Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 1.
Published in final edited form as: Brain Stimul. 2021 Mar 24;14(3):598–606. doi: 10.1016/j.brs.2021.03.004

Entrainment of Cerebellar Purkinje Cell Spiking Activity Using Pulsed Ultrasonic Stimulation

Ahmet S Asan 1, Qi Kang 1, Omer Oralkan 2, Mesut Sahin 1
PMCID: PMC8164992  NIHMSID: NIHMS1686923  PMID: 33774207

Abstract

Background:

Focused ultrasound (FUS) has excellent characteristics over other non-invasive stimulation methods in terms of spatial resolution and steering capability of the target. FUS has not been tested in the cerebellar cortex and cellular effects of FUS are not fully understood.

Objective/Hypothesis:

To investigate how the activity of cerebellar Purkinje cells (PCs) is modulated by FUS with varying pulse durations and pulse repetition frequencies.

Methods:

A glass microelectrode was inserted into the cerebellar vermis lobule 6 from the dorsal side to extracellularly record single unit activity of the PCs in anesthetized rats. Ultrasonic stimulation (500 kHz) was applied through a coupling cone, filled with degassed water, from the posterior side to target the recording area with varying pulse durations and frequencies.

Results:

Simple spike (SS) activity of PCs was entrained by the FUS pattern where the probability of spike occurrences peaked at around 1 ms following the onset of the stimulus regardless of its duration (0.5, 1, or 2 ms). The level of entrainment was stronger with shorter pulse durations at 50-Hz pulse repetition frequency (PRF), however, peri-event histograms spread wider and the peaks delayed slightly at 100-Hz PRF, suggesting involvement of a long-lasting inhibitory mechanism. There was no significant difference between the average firing rates in the baseline and stimulation periods.

Conclusion:

FUS can entrain spiking activity of single cells on a spike-by-spike basis as demonstrated here in the rat cerebellar cortex. The observed modulation potentially results from the aggregate of excitatory and inhibitory effects of FUS on the entire cortical network rather than on the PCs alone.

I. INTRODUCTION

Energy in the form of acoustic (1), magnetic (2), optical waves (3) and electric fields (4) can stimulate neurons or modulate their activity. Among these stimulation paradigms, ultrasonic stimulation comes into prominence due to its capability of providing high spatial resolution at deeper brain regions without disturbing the superficial and neighboring areas (5). The use of high intensity focused ultrasound (HIFU) goes back to early 20th century (6), whereas recent studies have focused on low intensity focused ultrasound (LIFU) and shown its modulatory effects on neural activity at the network level (7). In 2008, Tyler et al. reported that LIFU reversibly modulated the activity in hippocampal slice cultures (1). Studies in rats (8), mice (9), and rabbits (10) showed muscle contractions when LIFU is applied to the motor cortex. FUS has also been demonstrated to modulate neural activity in human subjects (11–14). As one of the first human studies, Legon and colleagues reported that sensory response evoked by median nerve stimulation are reduced by FUS, and also showed that targeting the somatosensory cortex with FUS improves the performance in the sensory discrimination task (12).

The need for non-invasive focal stimulation as a tool to modulate brain activity warrants further studies aimed at elucidating the mechanism of FUS, which are scarce in the literature. One possible explanation for the observed effects is that ultrasound waves activate the mechanosensitive ion-channels and alter the membrane potential (1). Also, radiation force generated by FUS is thought to induce changes in membrane capacitance and thereby manipulate membrane currents (15, 16). In addition, Oh et al. showed that FUS opens the Ca+2 channels of astrocytes and causes them to release glutamate into the extracellular space, which results in an increase in the overall excitability of the neural circuits (17).

The cerebellum is an ideal target to investigate the underlying mechanism of neuromodulation methods due to its unique neural circuitry that is stereotypically repeated across its cortex with only a few different cellular subtypes. The outgoing axons from the cerebellum that originate in the cerebellar nuclei (CN) have direct and indirect connections with many brain regions, in agreement with the diversity of the functions in which the cerebellum is involved, including motor control and cognitive functions (18–20). The Purkinje cells (PCs), located near the cortical surface, are responsible for sculpting the activity in the CN. This allows modulation of the cerebellar outputs by stimulation of the cerebellar cortex, a more superficial and easier target in contrast to the CN. In this respect, the cerebellar cortex offers an opportunity for investigation of FUS effects on neural circuits that indirectly modulate a whole host of brain function.

