Medical use of ultrasound imaging started in the early 20th century, and it was initially used to diagnose peripheral muscular disorders and rheumatoid arthritis or to guide neurosurgery or other therapeutics. A variety of mechanisms can explain the biological effects of therapeutic US (Miller et al., 2012). FUS has been utilized in medical ablation since the 1940s for its thermal effects, such as treating tumors, kidneys, essential tremors, and bladder. However, such effects may be irreversible (Ghanouni et al., 2015). FUS has been used for creating local anesthesia and revolutionized chronic pain management by blocking related neural functions through its thermal effects (di Biase et al., 2021). Some studies proposed non-thermal acoustic radiational effects of US as the working mechanism for low-intensity pFUS (Downs et al., 2018; Lee et al., 2020), while others proposed intramembrane cavitation as the working mechanism for general ultrasonic neuromodulation (Plaksin et al., 2016). Several hypotheses were proposed to explain the neuromodulation effects of FUS (Kamimura et al., 2020). Additionally, high-intensity US stimulation of rat sciatic nerves could induce inertia cavitation with strong acoustic forces (Lee et al., 2015a). Understanding these underlying working mechanisms would allow for more reliable and targeted interventions. This section summarizes the pre-clinical and clinical trials of both high and low intensities. Figure 2 indicates a gradual increase in interest in the neurostimulation effects of pFUS. Both animal and clinical studies have shown intriguing results for high-intensity peripheral pFUS in pain management. Additional clinical applications are still in the early stages and translational investigations are on the way.
2.1. Peripheral focused ultrasound for neuromodulation
Early studies in the 1950s demonstrated that the US could reversibly inhibit neuronal functions of the CNS in animals (Fry et al., 1958). Transcranial focused ultrasound (tFUS) has been shown to provide non-invasive neuromodulation of deep brain tissue in animals and humans over the past two decades, and it could work as a promising therapeutic tool (Tyler et al., 2008; Tufail et al., 2011). Compared to tFUS, the fundamental mechanisms of pFUS neuromodulation are only partially understood and even under debate, and its clinical applications are still in the experimental stage. Currently, the most common application of low-intensity pFUS is on the peripheral nerve fibers, sub-organs, and nerve endings, with varying stimulation effects depending on the specific parameter settings. The underlying mechanisms of pFUS are also remain diverse and still being studied. This subsection first discussed the metrics for assessing neural functions after pFUS. Afterwards, we summarized the development history and working mechanisms of pFUS applications in different target types. Finally, we provided a tentative prospect for the future directions of pFUS.
2.1.1. Metrics for assessing neural functions after pFUS
Quantitative evaluations of stimulation effects are essential for understanding the working mechanisms of pFUS. Although US is compatible with a wide range of neurophysiological measurement tools, it is challenging to capture immediate peripheral neuronal response to US stimulation due to the protective soft connective tissues surrounding mammalian peripheral nerve axons, making it difficult to complete electrophysiological measurement for individual nerve axons. In contrast, while neurons in the CNS are not protected by such tissues and can be reliably recorded using invasive electrodes placed within the skull (Kandel et al., 2012). Single-unit recording can measure temporal information of spike signals with dominant action potential propagation (Chen et al., 2017), but its application remained limited in pFUS applications due to technical difficulty. Only one study reported that pulsed pFUS increased peripheral nerve conduction velocity, as measured by single-unit recordings (Ilham et al., 2018). Compound action potentials (CAPs) are the summation of action potentials from multiple muscle fibers (Merletti and Farina, 2016). Previous studies have utilized CAPs as a primary neurophysiological metric for assessing peripheral neural activities after pFUS (Lee et al., 2015b, 2020; Wright et al., 2017; Downs et al., 2018). However, various parameters easily influence CAP properties, such as recording positions and electrode properties, making it sophisticated to evaluate peripheral nerve functions after pFUS accurately.
