A causal study of the phenomenon of ultrasound neurostimulation applied to an in vivo invertebrate nervous model

A review of the bioeffects of low-intensity focused ultrasound and the benefits of a cellular approach

This review article highlights the historical developments and current state of knowledge of an important neuromodulation technology: low-intensity focused ultrasound. Because compelling studies have shown that focused ultrasound can modulate neuronal activity non-invasively, especially in deep brain structures with high spatial specificity, there has been a renewed interest in attempting to understand the specific bioeffects of focused ultrasound at the cellular level. Such information is needed to facilitate the safe and effective use of focused ultrasound to treat a number of brain and nervous system disorders in humans. Unfortunately, to date, there appears to be no singular biological mechanism to account for the actions of focused ultrasound, and it is becoming increasingly clear that different types of nerve cells will respond to focused ultrasound differentially based on the complement of their ion channels, other membrane biophysical properties, and arrangement of synaptic connections. Furthermore, neurons are apparently not equally susceptible to the mechanical, thermal and cavitation-related consequences of focused ultrasound application-to complicate matters further, many studies often use distinctly different focused ultrasound stimulus parameters to achieve a reliable response in neural activity. In this review, we consider the benefits of studying more experimentally tractable invertebrate preparations, with an emphasis on the medicinal leech, where neurons can be studied as unique individual cells and be synaptically isolated from the indirect effects of focused ultrasound stimulation on mechanosensitive afferents. In the leech, we have concluded that heat is the primary effector of focused ultrasound neuromodulation, especially on motoneurons in which we observed a focused ultrasound-mediated blockade of action potentials. We discuss that the mechanical bioeffects of focused ultrasound, which are frequently described in the literature, are less reliably achieved as compared to thermal ones, and that observations ascribed to mechanical responses may be confounded by activation of synaptically-coupled sensory structures or artifacts associated with electrode resonance. Ultimately, both the mechanical and thermal components of focused ultrasound have significant potential to contribute to the sculpting of specific neural outcomes. Because focused ultrasound can generate significant modulation at a temperature <5°C, which is believed to be safe for moderate durations, we support the idea that focused ultrasound should be considered as a thermal neuromodulation technology for clinical use, especially targeting neural pathways in the peripheral nervous system.

Focused Ultrasound Stimulation of an ex-vivo Aplysia Abdominal Ganglion Preparation

BACKGROUND

A growing body of research demonstrates that focused ultrasound stimulates activity in human and other mammalian nervous systems. However, there is no consensus on which sonication parameters are optimal. Furthermore, the mechanism of action behind ultrasound neurostimulation remains poorly understood. An invertebrate model greatly reduces biological complexity, permitting a systematic evaluation of sonication parameters suitable for ultrasound neurostimulation.

NEW METHOD

Here, we describe the use of focused ultrasound stimulation with an ex-vivo abdominal ganglion preparation of the California sea hare, Aplysia californica, a long-standing model system in neurobiology. We developed a system for stimulating an isolated ganglion preparation while obtaining extracellular recordings from nerves. The focused ultrasound stimulation uses one of two single-element transducers, enabling stimulation at four distinct carrier frequencies (0.515 MHz, 1.l MHz, 1.61 MHz, 3.41 MHz).

RESULTS

Using continuous wave ultrasound, we stimulated the ganglion at all four frequencies, and we present quantitative evaluation of elicited activation at four different sonication durations and three peak pressure levels, eliciting up to a 57-fold increase in spiking frequency.

COMPARISON WITH ELECTRICAL STIMULATION

We demonstrated that ultrasound-induced activation is repeatable, and the response consistency is comparable to electrical stimulation.

CONCLUSIONS

Due to the relative ease of long-term recordings for many hours, this ex-vivo ganglion preparation is suitable for investigating sonication parameters and the effects of focused ultrasound stimulation on neurons.

Focusing in on the Future of Focused Ultrasound as a Translational Tool

This article summarizes the field of focused ultrasound for use in neuromodulation and discusses different ways of targeting, delivering, and validating focused ultrasound. A discussion is focused on parameter space and different ongoing theories of ultrasonic neuromodulation. Current and future applications of the technique are discussed.