Ultrasonic stimulation pattern plays a critical role determining the modulatory effects of LIFU. However, there is a gap in our knowledge regarding differential modulatory effects of stimulation parameters on local neural circuits. In this study, we mainly focused on the pulse duration (PD) and pulse repetition frequency (PRF) as the critical parameters that define the FUS pattern and looked into how these parameters affect the spike timing of the PCs in the cerebellar cortex. Our results showed that LIFU does not change the overall firing rate, but the timing of the PC simple spikes was locked to the stimulus cycle. The LIFU pattern entrained the spike activity of the PCs and the level of the entrainment was stronger with the shortest PD tested. This study presents unprecedented data on how LIFU affects the activity of a well-studied neural circuitry at the single cell level (PCs), and the importance of the pulse duration on entrainment of neural spikes.

II. METHODS

Animal Preparation

Six male Sprague Dawley rats (320–350 g) were used in this study. All procedures were approved and performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC), Rutgers University, Newark, NJ. Animals were initially anesthetized with the ketamine/xylazine mixture (100 mg/kg and 12 mg/kg, IP) and placed in a stereotaxic frame. The body temperature was measured by a rectal probe and regulated with a heating pad placed underneath the animal. Additional doses of ketamine were injected (25 mg/kg) as needed during the course of surgery. The blood oxygen level, monitored with a pulse oximeter probe attached to the hind paw, was made sure to stay above 90% during data collection.

An incision was made to the skin to remove muscle tissue and open a craniotomy hole over the cerebellum. Then, the skull over the lobules 6–10 in the posterior part of the cerebellum was completely removed. A small hole was made into the dura mater over the lobule 6 or 7 with the tip of a 31-g needle. A glass micropipette electrode (3–5 MΩ) filled with normal saline was slowly inserted for extracellular recording through the dura hole from the top into the cortex with the help of a 10-μm step-size micromanipulator. A Ag/AgCl wire was attached atop the skull near the frontal sinus as a recording reference electrode.

The neural recordings were performed in a large Faraday cage through a physiological amplifier (Model 1700, modified for high-input impedance, A-M Systems, Sequim, WA) with filter setting at 100 Hz – 5 kHz and a gain of 1,000 or 10,000, depending on the spike amplitudes. Neural signals were acquired onto the computer at 100-kHz sampling rate (PCI-6071, National Instruments, Austin, TX), while simultaneously being monitored on the oscilloscope and listened through an audio speaker. PCs were identified by their characteristic complex spikes and signal-to-noise ratio was maximized by fine adjustment of the electrode position.

The ultrasonic stimulation was applied using a 500-kHz immersion-type ultrasonic transducer (V-301-SU, focal length = 31.75 mm from the transducer rim, diam. = 31 mm, Olympus Scientific Solutions Americas Inc., Waltham, MA) at the posterior side of the cerebellum (Fig. 1A). In order to focus the output of the ultrasound probe into a smaller beam, a 3D printed cone with 30-mm height and 3-mm tip diameter was attached and filled with degassed double-distilled water. The tip of the cone was covered with a thin (~12 μm) plastic film and positioned about 1.75 mm away from the cerebellar surface. The gap between the cone and the cerebellar cortex was filled with ultrasound gel (Aquagel, Aquasonic, MI) as a coupling medium, as well as for keeping the cerebellar surface moist.

Figure 1.

Figure 1.

Open in a new tab

A) Schematic view of the placement of the recording electrode and ultrasound probe (50). B) A sample PC recording. The complex spike is marked with a red asterisk. C) The pulse train generated on the computer to control the ultrasonic stimulation pattern. Each train of pulses in the top figure is 500 ms long separated by 500 ms off-periods. PRF: Pulse repletion frequency.