The biomedical community has investigated various methods to investigate the effects of pFUS on the peripheral nervous system. Early studies also explored indirect measures such as refractory periods, muscle force, sensory action potentials and image-based techniques. For instance, a study demonstrated that 250 kHz focused US could suppress rhythmic bladder contractions in ten anesthetized rats, with longer latencies and refractory periods than conventional electrical stimulation, indicating a potential outpatient treatment for overactive bladder (Casella et al., 2017). Electroencephalography (EEG) and functional magnetic resonance imaging (MRI) analysis also revealed that pulsed FUS on the periphery could elicit a representative somatosensory response, as evidenced by evoked potentials and blood oxygen level-dependent responses (Legon et al., 2012). Another study found that pFUS on the sciatic nerves could induce EMG responses similar to that of conventional electrical stimulation in mice without harming the nerves or surrounding regions (Downs et al., 2018). Most prior FUS targeting modalities required MRI or B-mode US imaging to guide the treatment (Wu et al., 2005; Weeks et al., 2012). Notably, B-mode US imaging provided safe, accurate, and convenient real-time monitoring and targeting of peripheral nerve disorders (Carroll and Simon, 2020). For example, a US image-guided focused US study exhibited safe and controllable modulation of evoked motor neuron activity in vivo with PRF of less than 40 Hz by facilitating motor neuronal response on electrically evoked potential, while PRF more than 100 Hz could induce inhibition effects with irreversible temperature elevation of above 15°C (Kim et al., 2020). It was shown that US efficacy depended on PRF or duty cycle in well-characterized nervous system of Caenorhabditis elegans nematodes (Kubanek et al., 2018). Additionally, US-based displacement imaging illustrated a correlation between acoustic radiational forces and motor activations in vivo peripheral nerves during 4 MHz pFUS in mice, with EMG-based CAPs increasing with interframe nerve displacement (Lee et al., 2020). Moreover, pFUS could induce representative neurogenic axon reflexes of peripheral nerves in the mouse sciatic nerve, as evidenced by distinct blood flow changes after pFUS (Kim M. G. et al., 2021).
In summary, metrics to assess neural functions after pFUS are intriguing and under development. Further research based on these metrics is needed to fully understand the underlying mechanisms of various pFUS applications.
2.1.2. Peripheral focused ultrasound stimulation on nerve fibers
In Harvey (1929), a pioneering study reported that ultrasonic stimuli could cause rhythmic contractions of quiescent ventricular muscles in frogs and turtles. Subsequently, peripheral US stimulation has been tested in vivo, in vitro, and ex vivo animal models. By increasing the level of brain-derived neurotropic factor, unfocused pulsed ultrasound could facilitate nerve regeneration in a rat sciatic nerve crushed injury model (Raso et al., 2005; Jiang et al., 2016; Ni et al., 2017). The effects of the pFUS can vary depending on the stimulation parameters and may involve either mechanical or thermal effects. pFUS has been shown to inhibit and promote neural activities in various mammalian and invertebrate nerves, such as frog sciatic nerves, crab leg nerves, rat vagus and tibial nerves, mouse sciatic nerve, and human median and sciatic nerves (Young and Henneman, 1961; Lele, 1963; Mihran et al., 1990; Tsui et al., 2005; Dickey et al., 2012; Juan et al., 2014; Casella et al., 2017; Downs et al., 2018). Spike activities and conduction velocities are important early-stage metrics indicating that mammalian nerves respond differently to US stimulation. It was observed that the smallest C fibers are the most responsive while the largest A-α fibers are the least sensitive, and that the US may achieve neuromodulation through electro-mechanical resonance properties (Legon et al., 2012).
