Focused Ultrasound Neuromodulation

Ultrasound has been consistently reported for neuronal stimulation for several decades in both animals and humans including eliciting brain activity detected by functional MRI and electroencephalography. In addition, this knowledge can be used to understand the differences between normal and pathological brains to treat patients. In the peripheral nervous system, the leading technique to treat peripheral neurological disorders is implantation of electrodes along the peripheral nerve and stimulating the nerve with electrical current. A noninvasive alternative that could treat neuropathic pain and suppress nerve activity constitutes thus an important challenge in interventional neurology.Our group has been studying the noninvasive stimulation or inhibition of both the central and peripheral nervous system in live animals. In the brain, we have shown that focused ultrasound is capable of noninvasively stimulating paw movement as well as sensory responses such as pupil dilation and eye movement when different brain regions are targeted, showing for the first time that ultrasound can trigger both motor and sensory brain responses. In the periphery, when the ultrasound beam is focused on the sciatic nerve in a live, anesthetized animal, the thigh muscle becomes activated and muscle twitches can be induced at low ultrasonic intensities while the same twitches can be inhibited at higher intensities due to associated temperature rise that inhibits nerve firing. In humans, a peripheral sensory response has also been elicited. Cellular and fiber responses in excised tissue have confirmed the live animal responses. An overview of the aforementioned findings together with the most recent results will be presented.

We carried out an investigation in 18 healthy volunteers in which 500 kHz ultrasound was focused to two visual brain areas: MT and FFA.  Target sites in participants were identified using MRI.  Computer modelling was used to determine the optimal placement of a single element transducer in order to insonify target regions on the right hemisphere and to ensure exposure parameters were limited. In the trials, the transducer was mounted to an arm an positioned using a neuronavigation system.  Subjects were instrumented with EEG electrodes and asked to carry out behavioral tasks: either detecting motion of dots or identifying faces.  The 500 kHz ultrasound was modulated with a 1 kHz square wave and participants could perceive audible sound during FUN which was confirmed by the presence of an auditory evoked potential in EEG recordings.  Ex vivo skull experiments suggested that ultrasound was absorbed by the skull resulting in a 1 kHz flexural wave that propagated to the ear canals.  We were able to mask the auditory confound by playing a 1 kHz tone through earphones while FUN was applied.  For the behavioural task preliminary analysis suggests that ultrasound neurostimulation resulted in improvement  in performance at the threshold for movement coherence when MT was targeted.  When FFA was targeted there was an improvement in hit-rate for facial recognition just below threshold and an increase in the amplitude of the N180 ERP.  In conclusion our results show that auditory masking can be employed to blind participants to an auditory confound and that ultrasound neurostimulation can modulate the response to visual stimuli at threshold in both MT and FFA.

Transcranial focused ultrasound (tFUS)-based strategies for non-invasive and localized neuromodulation are rapidly expanding; however, the effects of plasma protein binding (PPB), which plays an important role in drug pharmacokinetics, particularly for central nervous system drugs, have long been overlooked. We demonstrate the non-invasive and localized unbinding of PPB from phenytoin, an anti-epileptic drugs with high affinity to albumin, using FUS delivered at a fundamental frequency of 250 kHz in a pulsed manner (55 ms tone burst duration, 4 Hz pulse repetition frequency), at 5 W/cm2 spatial peak pulse average intensity (Isppa). No microbubble-based ultrasound contrast agent was used. Equilibrium dialysis was performed on the sonicated buffered-saline containing therapeutic concentration of phenytoin and bovine serum albumin, and yielded a 27.7% elevation in the level of unbound phenytoin compared to an unsonicated control. Sonication of a brain hemisphere in rats (n = 10) that received intraperitoneal phenytoin, though immunohistochemistry (IHC) assessment, showed elevated regional phenytoin uptake compared to the unsonicated hemisphere, without actively disrupting the BBB. Temperature change was not seen. These findings demonstrate the use of FUS as a novel technique for spatially-selective unbinding of PPB, which may alter a wide range of drug–plasma protein interactions. The study suggests the utiliy of trancranial FUS in elevating local concentration of a free, unbound pharmaceutical agent and subseuqent neuromodulatory effects by the agent.

