaging.
For instance, Sieu et al. (Nature Methods, 2015) demonstrated the use of implanted electodes + functional ultrasound in the context of absence seizure in the rat, the local field potential signal being used to characterize seizures. Bergel et al. (Nature Communications, 2018) further pushed the concept to study locomotion and sleep states. By correlating the enveloppe of different EEG LFP bands with the functional ultrasound signals, the researchers have been able to reveal previously undescribed brain-wide spatiotemporal hemodynamics of single REM sleep episodes.
A multiarray silicon probe
Rungta et al. (Nature Communications, 2017) and Mace et al. (Neuron, 2018) have both demonstrated the combined use of optogenetics with functional ultrasound. Optogenetics can be used to stimulate specfiic neurons population while the unbiaised whole-brain response is simultaeously monitored.
Overall, functional ultrasound imaging is highly versatile and adaptable to different species.
Researchers worldwide have already obtained impressive results on many different species : mice, rats (including pups), pigeons, ferrets, rabbits and non-human primates.
By adapting the probe and the ultrasound frequency, we can scale from small to large animals, benefiting from the classical tradeoff between spatial resolution and imaging depth. However, as fUS imaging has a very good spatial resolution to start with (as compared to fMRI for exemple), even at 6MHz, which allows to visualize the whole human brain, we can observe and measure vasculature and cerebral blood flow with impressive details.
Different probes configuration allow to increase depth at the expanse of resolution. fUS imaging of the same visual cortex of a non-human primate seen with a different ultrasonic probe used at their central frequency. Dizeux et al, Nature Communications, 2019, Open Access, Courtesy M Tanter
1. Head-fixed (2D locomotion)
Mice, rats, ferrets as well as non-human primates can be imaged under head-fixation conditions.
Head-fixed configuration in mice with a Mobile HomeCage™ from Neurotar™, associated with an ICONEUS transducer via a compatible headplate.
2. Head-mounted (3D locomotion)
Head-mounted configuration in mice
Our probes are adapted to fit most configurations including standard primate electrophysiology chambers, such as the systems from Crist Instruments, for awake primate experiments.
Birds, ferrets, rats, rabbits and non-human primates require a specific preparation to create for an acoustic window. For chronic imaging, researchers have been using either a thin-skull window (Osmanski et al., 2015) or used an acoustically transparent polymer (Sieu et al. 2015) to replace the skull by adapting a protocol similar to that employed in chronic optical imaging.
If rats can also be imaged directly through the skull with off-the-shelf acoustic contrast agents (Errico et al., Neuroimage, 2016), mice do not require any specific preparation or contrast agents and can be imaged directly through the skull (Tiran et al., Ultrasound Med. Biol. 2017), although such preparation can further improve SNR for specific studies.
Better than BOLD. Functional ultrasound can be the right readout for the study of the effect of pharmacological agents on the brain, without the bias of anesthesia.
Essentially, functional ultrasound can be used as PharmacoMRI, fMRI and rs-fMRI at once, more sensitive, easier to use and adapted to awake configurations. fUS can be used to show time- and dose-dependent connectivity changes following pharmacological treatments, both in wild-type and genetically modified awake mice. Machine-learning approaches are are able to extract robust pharmacological scores or fingerprints of drug effects on the brain. This is a ideal tool to complement behavorial studies with solid results obtained in small cohorts of animals.
Better than BOLD. Three recent proof-of-concept studies suggest that functional ultrasound (fUS) imaging, a method capable to be used simultaneously as PharmacoMRI, fMRI and rs-fMRI, provides a powerful readout to study the effects of pharmacological agents on the brain, without the bias of anesthesia. Pharmacological-fUS or pharmaco-fUS is now validated in rats and mouse models as an interesting tool to complement behavorial studies, with solid results obtained in small cohorts of animals.
The first study, a collaborative work of Iconeus founders, (Rabut and al., NeuroImage, 2020) has shown that scopolamine, a major preclinical drug to model Alzheimer’s disease, leads to time- and dose depend changes of cerebral functional connectivity in awake mice. A machine-learning approach was used to extract a robust pharmacological score or fingerprint of the drug effects on the brain from a relatively small cohort of mice. The pharmacological pertinence of the scopolamine score was demonstrated by reversal through the antagonist milameline and the peripherally-restricted methyl-scopolamine.