The effects of US stimulation are affected by many factors, including the duration, intensity and frequency of the stimulation. Long pulses (5 s–5 min) can induce thermal effects or mechanical effects depending on the duty cycle of the US stimulation. In contrast, short pulses of US stimulation may induce changes in compound action potentials (CAPs) with either enhancement or suppression via mechanical effects (Mihran et al., 1990). According to the US Food and Drug Administration (FDA), the intensity of diagnosis US imaging should be within 0.72 W/cm2 in ISPTA, US stimulation of ISPTA below 1 W/cm2 is generally considered low-intensity, and transcranial focused ultrasound stimulation with such intensity could effectively introduce neuromodulation effects in both animals and humans (Tyler et al., 2018). However, whether US intensity < 1 W/cm2 could consistently modulate peripheral nerve functions in still controversial. A study reported that US intensity < 1 W/cm2 was insufficient for identifying significant neuromodulatory effects on the PNS in vitro (Ilham et al., 2018), while others showed US stimulation within FDA recommendations could elicit modulatory effects in the PNS (Kim et al., 2012; Downs et al., 2018; Riis and Kubanek, 2022). Regarding the working mechanism for pFUS, previous reviews summarized that pFUS with ISPTA between 1 and 200 W/cm2 is not intense enough to elicit temperature-driven neuromodulation, and acoustic radiation force is likely the dominant effect for pFUS with ISPTA less than 200 W/cm2 (Feng et al., 2019). Most studies on body-coupled peripheral nerve stimulation utilized acoustic frequencies ranging from 0.25 to 7 MHz, with a spatial resolution of roughly 3 mm to 0.1 mm. The selection of a spatial resolution depends on the stimulation location under the skin and the size of the stimulation targets. Furthermore, there was still no recognized standard for parameter settings. Descriptions of US intensities used in different studies include ISPPA, ISPTA, pressure and so forth, making it difficult to compare distinct stimulation settings. Further investigations are needed to determine the optimal US parameter space for safe and efficient US stimulation protocols.
A recent mouse ex vivo study found that high-intensity, milli-seconds FUS pulses (350–500 W/cm2) can evoke action potentials for both myelinated A fibers and unmyelinated C fibers in mouse dorsal root ganglia, with millisecond latencies compared with electrical stimulation, implying PIEZO2 ion channel-mediated mechanisms (Hoffman et al., 2022). The study manifested that transcutaneous FUS stimulates peripheral nerve activity by activating intrinsic mechano-transduction mechanisms in neurons, and the modulation performance varied in different types of neurons (Hoffman, 2019). Furthermore, decreased FUS-evoked action potentials at higher FUS stimulus intensities may account for the activation of potassium channels after FUS, subsequent study should also evaluate the hypothesis thoroughly. However, the modulation effects of FUS might be limited for peripheral neurons which are lacking mechanosensitive ion channels (Hoffman et al., 2022). A recent study showed that US could not excite an isolated pig sciatic nerve in vivo, but it could reliably inhibit nerve functions across a wide range of parameters through a thermal effect (Guo et al., 2022). These findings contradict previous research on non-isolated nerves. It might be accounted for the lack of mechanosensitive ion channels in pig axon membranes. Another explanation of Guo’s paper is that the mechanical environment of the neuron might change in the extracellular matrix, altering the response of those ion channels following pFUS. Higher pFUS intensities indicate more severe safety challenges, and the safe parameter space was not sufficiently explored for practical applications. Moreover, mathematical modelling could also play important role in understanding the modulation effects of US stimulation. Several models have been proposed to explain the US neuromodulatory effects regarding radiation force and neuronal action potentials, but no consensus has been reached (Plaksin et al., 2016; Lemaire et al., 2019).
Beyond previous efforts in animal models, pFUS studies were also conducted in human beings. Preliminary studies showed that pFUS could induce representative suppression of somatosensory evoked potentials in human subjects, as measured by EEG (Kamimura et al., 2019). Further, US imaging-guided 1.1 MHz pFUS could modulate median nerve functions and change thermal pain perception in healthy human subjects, implying potential FUS clinical applications in pain management (Lee et al., 2021). Later clinical trials in patients were indispensable to assess the functional role of pFUS in treating specific pain syndromes.
2.1.3. Peripheral focused ultrasound stimulation on sub-organs
In addition to sensorimotor functions of somatic PNS, US stimulation could also target autonomic PNS, such as the sub-organs which are innervated by nerve pathways. For instance, US stimulation prevented renal ischemia–reperfusion in mice by modulating of the splenic cholinergic anti-inflammatory pathway (Gigliotti et al., 2013). Further, spleen-targeted pFUS could modulate the cholinergic anti-inflammatory pathway and reduce cytokine (for example, TNF, IL-1β, and IL-6) response to endotoxin to the same levels as implanted vagus nerve electrical stimulation (Cotero et al., 2019a). Recent research has also demonstrated the effectiveness of spleen-focused US stimulation in the treatment of inflammatory diseases. Spleen-targeted pFUS could effectively improve the severity of arthritis in an arthritis mouse model by influencing CD8 + T cells (Hu et al., 2022). Single-cell RNA sequencing of splenocytes and experiments in genetically-immunodeficient mice models illustrated the functional role of T and B cells in the anti-inflammatory pathway (Zachs et al., 2019). In addition to the spleen, hepatic US stimulation could modulate pathways that regulate blood glucose and is as effective as vagus electrical stimulation in suppressing the hyperglycemic effect of endotoxin exposure (Cotero et al., 2019a). Several studies have manifested that selective activation of the hepatoportal nerve plexus via pFUS could improve glucose homoeostasis and glucose tolerance and utilization in diabetes rodent and swine models (Huerta et al., 2021; Cotero et al., 2022).