Transducer Calibration

The acoustic field generated by the transducer was characterized in a water tank using a calibrated hydrophone (HNA-0400, Onda Corporation, Sunnyvale, CA) coupled to a calibrated preamplifier (AH-2020–100, Onda Corporation, Sunnyvale, CA) to determine the effect of the cone on the acoustic field map and the peak pressure. To fully map the acoustic field, the hydrophone was scanned using a computer-controlled three-axis linear stage (Model PRO165, Aerotech Inc., Pittsburgh, PA). The axial plane at the transducer center was mapped with and without the cone attached. Furthermore, the horizontal planes at the focal point (z=0) and 2.5 mm beyond were mapped with the coupling cone attached. The acoustic field generated on these planes was measured at 500 kHz for mapping. As the minimum frequency, for which the hydrophone was calibrated, was 1 MHz, the absolute pressure generated at 500 kHz was calculated using the one-way frequency response of the transducer and calibrated measurements at 1 MHz.

Pattern Generation

The ultrasonic stimulation pattern was generated in Matlab (MathWorks, Natick, MA) and sent out in real time through a multifunction data acquisition board (PCI-6071, NI) from the computer to a function generator (Model 33120A, Agilent) to turn its sinusoidal output on and off at the desired pattern as shown in Fig. 1C. The carrier frequency of the sinusoidal signal was set to 500 kHz, corresponding to the peak efficiency of the transducer. The sinusoidal signal amplitude applied to the input of a linear power amplifier (Model 240L, 40 W, Power Gain=50 dB, E&I Ltd., Rochester, NY) that drives the ultrasonic transducer was manually adjusted between 100 mVpeak to 500 mVpeak, corresponding to 35 Vpeak and 175 Vpeak at the transducer. Stimulation intensity was started at a relatively low level, slowly increased until observing an appreciable level of entrainment in the spike activity and kept at the same intensity for all trials in the same cell. Thus, the best level of peak pressure was decided independently in each PC based on the modulation level. The pulse duration was adjusted to 0.5, 1, or 2 ms in sequential trials.

Data Collection and Analysis

Ten-second-long neural recordings were collected and each second of recording contained 500 ms of ultrasound stimulation followed by 500 ms of no-stimulation period. The PRF was set to either 50 Hz or 100 Hz to make it close to the spontaneous firing rates of PCs. Upon completion of the first set of neural activity, recording electrode was repositioned to obtain another set of recording from a different PC in some animals. Though there are multiple stimulation parameters that potentially can have an impact on the modulation, we focused on investigating the effects of PD and PRF.

Single unit activities from a total of 18 PCs (10 at 50-Hz PRF and 8 others at 100 Hz) were recorded in 6 animals. PCs were identified by their characteristic complex spikes (Fig. 1B) and simple spikes were detected using a peak-detection algorithm with a threshold value determined based on the signal-to-noise-ratio. The recordings were divided into no-stimulation and stimulation periods. To find the mean firing rate in these two periods, the number of spikes occurred in each time interval was divided by duration. To measure the level of entrainment, peri-event histograms of simple spikes were constructed with respect to the timing of the ultrasound pulses, and entropy and coefficient of variation (CV) were computed as metrics of modulation. We computed Shannon’s entropy and CV values in the 10 or 20-ms (depending on PRF) inter-pulse intervals during the stimulated and baseline periods in each second. The entropy was defined as E=−∑i=1Npilog(pi), where pi are the normalized probabilities at discrete time points along the inter-pulse interval.

III. RESULTS

Transducer Calibration

The peak pressure per unit excitation voltage at the focal spot without the cone was found to be 2.6 kPa/V. With the cone, the total emitted energy from the transducer surface was confined to a smaller area (Figs. 2E and 2F) and declined quickly beyond the focal point in the axial direction (Fig. 2D). The peak pressure per unit excitation voltage at the focal spot with the coupling cone attachment was measured as 3.94 kPa/V. Furthermore, the cone opening acts as a secondary aperture and a near-field diffraction pattern is observed on planes that is beyond the focal plane (Fig. 2G). At 2.5 mm away from the focal plane (Fig. 2G), the peak pressure was measured as 2.5 kPa/V and the peak pressure amplitudes were estimated to be between 88 kPa and 438 kPa for the transducer voltages of 35 Vpp and 175 Vpp respectively, with corresponding spatial-peak-pulse-average intensity (I_SPPA) levels of 0.26 W/cm2 and 6.4 W/cm2. Because the exact location of the recorded PCs within the pressure field generated could not be determined, these values should be considered as the upper bounds of the applied acoustic pressure and power.