White matter tract transcranial ultrasound stimulation, a computational study

Low-intensity transcranial ultrasound stimulation (TUS) is poised to become one of the most promising treatments for neurological disorders. However, while recent animal model experiments have successfully quantified the alterations of the functional activity coupling between a sonicated target cortical region and other cortical regions of interest (ROIs), the varying degree of alteration between these different connections remains unexplained. We hypothesise here that the incidental sonication of the tracts leaving the target region towards the different ROIs could participate in explaining these differences. To this end, we propose a tissue level phenomenological numerical model of the coupling between the ultrasound waves and the white matter electrical activity. The model is then used to reproduce in silico the sonication of the anterior cingulate cortex (ACC) of a macaque monkey and measure the neuromodulation power within the white matter tracts leaving the ACC for five cortical ROIs. The results show that the more induced power a white matter tract proximal to the ACC and connected to a secondary ROI receives, the more altered the connectivity fingerprint of the ACC to this region will be after sonication. These results point towards the need to isolate the sonication to the cortical region and minimise the spillage on the neighbouring tracts when aiming at modulating the target region without losing the functional connectivity with other ROIs. Those results further emphasise the potential role of the white matter in TUS and the need to account for white matter topology when designing TUS protocols.

Effects of Noninvasive Low-Intensity Focus Ultrasound Neuromodulation on Spinal Cord Neurocircuits In Vivo

Although neurocircuits can be activated by focused ultrasound stimulation, it is unclear whether this is also true for spinal cord neurocircuits. In this study, we used low-intensity focused ultrasound (LIFU) to stimulate lumbar 4–lumbar 5 (L4–L5) segments of the spinal cord of normal Sprague Dawley rats with a clapper. The activation of the spinal cord neurocircuits enhanced soleus muscle contraction as measured by electromyography (EMG). Neuronal activation and injury were assessed by EMG, western blotting (WB), immunofluorescence, hematoxylin and eosin (H&E) staining, Nissl staining, enzyme-linked immunosorbent assay (ELISA), immunohistochemistry (IHC), somatosensory evoked potentials (SEPs), motor evoked potentials (MEPs), and the Basso–Beattie–Bresnahan locomotor rating scale. When the LIFU intensity was more than 0.5 MPa, LIFU stimulation induced soleus muscle contraction and increased the EMG amplitudes (P < 0.05) and the number of c-fos- and GAD65-positive cells (P < 0.05). When the LIFU intensity was 3.0 MPa, the LIFU stimulation led to spinal cord damage and decreased SEP amplitudes for electrophysiological assessment (P < 0.05); this resulted in coagulation necrosis, structural destruction, neuronal loss in the dorsal horn by H&E and Nissl staining, and increased expression of GFAP, IL-1β, TNF-α, and caspase-3 by IHC, ELISA, and WB (P < 0.05). These results show that LIFU can activate spinal cord neurocircuits and that LIFU stimulation with an irradiation intensity ≤1.5 MPa is a safe neurostimulation method for the spinal cord.

Effect of Therapeutic Ultrasound on the Release of Insulin, Glucagon, and Alpha-Amylase from Ex Vivo Pancreatic Models

OBJECTIVES

Our previously published studies showed the potential of therapeutic ultrasound (US) as a novel non-pharmacological alternative for the treatment of secretory deficiencies in type 2 diabetes. Despite showing enhanced insulin release from beta cells, these studies did not explore the potential effects of US treatment on other cells in the islets of Langerhans such as glucagon-secreting alpha cells or acinar cells of the exocrine pancreas.

METHODS

We applied US parameters found capable of safely stimulating insulin secretion from pancreatic beta cells (f = 800 kHz, ISPTA  = 0.5-1 W/cm2 , 5 minutes) to a diced rabbit pancreas model in culture plates (n = 6 per group). Released quantities of insulin and glucagon in response to US treatment were measured by collecting aliquots of the extracellular medium prior to the start of the treatment (t = 0 minute), immediately after treatment (t = 5 minutes) and 30 minutes after the end of treatment (t = 35 minutes). Potential release of digestive enzyme alpha-amylase as a result of US treatment was evaluated in rabbit pancreas experiments. Preliminary studies were also performed in a small number of human pancreatic islets in culture plates (n = 3 per group). The general integrity of the US-treated rabbit pancreatic tissue and human pancreatic islets was evaluated through histological analysis.