When combined with imaging guidance focused ultrasound (FUS) provides means for localized delivery of mechanical energy deep into tissues. This focal energy deposition can modify tissue function via direct thermal or mechanical interactions with the tissue.  The impact of an ultrasound exposure can be  potentiated by intravascular microbubbles (MB). FUS in combination with MBs has been extensively explored for the enhancement of the permeability of the blood-brain barrier (BBB) and it has been shown to be  an effective method for local delivery of molecular agents, particles and even cells into targeted locations of the brain. The first clinical trials will be reviewed and demonstrate feasibility and safety in a clinical setting. The potential of this method for neuromodulation will be discussed.

For historical reasons inherited from the field of biomedical ultrasound the most common working hypotheses to explain neuromodulation produced by low intensity focused ultrasound (FUS) are thermal rise and cavitation. However recent evidence has not confirmed these hypotheses and points towards other mechanisms such as changes in the action potential produced by the electrophysiological-mechanical coupling in the neuronal membrane to explain neuromodulation by FUS (Jerusalem et al. Acta Biomaterialia 2019). Recent advances in atomic force microscopy (AFM) has made it possible to investigate the effect of acoustic signals on membranes and living cells with nanometre resolution (Al-Rekabi, Contera, PNAS 2018) in a large range of time scales (from Hz to MHz) . In my talk I will introduce the techniques we are developing to study the effect of ultrasound on membranes and cells mechanical properties, I will also discuss how ultrasound can potentially modulate the synchronisation of neurons by altering the mechanical properties of brain tissues across the scales.

Low intensity transcranial focused ultrasound (LIFU) is a promising non-surgical low-energy technique for transient neuromodulation with high spatial resolution and adjustable focus suitable for targeting cortical and sub-cortical human brain regions. Here, we will discuss current state of the art as well as progress and challenges for developing LIFU for human clinical applications. 

Research by Professor Antoine Jerusalem and his team focusses on electrophysiological-mechanical coupled pulses in neural membranes. This endeavour is set to benefit the medical community in the diagnosis, prognosis, and treatment of Traumatic Brain Injury and Spinal Cord Injury, both major, global public health issues, while providing new avenues for non-invasive electrophysiological control, such as pain management.

Computational methods for dosimetry allow estimating and optimizing the spatial distribution and strength of the ultrasound waves transmitted to the brain. These calculations are based on detailed models of the head anatomy that are derived from medical imaging data. In this talk, I will review the state-of-the-art in computational dosimetry for focused ultrasound neuromodulation, and discuss the remaining challenges.

To understand brain circuits we need to more than measurement; we need to modulate neural activity as well. This is even more critical in a clinical context. Current interference techniques are either highly invasive, or restricted to the surface of the brain. Focused transcranial ultrasound has the potential to overcome these limitations. I will discuss how we can use low-intensity ultrasound to safely stimulate deep brain structures, such as the amygdala, with high precision. To date, effects were short-lived, limiting its relevance to the clinic. In pioneering non-human primate work, we showed how a novel repetitive protocol induces hour-long plastic changes. Whole-brain fMRI allowed us to track how ultrasound impacted both local and remote circuits. Behavioural measurements revealed the specific consequences of a focal neural perturbation on cognitive computations. I will discuss work on the causal roles of the anterior cingulate cortex, amygdala, basal forebrain, and ventromedial prefrontal cortex in decision-making and learning.