Figure 2.

Figure 2.

Open in a new tab

Transducer pressure profiles measured in water tank. A and B show the areas mapped in the axial (z) direction with and without the coupling cone present. The same maps are replotted in C and D in an expanded scale and containing the focal spot (z=0). The bottom plots are the intensity profiles in the horizontal plane with no cone at z=0 (E), and with the cone at z=0 (F) and z=2.5mm (G). The first and second numbers on the color bar scales of the heatmaps indicate the pressures (in units of kPa) that correspond to peak transducer voltages of 35 Vp and 175 Vp, respectively. The white lines are the full width at half maximum (FWHM) contours. Note that the focal area in horizontal plane is much smaller with the cone present (F) than without (E), but the pressure amplitude diminishes quickly in the z direction and a near-field diffraction pattern is observed due to the opening at the cone tip (the halo in G).

Overall Entrainment of PCs with LIFU

Spontaneous PC activity presented fluctuations in the inter-spike interval (ISI) pattern. During application of ultrasonic pulses to the cerebellar cortex, ISIs became more regular and the PCs were synchronized to the ultrasound pulse train.

Figure 3 shows a sample PC recording during 10-s long LIFU application with a train of 0.5-ms stimulus durations at a repetition frequency of 50 Hz (top left and bottom) and 100 Hz (top right). Some pulses were skipped by the simple spikes at 50 Hz and more skipping occurred at 100 Hz (top right).

Figure 3.

Figure 3.

Open in a new tab

The lower row shows the simple spike activity of a PC during application of LIFU train, marked by the red trace. The top left row contains a short episode in an expanded time scale of the recording during 50 Hz stimulation. The top right is from another trial at 100 Hz PRF where some pulses were skipped by the neural spikes.

Peri-event histograms were constructed with respect to the timing of individual pulses (Fig. 4). Similar histograms made for non-stimulated periods in between the stimulation trains showed uniform distribution of the spike activity. During application of ultrasound pulse train, PC spikes synchronized to the stimulus pattern, indicated by a peak around 0.5 ms in the histogram, and followed by a few milliseconds of reduced probability of spike firing.

Figure 4.

Figure 4.

Open in a new tab

Peri-event histograms shown as a normalized probability distribution of spikes as a function of time during the baseline and stimulation periods in all 10 PCs at 50 Hz PRF. The red-dash line represents the 0.5-ms FUS duration in which the 500 kHz ultrasonic stimulation is applied. The red line in the baseline plot is for comparison only.

We also compared the firing rates during the stimulation and baseline periods to determine if LIFU increased the overall activity. Paired t-test results showed no significant difference between the average rates (baseline: 39.8±1.4, stimulation: 41.3±1.5 (mean±SEM), p> 0.32), suggesting that observed increase in the peri-event histogram results from redistribution of spike timings without an overall increase in the firing rate.

Effect of Pulse Duration on Entrainment

Various LIFU parameters may have differential effects on different compartments of the neural circuit and the net effect may be in the direction of excitation or inhibition (10, 21). Here, we primarily focused on the effects of stimulus PD and PRF on neural entrainment.