Beyond previous efforts, additional explorations are required to elucidate the modulation mechanism in human subjects and translate these state of art techniques into clinical practice for inflammatory and metabolic disorders. pFUS has the potential to be a promising bioelectronic medicine method as an alternative to conventional electrical stimulation modalities.
2.1.4. Peripheral focused ultrasound stimulation on nerve endings
Sensorimotor functions can be encoded by neurosensory receptors and nerve endings beneath the epidermis. Unlike most pFUS studies on nerve fibers or sub-organs, which were in the pre-clinical stage, the US applications in peripheral nerve endings have been successfully conducted in human subjects. As early as 50 years ago, classical electrophysiological experiments in animals and humans showed that 1–3 MHz pFUS could activate skin mechanoreceptors and induce tactile sensations. Meanwhile, the stimulation of nerve mechanoreceptors such as Pacinian corpuscles can generate temperature and pain sensations (Gavrilov et al., 1976). Several pioneering studies by Soviet Union scientists have demonstrated that pFUS of short duration and relatively high intensity can induce various somatic sensations such as tactile, thermal, hearing, and pain without adjunct harm to surrounding tissues. The modulation effects might attribute to the mechanical effects of US stimulus (Gavrilov et al., 1996). The US intensity (1–200 W/cm2) can activate low-threshold mechanoreceptors in the skin, muscles, joints, and other typical components. As the stimulus intensity increases, the sensations progress from tactile to thermal and finally to pain perception. These sensations can be elicited by pulsed mode, amplitude-modulation or pulse-amplitude modulated pFUS. However, continuous US stimulation failed to elicit tactile sensations like pulsed stimulation. The most representative thermal sensations were found to occur at an intensity range of 10–30 W/cm2 (ISPTA), while vibrotactile and nociception could happen in the range of up to 100 W/cm2 (ISPTA; Lee et al., 2014). Furthermore, US on acupuncture points (LI4) can elicit deqi sensations, implying its potential medical applications (Yoo et al., 2014).
Moreover, a subsequent study manifested that the sensations induced by pFUS were related to the density of mechanoreceptors, and different neuroreceptors respond differently to the same US dosage (Dickey et al., 2012). It could explain that the sensations induced by pFUS differ in different parts of the human body and the sensations of fingers require lower US stimulation intensity than that of palms with the same stimulation frequency. Further, sensations appear to be more effectively triggered at lower frequencies of US stimulation (Gavrilov, 2016), which might account for increased radiational force and neuronal displacements after lower frequency US, and mechanoreceptors and ion channels could be better modulated (Coste et al., 2012). An underwater human study also determined that a 300 kHz stimulus is more effective than a high 900 kHz stimulus in stimulating excitable mechanoreceptors and nerve fibers in the human PNS, providing transducer selection guidelines for human peripheral nerve stimulation (Riis and Kubanek, 2022). The study excluded the effects of heating and cavitation from their experimental results, and future studies are required to determine the functional role of US in peripheral neural receptors.
Peripheral focused ultrasound stimulation on peripheral nerve endings could diagnose various neurological, dermatological and hearing disorders (Gavrilov and Tsirulnikov, 2012). For example, neurological patients have higher tactile thresholds after pFUS than healthy controls, and this pattern could be used to diagnose of neurological diseases (Gavrilov, 1984). Similarly, pFUS could diagnose hearing disorders by comparing standard threshold ultrasonic audiograms in healthy controls to audiograms measured by pFUS in hearing diseases (Tsirulnikov et al., 1988). However, the potential risks of prolonged US exposure are non-negligible and worth further explorations. Due to high-frequency US transmission problems in the air and inherent safety challenges for surrounding organs, pFUS applications on nerve endings remain limited. Furthermore, optimized and appropriate stimulation settings for diverse applications are not well characterized for diverse applications, and a wide range of acoustic parameters may influence peripheral neuroreceptors’ responses after stimulation.