In a large animal model and using MR-ARFI targeting, we studied the effect of transcranial focused ultrasound sonication of the lateral geniculate nucleus on visual evoked potentials. Our study included a histological evaluation of all sonicated locations with neuromodulatory pulses and with MR-ARFI pulses. We found that focused ultrasound reversibly suppressed the VEPs peak-to-peak amplitude compared to controls, and returned to baseline gradually over 1 hour, with the suppression proportional to the focal intensity. Our histological study showed no differences between sonication locations and control.

The soliton theory provides an alternative explanation for the nervous impulse based on cooperative melting transitions in the biomembrane.  It relies on reversible electromechanical processes and the thermodynamics of membranes. During the nervous impulse, the membrane is shifted through its transition, with associated changes in heat capacity, compressibility and inherent time-scales. The latent heat of the transition explains the sign and the magnitude of the experimentally observed reversible heat production of the action potential and the observed changes in nerve dimension (shortening and increase in membrane thickness). The dimensional changes give rise to the possibility to excite the nerve mechanically or thermally.

In the transition, one finds fluctuations in the membrane permeability corresponding to spontaneous pore formation, which in a patch clamp experiment are indistinguishable from the quantized current events usually attributed to protein channels.  They show most features of protein channels including temperature- and mechano-sensitivity, and voltage-gating. Their characteristic open time scales can be understood in the context of the fluctuation theorems applied to a membrane.

Membrane fluctuations therefore provide an estimate for the most suitable time-scale of membrane excitation. After a perturbation, relaxation processes in artificial membranes may display relaxation time scales up to several seconds. In biomembranes close to transitions, the expected relaxation time scales are expected to be in the millisecond regime. This is also the time scale of the open-lifetimes of membrane pores - and as it happens, the typical lifetime of protein channels. Therefore, we argue that this is also the time-scale of the most effective nerve membrane excitation.

Focused Transcranial Ultrasound Stimulation enables us to non-invasively and transiently modulate activity in subcortical and deep cortical brain structures, without the need of any surgical procedure. I am particularly interested in using low-intensity ultrasound stimulation to investigate the neural and behavioural effects resulting from a transient interference of neural communication between different brain regions within a circuitry. The capacity of transcranial focused ultrasound stimulation to alter activity in regions located deep below cortex will allow us to better understand the contribution of these structures to normal and pathological cognition.

Ultrasound has the ability to focus energy to a small point beyond the skull and is being widely explored by researchers as a tool for non-invasive neuromodulation. When combined with magnetic resonance imaging (MRI), focused ultrasound (FUS) can be precisely guided while the effects of FUS can be visualized at the network level using fMRI. In this talk, I will discuss our ongoing work in developing systems to apply image-guided FUS neuromodulation in the MRI environment while imaging functional activity. Specifically, I will cover the development of optical tracking as a method to guide FUS neuromodulation, the creation of transducer arrays for steerable FUS neuromodulation, and the development of MR acoustic radiation force imaging methods to visualize the acoustic focus. We have used these methods to modulate the somatosensory network in non-human primates, demonstrating that MRI-guided FUS is capable of exciting precise targets in somatosensory areas 3a/3b, causing downstream activations in off-target brain regions within the circuit which we simultaneously detected with fMRI. Our observations are consistent with others’ work in the field of FUS neuromodulation; however, questions remain about mechanisms underlying FUS neuromodulation and potential confounds. The talk will conclude by reporting on recent work at the cellular level where we are measuring calcium signaling in mouse brain slices with optical markers during FUS neuromodulation.

Accurate computational models of ultrasound propagation through the skull are essential for optimising, planning, and delivering ultrasonic neuromodulation in the brain. To evaluate their accuracy, models of transcranial ultrasound propagation are validated against measurements of ultrasound fields after propagation through human skulls. I will discuss our approach to performing these validation studies, the sources of uncertainty, and the challenges involved. I will also give an overview of the datasets used for validation. These include micro CT and clinical CT scans of skulls, and ultrasonic source characterisations, and will be made publicly available via the UCL open access data repository.