Fig. 5 provides an overview of how the probability of spike timings change with respect to the stimulus pulse at different PDs and PRFs. Spikes are uniformly distributed during the no-stim period whereas the activity is strongly entrained with the ultrasonic stimulation. The level of entrainment seems higher with 50-Hz PRF than 100 Hz at 0.5-ms pulse duration. However, the discrepancy between the spike probability distributions for the two PRFs decreases at 1-ms and 2-ms pulse durations. The column on the right end also provides a comparison between different PDs at 50-Hz and 100-Hz PRF separately. As a metric of entrainment, we used the ratio of the spike numbers during the first 3 ms for 2 ms PD, and 2 ms for 0.5 and 1 ms PDs over the stimulation interval. There was a decrease in spike ratio as the pulse duration was increasing at 50 Hz PRF (67±7%, 64±8%, and 59±7% for 0.5, 1, and 2 ms respectively), but the opposite trend was observed at 100 Hz (40±5%, 61±8%, and 69±8% (mean +/− sem)). As a measure of randomness, we computed the entropy of the spike timings within the stimulus cycle (1/PRF) (10 or 20 ms) during the stimulated and baseline periods. The average entropy had tendency to decrease at all PRFs and PDs compared to the entropy of the baseline from the same trials (p<0.05, paired t-test, two-sided), but did not reach a statistical significance in two cases (at 50 Hz with PD=2 ms, and at 100 Hz with PD = 0.5 ms, p>0.5 for both). These two cases presented weaker entrainment as indicated with smaller peaks in the peri-event histograms (Fig. 5) as well. The CV also decreased significantly (p<0.05, paired t-test, two-sided) in most cases of stimulation except at 50Hz with PD of 2ms (p=0.059). We also performed statistical analysis to measure if the probability distributions for different PDs and PRFs were statistically different. No significant difference was observed between any of the conditions (p>0.13, z-test, two-sided).

Figure 5.

Figure 5.

Open in a new tab

B) Peri-event histograms of unstimulated and stimulated periods at 50 Hz and 100 Hz pulse repetition frequencies and pulse durations of 0.5, 1, 2 ms. Plots are normalized by dividing the spike probabilities by their sum (so the sum is equal to 1). The shaded rectangles on the peri-event histograms specify the time frames that were used to show the spike distribution on an expanded time scale on the top (A) and right (C).

The 50-Hz and 100-Hz data did not show a significant difference in terms of the mean firing rate and spike distribution (p>0.05); therefore, we combined them together to strengthen the statistics and be able to see temporal trends in firing probability (Fig. 6). The probability of spike distribution has a peak around 1 ms regardless of the LIFU duration. However, the longer ultrasound pulse durations (1 ms and 2 ms) result in a wider spread of spike timings as shown in the averaged plots fitted by a Gaussian curve on the right. The high probability window, which lasted longer with longer stimulations, was followed by a reduced probability interval that seems to follow the falling edge of the ultrasound with similar delays in all three cases. Then, the spike probability recovers to that of the unstimulated periods. The comparison of the probability distributions showed a significant difference only between the 0.5ms vs 2ms (p>0.019, z-test, two-sided). Paired t-test results also showed no significant difference in the average firing rates between stimulated and unstimulated periods for any PD (p= 0.36, p= 0.9, p= 0.25 for 0.5, 1, and 2 ms, respectively).

Figure 6.

Figure 6.

Open in a new tab

Peri-event time spike distribution with 0.5, 1, and 2-ms ultrasonic pulse durations combining data at 50 Hz (n:10 cells) and 100 Hz (n:8 cells). Red dash line represents the applied pulse. The Gaussian distribution plot on the right shows the spike timings during the first 3 ms following the onset of ultrasound stimulation. The bin size is 100 μs.

IV. DISCUSSION

Neuro-modulatory effects of ultrasonic stimulation on various brain circuits have been reported in animal preparations. This study is the first of its kind showing the effect of LIFU on the spiking activity of a single cell. To the best of our knowledge, there is only one group that employed LIFU for cerebellar stimulation (22–25). The authors stimulated the lateral cerebellar nucleus in a stroke mouse model while recording the motor evoked potentials (23). The animals exposed to LIFU had significantly better somatosensory recovery than the control group. In another study by the same group, cerebellar LIFU restored the functional imbalance between the cerebral hemispheres caused by unilateral middle cerebral artery stroke (24).

