Everything flows: Functional ultrasound imaging and neuromodulation of the brain and spinal cord

Man has always been fascinated by exposing and manipulating the human brain and spinal cord: from the Incas performing ritualistic trepanations, to the first modern neurosurgeons introducing systematic use of electrical stimulation to guide surgeries.

In Chapter 1 I argue that much of what is current neurosurgical practice remains similar to techniques used over a century ago, despite technological advancements introducing new tools into the neurosurgical operating room. The work in this thesis is fueled by a fascination for this ambiguity in neurosurgery, aiming to introduce new technological tools to an old-school form of craftsmanship.

Specifically, this thesis focusses on two types of neurotechnology: neuromodulation and (functional) ultrasound imaging of the brain and spinal cord, both applied to clinical and neurosurgical contexts. In three parts, the work clusters itself around specific ‘types of flow’: flow of current, flow of blood and the combination of current & blood flow.

In the first part (Current Flows), I set out to study the Dorsal Root Ganglion (DRG) as a new anatomical target for electrical neuromodulation in Spinal Cord Injury (SCI), showing theoretical potential for higher selectivity as compared to Epidural Electrical Stimulation (EES), the current gold standard. Chapter 2 describes a first in-human study looking at motor recovery using DRG-stimulation in patients with motor complete SCI. For this study, we made off-label use of commercially available DRG-leads to treat chronic pain symptoms. In five patients with SCI, we showed how bilateral L4-level DRG-stimulation could evoke reproducible knee extension movements, strong enough to facilitate assisted weight-bearing. In Chapter 3 we discuss a second series of five patients, in which three unexpectedly presented as non-responders to the DRG-stimulation. Using a test battery including clinical and neurophysiological measurements, we determined post-hoc that the complete absence of plasticity-related complaints was a distinguishing factor between responders and non-responders, in line with prior reports on EES. In Chapter 4, we discuss two unique cases with unexpected effects upon DRG-stimulation. Firstly, activation of bilateral rhythmic motor response in the legs upon unilateral L2-level DRG-stimulation, mimicking a Central Pattern Generator (CPG). Secondly, a case of suppression of lower limb spasticity upon bilateral L2-level DRG-stimulation over the course of five days. Both cases provided insights on mechanisms underlying the effect of DRG-stimulation in SCI. Finally, in Chapter 5, we shift gears to a pre-clinical study using a wireless, closed-loop optoelectronic system to perform optogenetic neuromodulation in mice-models of SCI. The ultimate goal of this pre-clinical optogenetic model was similar to the efforts of in-human DRG-stimulation in the previous chapters: increasing (spatial) selectivity of stimulation and increasing our mechanistic understanding of neuromodulation in SCI. The optoelectronic device was able to reveal the role of various neuronal subtypes, sensory pathways and supraspinal projections in the control of locomotion in healthy and SCI-model mice.

In the second part (Blood Flows), I studied μDoppler-imaging in pre-clinical and neurosurgical context. Hemodynamic μDoppler-imaging makes use of a so-called high-frame-rate (HFR) ultrasound acquisition scheme to boost the sensitivity of conventional Doppler ultrasound. Chapter 6 and Chapter 7 describe the murine and in-human application of μDoppler-imaging on the spinal cord. Both in mice, and in a human subject undergoing resection of a hemangioblastoma, μDoppler was able to capture in real-time the hemodynamic features of the healthy and tumorous spinal cord tissue with submillimeter resolution. In Chapter 8 we describe the application of μDoppler-imaging in the context of in-human cerebral pathology, describing a case of an arteriovenous malformation (AVM) in which 2D- and 3D-μDoppler-imaging was able to reveal unique vascular details of the pathological and healthy brain tissue. We discuss the potential of μDoppler-imaging as future imaging technique useful for real-time surgical feedback or even hemodynamics-based tumor delineation.

In the third part (Current & Blood Flows), I focus on functional Ultrasound (fUS)-imaging, the functional equivalent of μDoppler-imaging, which relies on the phenomenon of Neurovascular Coupling (NVC). Because of NVC, we can use hemodyamics (‘blood flows’) as a proxy of neuronal activity (‘current flows’). This same principle underlies currently established techniques such as functional Magnetic Resonance Imaging (fMRI). Chapter 9 starts out with an extensive review of all currently available clinical and experimental techniques which shows potential for functional brain imaging in the intra-operative context. Comparing and contrasting these techniques on underlying biological substrate, technical characteristics, and clinical applicability, clearly points out the unique position that fUS takes up within the imaging landscape. In Chapter 10 we describe our first in-house fUS-experiments in ten patients undergoing awake brain surgery for tumor removal. We demonstrate fUS’ ability to image functional motor- and language-related brain areas, with high spatiotemporal resolution at large fields of view, all with the same ease of use and mobility as conventional ultrasound. Chapter 11 marks important technical developments, which allowed us to build a surgical ecosystem in which we could perform co-registered functional imaging using ESM, fMRI and fUS in the same human subject. With the help of three patients undergoing awake brain surgery, we were able to consistently confirm overlap between fUS-defined functional brain regions and those defined by ESM and fMRI for a range of motor, language and visual tasks. This marks the first-ever in-human confirmation of spatial overlap between these three imaging modalities, an important milestone towards the actual clinical maturity of fUS. In Chapter 12 we undertake an important technical challenge towards this same clinical maturity: finding new ways to improve the functional sensitivity of our 2D-fUS maps. Chapter 13 marks an exciting migration from the surgical room to the real word. In an effort to work towards actual transcranial applications of fUS, we imaged two participants with a skull bone defect covered by a sonotransparent plastic (PEEK). Our experiments show our ability to image functional activity in the sensorimotor cortex of the mouth with the help of 3D-printed, personalized fUS-helmets to fixate the probe on the subject’s head. We confirm our fUS-based functional brain regions using co-registered fMRI and show the robustness and reproducibility of these fUS- signals, across subjects and over time.

In Chapter 14 I discuss the future for both DRG-stimulation in the context of SCI and (functional) Ultrasound-imaging for the clinical context. I discuss a combination of technical, neuroscientific and clinical challenges which will need to be overcome synergistically for either of the techniques to see clinical maturity. Finally, I discuss the synergy between the three separate parts of this thesis, highlighting recurrent themes such reproducibility, resolution and ecology. The latter concept forms the heart of my ultimate dream: using ecological fUS brain mapping to guide a patient’s surgical procedure, as well as their post-operative rehabilitation and neuromodulation trajectory. In the conception of this ambition, we see the three types of flow discussed in this thesis unite, truly exemplifying that indeed, πάντα ρεῖ, everything flows.