Localization of Ultrasound-Induced In Vivo Neurostimulation in the Mouse Model

Abstract

Developments in the use of ultrasound to stimulate and modulate neural activity have raised the possibility of using ultrasound as a new investigative and therapeutic tool in brain research. Although the phenomenon of ultrasound-induced neurostimulation has a long history dating back many decades, until now there has been little evidence of a clearly localized effect in the brain, a necessary requirement for the technique to become genuinely useful. Here we report clearly distinguishable effects in sonicating rostral and caudal regions of the mouse motor cortex. Motor responses measured by normalized electromyography in the neck and tail regions changed significantly when sonicating the two different areas of motor cortex. Response latencies varied significantly according to sonication location, suggesting that different neural circuits are activated depending on the precise focus of the ultrasound beam. Taken together, our findings present good evidence of the ability to target selective parts of the motor cortex with ultrasound neurostimulation in the mouse, an advance that should help to set the stage for developing new applications in larger animal models, including humans.

Introduction

In recent years, the idea of directly modulating neural activity in brain using ultrasound has triggered considerable scientific interest. Such direct modulation, if applied with sufficiently high temporal and spatial resolution, could make possible new treatments for neurologic and psychiatric diseases, and could provide powerful new tools for investigating mechanisms of the human brain from sensory processing to the mysteries of cognitive function. Current techniques for inducing neural modulation, which include pharmacologic treatments, electrical stimulation techniques such as deep brain stimulation and transcranial magnetic stimulation, generally have significant limitations. Pharmacologic agents usually need to be applied directly to the site of interest, which may require surgical intervention. Similarly, although deep brain stimulation has already been employed for the effective treatment of clinical conditions including Parkinson's disease (Lyons 2011), it also requires surgery. transcranial magnetic stimulation has the advantage of being non-invasive and is employed clinically in the treatment of acute depression (Fitzgerald et al. 2011), though it has relatively poor spatial resolution and is unable to focus on structures deep in the brain. Optogenetic techniques represent one of the latest acquisitions in the neuroscientist's toolkit, and although they can be applied with great precision, attempts to use them in a clinical setting will require invasive procedures (Lalumiere 2011). This puts ultrasound neuromodulation in a tantalizing position because, although being non-invasive, it could, with current array technology, reach deep brain structures with good spatial resolution (Clement et al., 2005, Marsac et al., 2012, Pernot et al., 2003). With recent developments, researchers have found that it is possible to perform brain surgery in humans non-invasively using magnetic resonance imaging-guided focused ultrasound (Martin et al., 2009, McDannold et al., 2010). By capitalizing on these techniques, ultrasound neuromodulation has the potential to become a powerful tool for localized neurostimulation. However, at the moment, ultrasound neuromodulation has yet to be used clinically, and its limitations and capabilities remain uncertain. Although evidence for the stimulating effects of ultrasound on neural tissues has been accumulated over many decades (e.g., Bystritsky et al., 2011, Fry et al., 1958, Gavrilov et al., 1996, Harvey, 1929, King et al., 2013, Tufail et al., 2010, Yoo et al., 2011), little is known about how it actually works, and it is only in the last few years that the possibility of using it to stimulate brain structures in living organisms has even begun to look feasible. Tufail et al. (2010) provided evidence of transcranial ultrasound activation in vivo using low-frequency ultrasound stimulation in the intact mouse, reporting activation of motor cortex with no apparent damage to the brain after repeated ultrasound stimulation experiments. More recently King et al. (2013) established effective driving parameters and conditions for achieving transcranial stimulation of the nervous system in vivo, using a similar mouse somatomotor response.

These developments have been encouraging, but for ultrasound neuromodulation to become a truly useful technique, compelling evidence of localized effects in the brain is needed. So far, such localization has proved difficult to establish, although there have been some hints. Using functional magnetic resonance imaging, Yoo et al. (2011) measured blood oxygen level-dependent contrast signals while sonicating rabbit motor cortex and reported a response that was localized to the sonicated hemisphere. To confirm that the effect was indeed a form of neurostimulation, they performed additional experiments while the animal was outside the scanner and obtained forepaw movements contralateral to the sonication site, though no quantifiable details were provided on the magnitude of any differences between contralateral and ipsilateral activity.

In a follow-up study, the same laboratory (Kim et al. 2012) used ultrasound to stimulate the abducens nerve of a rabbit. The localized effect was confirmed by the occurrence of ipsilateral eye movements while the contralateral eye remained unaffected, as would be expected for stimulation of the nerves directly rather than the motor cortex. Although the researchers did not quantify the effect, the ensuing lateralized abductive eye movements were, at least, indicative of nerve stimulation.

In this study, we wanted to see whether we could generate different patterns of muscle activity by sonicating different parts of the motor cortex. From published maps of the topographic organization of the motor cortex in mice obtained using electrical intracortical microstimulation, distinct areas are known to control different muscle groups. Specifically, the areas controlling the neck and tail regions of the mouse are located at the rostral and caudal regions of the motor cortex (Tennant et al. 2011) (see Fig. 1). With this in mind, we placed an ultrasound transducer at different positions along a rostral-caudal axis aligned to the midline of the brain and were able to produce robust behavioral differences in twitches elicited in the tail and neck regions. Although it was difficult to discern by eye the differences in the twitches elicited in the neck muscle, the differences in tail responses were quite obvious, with sonication over the caudal region producing markedly larger deflections in the tail response. Furthermore, measurements of the associated electromyography (EMG) responses in both regions revealed distinctly different patterns of responses that changed according to the position of the transducer. This differentiation in rostral/caudal effects closely matched what might be expected from the published maps of the motor cortex. We explain how the observed patterns of response provide good evidence of localized activations occurring within different parts of the motor cortex. Somewhat surprisingly, attempts to induce lateralized responses by moving the transducer from left to right over the motor cortex produced little change in muscle activity on each side of the animal.