2.1.5. Future directions of pFUS
To further improve the neuromodulatory effects of pFUS and facilitate practical applications, the transducer designs, measurement, and guidance tools are among the primary technical challenges. Most previous studies utilized single-element transducers with a fixed focal point, poor spatial resolution, limited transducer bandwidth, and no image guidance (Gao et al., 2019; Luo et al., 2020). Current US stimulation systems cannot satisfy the all-around requirements of stimulation targets. To address these limitations, the development of dual-mode 2D transducer arrays allow for simultaneous recording and neuromodulation (Zhang et al., 2021), and multifrequency transducers may also contribute to more practical pFUS applications (Constans et al., 2017). Additionally, acoustic metamaterials can manipulate and control sound waves more effectively than conventional materials, allowing for more focused and reliable sound manipulations (Li et al., 2014; Cummer et al., 2016). Next-generation US transducers, such as capacitive or piezoelectric micromachined ultrasonic transducers, could be manufactured with flexible and wearable materials to produce broadband transducers (Khuri-Yakub and Oralkan, 2011). However, more efforts are necessary to evaluate these new peripheral ultrasonic neuromodulation techniques in real-world applications. Looking forward, we envision that US stimulation has the potential to serve as a new type of bioelectronic medicine, offering an alternative to traditional pharmaceutical treatments (Cotero et al., 2019b).
2.2. High-intensity pFUS for pain management
Pain is an unpleasant sensory and emotional experience associated with, or resembling that associated with actual or potential tissue damage (Walker et al., 2021). According to the International Association for the Study of Pain (IASP), chronic pain is defined as pain lasting longer than 3 months (Ashburn and Staats, 1999). Notably, chronic pain could be categorized into three types of damage or disease: nociceptive pain of direct tissue, neuropathic pain of somatosensory system, or mixed pain of both. When conventional treatments fail to relieve painful symptoms, temporary or permanent disruptions of functional pathways may be necessary for pain relief. Compared to other minimally-invasive or invasive nerve block techniques, HIFU-based pain management could reduce the risk of invasive surgery due to its non-invasive nature. HIFU could be targeted on both the central and peripheral nervous systems. Unlike central stimulation, which targets on specific neural pathways or regions in the brain or spinal cord, peripheral HIFU delivers local heating on peripheral nerves or regions via either ablation or reversible inhibition effects. Temperature elevation is the dominant mechanism of nerve conduction inhibition by HIFU (Lele, 1963), and non-thermal effects are not clear (Colucci et al., 2009). Other studies also reported that HIFU can induce mechanical effects on the giant axon of live earthworms (Wahab et al., 2012).
The early FUS investigations in the 1960s indicated that FUS effects were temperature dependent, with minor temperature differences between ablation and reversible effects (Lele, 1963). FUS was found to modulate C fibers without affecting A fibers, and reversible effects on conduction were observed after 0.4–1.0 s of US exposure in ex vivo frog sciatic nerves (Young and Henneman, 1961). Another study revealed that FUS stimulus on the optic nerve induced partial or total inhibition of visually evoked potentials for 4–5 min (Adrianov et al., 1984). Additionally, 30 s continuous wave sonication of HIFU could inhibit nerve conduction and provide a complete and temporary conduction block of normal bullfrog sciatic nerves (Colucci et al., 2009). In contrast, US-imaging-guided peripheral HIFU can only partially and temporarily block the CAPs of sciatic nerves in rat plantar foot muscles, and histological evidence of axonal demyelination and necrosis of Schwann cells or axons lesioning was reported (Foley et al., 2007, 2008). A recent study showed that some CAP parameters are altered similarly by HIFU and local anesthetics in an ex vivo rat sciatic nerve model, with minor but significant differences (Anderson et al., 2022). Overall, FUS has been studied for its potential in pain management, but more research is needed to fully understand the effects and applications of HIFU in pain management.