RESULTS

While sham-treated rabbit pancreas samples showed decreased extracellular insulin content, there was an increase in insulin release at t = 5 minutes from samples treated with US at 800 kHz and 1 W/cm2 (P <.005). Furthermore, no further insulin release was detected at t = 35 minutes. No statistically significant difference in extracellular glucagon and alpha-amylase concentrations was observed between US-treated and sham rabbit pancreas groups. Preliminary studies in human islets appeared to follow trends observed in rabbit pancreas studies. Islet and other pancreatic tissue integrity did not appear to be affected by the US treatment.

CONCLUSION

A potential US-based strategy for enhanced insulin release would require optimization of insulin secretion from pancreatic beta cells while minimizing glucagon and pancreatic enzyme secretions.

Neurostimulation success rate of repetitive-pulse focused ultrasound in an in vivo giant axon model: An acoustic parametric study

PURPOSE

Focused ultrasound (FUS) is a promising tool to develop new modalities of therapeutic neurostimulation. The ability of FUS to stimulate the nervous system, in a noninvasive and spatiotemporally precise manner, has been demonstrated in animals and human subjects, but the underlying biomechanisms are not fully understood yet. The objective of the present study was to investigate the bioeffects involved in the generation of trains of action potentials (APs) by repetitive-pulse FUS stimuli in a simple invertebrate neural model.

METHODS

The respective influences of different acoustic parameters on the neurostimulation success rate (NSR), defined as the rate of FUS stimuli capable of evoking at least one AP, were explored using the system of afferent nerves and giant fibers of Lumbricus terrestris as neural model. Each parameter was studied independently by administering random FUS sequences while keeping all but one FUS parameter constant. The NSR was evaluated as a function of (i) the spatial-average pulse-average intensity (I_sapa ); (ii) the pulse duration (PD); (iii) the pulse repetition frequency (PRF); iv) the number of cycles per pulse (N_cycles ); (v) two ultrasound frequencies, f₀  = 1.1 MHz and f₃  = 3.3 MHz, corresponding to the fundamental and third-harmonic resonant frequencies of the FUS transducer, respectively (spherical, radius of curvature: 50 mm); and (vi) levels of emerging stable cavitation and inertial cavitation.

RESULTS

The NSR associated to 1.1 MHz repetitive-pulse FUS stimuli was found to increase as a function of increasing I_sapa , PD, PRF, and N_cycles . When evaluating each parameter at f = 1.1 MHz, it was observed that NSRs close to 100% were achieved when sufficiently elevating their respective values. When computing the NSR as a function of the spatial-average, temporal-average intensity (I_sata ), defined as the product of PRF, PD, and I_sapa , a significant elevation of the NSR from 0% to close to 100% was measured by increasing I_sata from values approximate to 4 W/cm² to values higher than 12 W/cm² . No clear and consistent trend was observed in trials aimed at exploring the effects of different levels of stable and inertial acoustic cavitation on the NSR. Finally, the feasibility of inducing neural responses with 3.3 MHz repetitive-pulse FUS stimuli was also demonstrated with NSRs reaching up to 60%, in the range of FUS parameters studied.

CONCLUSION

The time-averaged value of the radiation force per unit volume of tissue is proportional to the acoustic intensity. As a result, the observations from this study suggest that the neural structure responding to the stimulus is sensitive to the mean radiation force carried by the FUS sequence, regardless of the combination of FUS parameters giving rise to such force. The results from this study further revealed the existence of a minimal activation threshold with regard to I_sapa .