Pharmacological validation for the automatic detection of scopolamine-induced connectivity changes. a)Schema of the treatment protocol. b)Pharmacological scores obtained in the 5 treatment groups. Groups either received saline (group 0) or increasing doses of scopolamine (groups 1-4) or a single dose of the peripherally- restricted methyl-scopolamine (group 5), followed by a single dose of the antagonist milameline. Paired T-tests were used to compare baseline, post-scopolamine and post-milameline injection scores for the different studied groups (∗ p < 0.05; ∗∗ p < 0.01). Bar graph: mean ± SEM. Rabut and al., NeuroImage, 2020 Open Access
The second and third studies, both from the pharmaceutical company Theranexus, demonstrated how fUS can be useful for drug development. In the first study (Vidal et al., Frontiers in Neuroscience, 2020) the authors compared the hemodynamic response of the brain to the administration of donepezil alone (potent acetylcholinesterase inhibitor, largely used worldwide to alleviate cognitive symptoms in Alzheimer's disease) versus donepezil combined with mefloquine (a new drug targeting connexins, the proteins involved in astrocyte network organization). They have showed that although administration of donepezil or mefloquine alone at low dose had only very limited effects on the signal compared to the baseline, their combination produced marked hemodynamic effects in the hippocampus, in line with previously published behavioral data demonstrating a synergic interaction between both drugs. These results confirmed the potential of the combined treatment and provided new insights on how both drugs could interact. In the third study (Vidal et al., Neuropharmacology, 2020), Theranexus investigated the effect of atomoxetine, a potent norepinephrine reuptake inhibitor and non-stimulant treatment of attention-deficit/hyperactivity-disorder, in anesthetized rats at increasing doses (0.3, 1 and 3 mg/kg). Dose-dependent temporal variations of the cerebral blood flow were observed, particularly in brain regions involved in vision.
These three studies demonstrate that pharmaco-fUS could improve our understanding of the mechanism of action of drugs in the brain and might accelerate the development of new drugs targeting neurological disorders. Pharmaco-fUS does not require the animals to be anesthetized, and therefore yields a quantitative assessment of brain functions without anesthesia-induced bias. Pharmaco-fUS currently acquires cross-sectional (2D) images of the rodent brain. Current development of volumetric (3D) imaging is expected to enable the assessment of drug effects in the entire rodent brain within a single imaging session.
Functional ultrasound enabled 3D microvascular imaging of brain hemodynamics in rodents by combining ultrasound ultrasensitive Doppler and tomographic reconstruction for 3D imaging (Demene et al., NeuroImage, 2016). The technique reaches an exceptional 100 μm isotropic spatial resolution and is sensitive enough to image vessels displaying fast to very slow blood flow (1 mm/s). Ultrafast Doppler angiography in vivo imaging could become a valuable tool for the study of brain hemodynamics, such as cerebral flow autoregulation or vascular remodeling after ischemic stroke recovery, and, more generally, brain tumor vasculature response to therapeutic treatment.
4D ultrasensitive Doppler was proposed to map dynamically relatively slow events (<20 seconds) in several slices of the brain automatically (Demene et al., NeuroImage, 2016). The approach allows to perform 3D functional ultrasound but also to monitor hemodynamics events such as stroke monitoring in the whole-brain in an efficient and easy manner. The trade-off betwee volume rate and number of slice can be adjusted to best fit the intended applications.
Functional ultrasound measures relative changes in CBV and can be used to monitor recanalisation and reperfusion in stroke models with high spatio-temporal resolution.
Martin et al. (J Cereb Blood Flow Metab., 2012) detected longitudinal changes in tissue perfusion and elasticity by using ultrafast ulrasound imaging in parallel with positron emission tomography (PET) and single photon emission computed tomography (SPECT). Brunner et al. (J Cereb Blood Flow Metab., 2018) used a rat intraluminal middle cerebral artery occlusion model to map the reperfusion in 2D and extracted time course of the occluded-side and non-occluded side somatosensory and motor structures. The authors investigated the slopes of the reperfusion. Functional ultrasound allows the characterization of the ischemic area with respect to regional differences in blood volume during transient occlusion and subsequent reperfusion.
Finally, a collaborative study (Hingot et al., Theranosics, 2020) between Iconeus founders and the team of Prof. Denis Vivien (Normandie University, Caen, France) provides the proof-of-concept, by using ultrafast Doppler and Ultrasound Localization Microscopy to image the cerebral blood volumes at early and late times after stroke onset and compare with the formation of ischemic lesions in a mouse model of thromboembolic stroke with an intact skull, that early ultrafast ultrasound can be indicative of responses to treatment and cerebral tissue fates following stroke.
Predicting the outcomes and evaluating responses to treatment with transcranial ULM. A. Timeline for ULM acquisitions. B. For mice treated with NaCl, a large lesion can be seen in ULM just after ischemia and very little reperfusion can be observed 2 h after the onset. In ultrafast Doppler, the hypoperfused area is the infarcted lesion seen in MRI. C. On rtPA treated mice, the reperfusion can be evaluated with ULM and predicts the formation of the lesion. Scale bar 1 mm. From Hingot et al, Theranostics, 2020. Open Access
In conculsion, fUS provied new and powerful tools to study ischemic stroke in preclinical models as the first step prior translation to the clinical settings.