Most prior pre-clinical studies have stimulated sensory and motor nerves together, and thus it is necessary to distinguish the functional roles of high-intensity pFUS in motor or sensory nerves separately. In vitro experiments showed that CAPs of the sciatic nerve and sensory action potentials of the sural nerve were temporarily and incompletely blocked by HIFU and fully recovered after pFUS with appropriate stimulation parameters in normal and diabetic rats. The blocking effects lasted 10 to 30 min. These results indicated that high-intensity pFUS could safely and reversibly suppress nerve conduction in diabetic rats for analgesic applications, implying application in blocking sensory nerves reversibly and providing peripheral pain relief in diabetic rats (Lee et al., 2015b). Earlier research examined the relationship between thermal doses and changes in peripheral nerve histology and neural functions (Vujaskovic et al., 1994), and it appears that both safety and intervention efficiency can be achieved. Patients with vocal cord paresis after single-session HIFU of thyroid nodules could recover within 6 weeks, suggesting that thermal power did not induce irreversible nerve damage in these cases (Lang et al., 2017). Further, medial branch nerves could be directly ablated with MRgHIFU in swine models, effective nerve thermal necrosis was created without damaging adjacent tissues (Kaye et al., 2016). Similarly, HIFU treatment on the occipital nerves was investigated in a validated rodent headache model of chronic migraine, pulsed HIFU but not ablative HIFU enhanced mechanical thresholds post-therapy like sumatriptan at day three after treatment in the periorbital region, suggesting the potential of pulsed pFUS in treating migraine (Walling et al., 2018).
To guarantee safety during high-intensity pFUS treatment, it is necessary to monitor changes in peripheral nerves during treatment. While there was minimal evidence that HIFU-induced heat induces adverse effects of on peripheral structures, US guidance allows for target-specific thermal energy deposition to the peripheral nerve, making US-guided HIFU for nerve ablation possible (Worthington et al., 2016). Furthermore, MRI thermometry could non-invasively detect body temperature and monitor organic changes following pFUS (Jiang et al., 2020). In a cadaveric and laboratory feasibility study, MRI-guided FUS (MRgHIFU) illustrated temperature elevations of the trigeminal nerve using a gradient echo-sequence (Monteith et al., 2013). Another pilot pig study showed that MRgHIFU with 3D MR neurography guidance could be used for targeted peripheral nerve ablation, such a system could be applicable for post-treatment thermal tracking without contrast injection (Huisman et al., 2015a). Diffusion-weighted imaging and tractography can also effectively visualize target peripheral nerve segments and assess the microstructural changes after MRgHIFU (Walker et al., 2021). Nevertheless, the spatial resolution of imaging techniques such as US and MRI may still be limited for peripheral nerves such as sciatic nerve with a millimeter lever or even smaller, highlighting the need for future advancements in this area.
Though the underlying mechanisms of pFUS-based pain management have not been thoroughly investigated, preliminary clinical applications have been already performed in typical pain symptoms. With CE and FDA approval, pFUS has been applied in human pain relief for uterine fibroids, bone metastases (Huisman et al., 2015b), and low back pain due to facet joint osteoarthritis (di Biase et al., 2021). Furthermore, pFUS could alleviate pain in cancer patients by inducing tissue denervation, tumor mass reduction, and neuromodulation, all of which can influence pain-related neural pathways (Dababou et al., 2018). It outperforms conventional analgesic therapies because of its non-invasiveness, rapid pain control, safe repetition, and easy combination with chemotherapy or radiation therapy. A phase I single-arm clinical trial indicated that HIFU on the nerve terminals of the facet joints could reduce pain and improve functional abilities in patients with facet joint arthritis (Weeks et al., 2012). Phantom limb pain was also a significant concern, potentially influencing the life quality of almost half of the amputee patients. An exploratory case study exhibited MRgHIFU-mediated ablation of stump neuromas could reduce pain intensity in patients with postamputation neuropathic pain (Nachtigal et al., 2022). The cause of such postamputation neuropathic pain is complex, stemming from both the CNS, the spinal cord and peripheral nerves. Therefore, more research is required regarding the mechanism, safety and long-term efficacy. Further randomized controlled clinical trials with large sample sizes measures are necessary to determine the clinical efficacy of the latest HIFU-based pain management strategies.