When it comes to the human body, few organs are as appealing and cryptic as the brain. Much of our current knowledge on how the brain works comes from studying its electrical activity, that is the product of hundreds of thousands of neurons firing information via neurotransmitters across axons and synapses. When acquired through electroencephalography (EEG), such data has proven particularly useful for monitoring of seizures, limb control and other neuropathologies such as dementia and memory loss. Functional magnetic resonance imaging (fMRI) is instead the go-to non-invasive imaging method for monitoring neurological function because of its sensitivity to changes in relaxivity, as dictated by the local concentrations of water and other compounds in solution. For instance, fMRI contrast agents can be used to detect brain lesions such as the presence of a tumour, intracranial hypotension and to assess blood-brain barrier permeability and rupture, such as hypotension and haemorrhage following traumatic brain injury.
Reporting in the October issue of Science Translational Medicine, Olivier Baud and colleagues now show that an ultrasound imaging modality is suitable for fast dynamic monitoring in the brains of newborns. The authors adapted the use of ultrafast Doppler (UfD) imaging, to date used to map functional brain connectivity in small animal models, to monitor cerebral activity by measuring cortical haemodynamics at high resolution (Figure 1). The technique, capable of detecting up to 10,000 frames per second, allows mapping the microvasculature with an accuracy that was so far unachievable by other haemodynamic methods, such as those provided by diffuse optics. The setup (Figure 1a-c) was built to gather functional ultrasound imaging (fUSI) data simultaneously to EEG data and, as a proof-of-concept, used to detect changes between active and quiet sleeping states, where the two methods show strong changes in fUS signal concomitantly to increased brain electrical activity (Figure 1d,e).
Figure 1: fUSI and EEG recording in neonates. (A and B) Representation of imaging plane, ultrasonic probe, and EEG electrode positions used in the study. (C) The fUSI setup that includes the EEG at the bedside, coordinated by commercial components. (D) Representative coronal view of the brain through the fontanel obtained with a conventional ultrasound device (black and white image) and stack of UfD images of the same area. (E) Example of EEG recording and respiratory trace (green) obtained during quiet and active sleep. From Demene, C. et al., Sci. Transl. Med., eaah6756 (2017). Image reproduced and caption adapted with permission from AAAS.
Baud and co-authors then moved to imaging brain activity in neonates with abnormal cortical development in a case of tuberous sclerosis complex (TSC). The technique mapped the spatial increases in blood volume to ictal electrical activity during seizures in the affected hemisphere, compared to the contralateral area. In a patient with cortical dysplasia, the authors found a correlation between the fast spikes in brain activity and slower waves of vascular dynamics where the frequency and amplitude of the signal propagation was higher in the ipsilateral hemisphere, than in the contralateral area. The results point to the ability of fUS data to spatially map the anatomical location of the source of seizures. A main point of interest in fUSI is that the data is three-dimensionally mapped with high fidelity (no motion artefacts were detected in this study) and at millimetre resolution. Instead EEG signals cannot be mapped and compared to an fMR or B-mode US image (Figure 2); rather, the detection of spatial differences relies on the positioning of the electrodes ahead of taking the measurements.
Since fUSI does not require the injection of contrast agents, the fUS imaging data can be acquired for long periods of time without loss of signal, however the transducer does need to be fitted to the neonates, thereby only allowing for minimal movement in a supine position, such as that during active sleep. Nevertheless, the ease of use and low cost of the device make for a remarkable value proposition when continuous monitoring of brain activity is desired, such as monitoring antiepileptic drug efficacy. The authors further suggest that the method could be used for mapping cortical function during brain surgery. The current setup is designed to operate specifically on neonates, for which ultrasound can travel through the anterior fontanel. Signal attenuation by the intact skull in the adult would hinder the transducer’s sensitivity. The use of microbubbles as contrast agents would however make fUSI compatible with adults, albeit for shorter periods of time.
Figure 2: Spatial correlation between UfD signal and ictal events in a newborn with TSC. (A and B) Coronal T2–weighted (A) and axial T1–weighted (B) MRI images where that arrows indicate enlargement of the right hemisphere. Diffuse abnormal white matter signal (*) and enlarged and dysmorphic right lateral ventricle (dashed arrow) are also evident. The red frame depicts the area imaged with fUSI. (C) Representative UfD image (colored) overlaid onto an ultrasound B-mode image (black and white), coloured circles represent three regions in the cortex: #1 and #2 in the right hemisphere and #3 in the left hemisphere (contralateral). From Demene, C. et al., Sci. Transl. Med., eaah6756 (2017). Image and caption adapted with permission from AAAS.
Demene, C. et al., Functional ultrasound imaging of brain activity in human newborns. Sci. Transl. Med. 9, eaah6756 (2017).