We chose the cerebellum as a target brain site because of variety of functions that it is involved in, and the stereotypical organization of the neural circuits repeated across its cortex. Thus, the observed neural response to FUS at a given site in the cerebellar cortex will most likely generalize to other cerebellar cortical areas and modulate many cerebellar functions in similar fashion. In this study, we specifically investigated the effects of PD and PRF on the PC simple spike activity, because they were shown to be critical parameters in modulation of neuronal activity. PCs have a spontaneous firing rate around 45±20 Hz (26, 27) and the PRF was chosen in the same approximate range.

Acoustic Intensity

We adjusted the ultrasound intensity based on the cellular response. In our setup, a high intensity ultrasonic stimulation disturbed the electrode-cell interface mechanically and resulted in signal loss. Hence, the testing started with low intensities of ultrasonic stimulation and then slowly increased until detecting modulation in the PC simple spike firing pattern. It was not possible to sustain the signal for a long time from PCs closer to the cortical surface because the superficial layers of the cortex are more susceptible to mechanical disturbance, such as dimpling caused by penetration of the glass electrode. Hence, most of recordings were made from 2–3 mm depths inside the walls of the folia. It should also be noted that positioning the FUS probe for maximum modulatory effect was challenging, and not possible in some cases. A ‘sweet spot’ had to be found for the relative position of the transducer with respect to the recording electrode where the reported effects in this study could be observed. Thus, based on the transducer calibration plots, we can only specify a range for the acoustic intensities for the targeted sites in the tissue.

Direct vs. Indirect Pathway

Two recent studies pointed out an indirect pathway through the auditory system that can cause the reported neuro-modulatory effects (28, 29). These two studies used deafened animals as control and reported the absence of modulation with FUS in the cortical regions. In our study, we directly measure the simple spike activity of the PCs and observed that spikes are locked to FUS pulses within a millisecond window. The fast spike response to the FUS pulses reported in this study and the minimum time delay between the auditory stimulus and the cerebellum refutes the possibility of an indirect modulation pathway in our recordings. A direct evidence was also observed by moving the transducer away from the cerebellum by a few millimeters during recording while maintaining the acoustic coupling through the gel, which resulted in disappearance of neural entrainment since the targeted PC was no longer in the high-intensity acoustic field.

Mechanism of Modulation

FUS is divided into two categories based on the acoustic intensity level; this also defines the putative mechanism underlying neural modulation. High intensity ultrasound stimulation causes temperature elevation (30) and cavitation (31), which in turn damages neural tissue. Hence, it is not considered safe and mostly used for surgical applications such as tissue ablation (32). On the other hand, LIFU reversibly modulates the neural activity without causing any tissue damage.

Several hypotheses have been developed to explain the underlying mechanism of LIFU. Tyler et al. showed that LIFU modulate neurons by triggering the voltage gated ion channels (1). Even though large cavitations are not expected to occur with low intensity applications, Krasovitski et al. showed that LIFU generates nano-cavities within the cellular membrane, which alters the membrane capacitance and drives the cell toward its firing threshold (33). Yoo et al., on the other hand, have shown the absence of nano-cavity formation from LIFU, and ascribed the observed excitatory response to the mechanosensitive Ca+2 channels (34). Various studies have also reported modulatory effects of FUS on the extracellular neurotransmitter levels (17, 35, 36). Yang et al. observed a decrease on the gamma-aminobutyric acid (GABA) level when FUS is applied to the thalamus without a change on the glutamate level. Oh et al., on the other hand, demonstrated that LIFU stimulates Ca+2 channels in astrocytes resulting in glutamate release to the extracellular environment, thereby elevating overall excitability. Accumulation of neurotransmitters in the extracellular space can explain the observed lasting effects subsequent to FUS termination.

Although the modulatory effect of FUS is mostly attributed to the excitatory mechanisms (1, 17, 37, 38), some studies reported inhibitory effects of FUS in different cortical regions (10, 39–41). In many studies, the outcome variable is usually a muscle response or a change in behavior. The net effect of ultrasonic stimulation applied to an entire network of neurons can be different depending on the circuity. An excitatory effect can be mediated by either direct excitation of the targeted neuron, or disinhibition of it through suppression of the inhibitory neurons projecting to the targeted neuron. Interestingly, Menz et al., showed that the high frequency (43 MHz) FUS sonication modulates the activity in the retina by activating interneurons while leaving the ganglion cells unaffected (42). In this regard, investigation of the differential effects of various FUS parameters, such as pulse width, frequency, and train duration, on different neuronal subtypes may be critical for better understanding of the circuit level outcomes.