Can ionic concentration changes due to mechanical deformation be responsible for the neurostimulation caused by focused ultrasound? A simulation study

Objective.Ultrasound stimulation is an emerging neuromodulation technique, for which the exact mechanism of action is still unknown. Despite the number of hypotheses such as mechanosensitive ion channels and intermembrane cavitation, they fail to explain all of the observed experimental effects. Here we are investigating the ionic concentration change as a prime mechanism for the neurostimulation by the ultrasound.Approach.We derive the direct analytical relationship between the mechanical deformations in the tissue and the electric boundary conditions for the cable theory equations and solve them for two types of neuronal axon models: Hodgkin-Huxley and C-fibre. We detect the activation thresholds for a variety of ultrasound stimulation cases including continuous and pulsed ultrasound and estimate the mechanical deformations required for reaching the thresholds and generating action potentials (APs).Main results.We note that the proposed mechanism strongly depends on the mechanical properties of the neural tissues, which at the moment cannot be located in literature with the required certainty. We conclude that given certain common linear assumptions, this mechanism alone cannot cause significant effects and be responsible for neurostimulation. However, we also conclude that if the lower estimation of mechanical properties of neural tissues in literature is true, or if the normal cavitation occurs during the ultrasound stimulation, the proposed mechanism can be a prime cause for the generation of APs.Significance.The approach allows prediction and modelling of most observed experimental effects, including the probabilistic ones, without the need for any extra physical effects or additional parameters.

An Integrated 2D Ultrasound Phased Array Transmitter in CMOS With Pixel Pitch-Matched Beamforming

Emerging non-imaging ultrasound applications, such as ultrasonic wireless power delivery to implantable devices and ultrasound neuromodulation, require wearable form factors, millisecond-range pulse durations and focal spot diameters approaching 100 μm with electronic control of its three-dimensional location. None of these are compatible with typical handheld linear array ultrasound imaging probes. In this work, we present a 4 mm × 5 mm 2D ultrasound phased array transmitter with integrated piezoelectric ultrasound transducers on complementary metal-oxide-semiconductor (CMOS) integrated circuits, featuring pixel-level pitch-matched transmit beamforming circuits which support arbitrary pulse duration. Our direct integration method enabled up to 10 MHz ultrasound arrays in a patch form-factor, leading to focal spot diameter of ∼200 μm, while pixel pitch-matched beamforming allowed for precise three-dimensional positioning of the ultrasound focal spot. Our device has the potential to provide a high-spatial resolution and wearable interface to both powering of highly-miniaturized implantable devices and ultrasound neuromodulation.

Spatio-temporal characterization of causal electrophysiological activity stimulated by single pulse Focused Ultrasound: an ex vivo study on hippocampal brain slices

OBJECTIVE

The brain operates via generation, transmission and integration of neuronal signals and most neurological disorders are related to perturbation of these processes. Neurostimulation by Focused Ultrasound (FUS) is a promising technology with potential to rival other clinically-used techniques for the investigation of brain function and treatment of numerous neurological diseases. The purpose of this study was to characterize spatial and temporal aspects of causal electrophysiological signals directly stimulated by short, single pulses of focused ultrasound (FUS) on ex vivo mouse hippocampal brain slices.

APPROACH

MicroElectrode Arrays (MEA) are used to study the spatio-temporal dynamics of extracellular neuronal activities both at the single neuron and neural networks scales. Hence, MEAs provide an excellent platform for characterization of electrical activity generated, modulated and transmitted in response to FUS exposure. In this study, a novel mixed FUS/MEA platform was designed for the spatio-temporal description of the causal responses generated by single 1.78 MHz FUS pulses in ex vivo mouse hippocampal brain slices.

MAIN RESULTS

Our results show that FUS pulses can generate local field potentials (LFPs), sustained by synchronized neuronal post-synaptic potentials, and reproducing network activities. LFPs induced by FUS stimulation were found to be repeatable to consecutive FUS pulses though exhibiting a wide range of amplitudes (50 - 600 µV), durations (20 - 200 ms), and response delays (10 - 60 ms). Moreover, LFPs were spread across the hippocampal slice following single FUS pulses thus demonstrating that FUS may be capable of stimulating different neural structures within the hippocampus.

SIGNIFICANCE

Current knowledge on neurostimulation by ultrasound describes neuronal activity generated by trains of repetitive ultrasound pulses. This novel study details the causal neural responses produced by single-pulse FUS neurostimulation while illustrating the distribution and propagation properties of this neural activity along major neural pathways of the hippocampus.