Some discrepancies exist in the published studies regarding pulse duration, which need to be resolved for better understanding of how FUS operates. King et al. demonstrated that continuous ultrasound is more effective than pulsed stimulation (21) while this finding was contradicted by others (43, 44). In a separate study, Plaksin et al. demonstrated that T-type voltage-gated calcium channels are more sensitive to short pulse widths (37) and proposed that selective stimulation can be achieved by targeting different ion-channels with different pulse widths. Considering that T-type channels are abundant in the cerebellar PCs (45), the proposed mechanism by Plaksin et al. agrees with our findings that entrainment is stronger with shorter pulses.

In this study we observed that short pulse duration is more effective for neural entrainment at 50 Hz. Shorter pulse durations may be more effective on PCs due to presence of the T-Type channels and this effect may be declining with longer pulses. However, if the inhibitory cells of the cerebellar cortex are also excited, rather than being suppressed, their inhibitory effect on PCs is expected to become stronger with longer pulses that carry more power. Since interneurons play a role in adjusting the activity pattern of PCs at a longer time scale (46), their stimulation can potentially cause wider spreading of the bell-shaped distribution of the spike timings. So, the shift and reduction of the peak probability with increasing FUS durations can be explained by the possibility that the inhibitory effects may last longer than the inter-pulse interval and the network does not have sufficient time to recover between the pulses. Alternatively, the PC itself may not be as excitable, after generating a spike, at subsequent pulses that come earlier than the interval dictated by the spontaneous firing rate. In both cases, the probability of firing is somewhat less than the baseline level (due to the previous pulse) at the time the next FUS pulse is delivered 10 or 20 ms later. The probability of firing is less, and the delay is longer with longer FUS pulses that contain more power.

Another possible explanation for widening of the distribution is that the probability of producing a simple spike may be reduced as a rebound effect from releasing the stimulation during the interval following FUS termination. When the ultrasonic stimulus is turned off early (at 0.5 ms), this sharpens the histogram by starting the lowered-probability-of-firing interval earlier. If the ultrasonic stimulation is kept on for 1 ms or 2 ms, the entrainment continues with heightened but decreasing firing probabilities for another millisecond after the onset of the ultrasonic stimulation. However, the rebound mechanism alone does not explain why the probability distribution has a lowered and delayed peak with increasing pulse durations assuming that the excitatory effects operate at shorter time scales than the inter-pulse interval (10–20 ms).

In conclusion, we cannot claim that ultrasonic stimulation directly stimulates the PCs based on the present data. The observed modulation in the PC simple spike pattern could be due to inhibition of several neuronal subtypes found in the cerebellar cortex (basket, stellate, and Golgi cells) that are inhibitory to the PCs. On the other hand, the immediate net effect is clearly excitatory as evidenced by strong entrainment of the simple spikes to the FUS pulses at millisecond scale. The fundamental difference between entrainment and supra-threshold neural stimulation is in the intensity of the excitatory effect. It is not clear if this excitatory effect would be able to drive the cell over the excitation threshold in the absence of spontaneous firing. This hypothesis can be tested by local injection of neurotransmitter blocking agents to eliminate synaptic connections as well as stopping spontaneous firing in future studies. Regardless of whether the observed modulation is a direct or indirect effect of FUS, the net effect being excitatory is an important finding considering that the PC is the sole output of the cerebellar cortex that project to the cerebellar nuclei. The fact that the entrainment occurs on a pulse-by-pulse basis, spikes locking to the stimulus pattern, is also encouraging for neuromodulation applications that require high temporal resolution. Simple spike entrainment may be more useful as a tool compared to supra-threshold stimulation in the cerebellum because PC simple spike synchrony has been thought of as a mechanism for modulating the nuclear cells (47–49).

V. CONCLUSIONS

Our study is the first of its kind showing the effect of FUS at a single cell level in the cerebellar cortex. Our results demonstrate that the simple spike activity of the PCs is effectively entrained by ultrasound on a pulse-by-pulse basis and locking to the stimulus pulse is stronger with shorter pulse durations. The cerebellar PCs could be entrained both at 50 Hz and 100 Hz ultrasound pulse repetition frequencies. A larger number of stimulus cycles were skipped at 100 Hz and there was no overall increase in firing frequencies at either 50 Hz or 100 Hz. The ultrasonic modulation of the cerebellar cortex may be a useful technique for indirect modulation of the cerebellar nuclear cells at millisecond temporal resolution.

  • Simple spike activity of the cerebellar Purkinje cells (PCs) are effectively entrained by the ultrasound stimulation pattern.

  • Entrainment of PC spikes reduces the randomness in PC spike timings (lower entropy), however, does not change the average firing rate.

  • The level of entrainment is highly dependent on ultrasound pulse duration (PD).

ACKNOWLEDGMENT

Supported by NIH/NEI Grant R21 EY028456. We thank Ali Onder Biliroglu for his help with transducer characterization.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

DECLARATIONS OF INTEREST: None

REFERENCES

  • 1.Tyler WJ, Tufail Y, Finsterwald M, Tauchmann ML, Olson EJ, Majestic C. Remote excitation of neuronal circuits using low-intensity, low-frequency ultrasound. PloS one. 2008;3(10):e3511–e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Hallett M Transcranial magnetic stimulation and the human brain. Nature. 2000;406(6792):147–50. [DOI] [PubMed] [Google Scholar]
  • 3.Thompson AC, Stoddart PR, Jansen ED. Optical Stimulation of Neurons. Curr Mol Imaging. 2014;3(2):162–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Nitsche MA, Paulus W. Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. J Physiol. 2000;527(3):633–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Rezayat E, Toostani IG. A Review on Brain Stimulation Using Low Intensity Focused Ultrasound. Basic Clin Neurosci. 2016;7(3):187–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Harvey EN. THE EFFECT OF HIGH FREQUENCY SOUND WAVES ON HEART MUSCLE AND OTHER IRRITABLE TISSUES. American Journal of Physiology-Legacy Content. 1929;91(1):284–90. [Google Scholar]
  • 7.Wang P, Zhang J, Yu J, Smith C, Feng W. Brain Modulatory Effects by Low-Intensity Transcranial Ultrasound Stimulation (TUS): A Systematic Review on Both Animal and Human Studies. Frontiers in Neuroscience. 2019;13(696). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.King RL, Brown JR, Pauly KB. Localization of Ultrasound-Induced In Vivo Neurostimulation in the Mouse Model. Ultrasound in Medicine and Biology. 2014;40(7):1512–22. [DOI] [PubMed] [Google Scholar]
  • 9.Tufail Y, Matyushov A, Baldwin N, Tauchmann ML, Georges J, Yoshihiro A, et al. Transcranial Pulsed Ultrasound Stimulates Intact Brain Circuits. Neuron. 2010;66(5):681–94. [DOI] [PubMed] [Google Scholar]
  • 10.Yoo S-S, Bystritsky A, Lee J-H, Zhang Y, Fischer K, Min B-K, et al. Focused ultrasound modulates region-specific brain activity. Neuroimage. 2011;56(3):1267–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Lee W, Kim H-C, Jung Y, Chung YA, Song I-U, Lee J-H, et al. Transcranial focused ultrasound stimulation of human primary visual cortex. Scientific reports. 2016;6:34026-. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Legon W, Sato TF, Opitz A, Mueller J, Barbour A, Williams A, et al. Transcranial focused ultrasound modulates the activity of primary somatosensory cortex in humans. Nature Neuroscience. 2014;17(2):322–9. [DOI] [PubMed] [Google Scholar]
  • 13.Hameroff S, Trakas M, Duffield C, Annabi E, Gerace MB, Boyle P, et al. Transcranial Ultrasound (TUS) Effects on Mental States: A Pilot Study. Brain Stimulation: Basic, Translational, and Clinical Research in Neuromodulation. 2013;6(3):409–15. [DOI] [PubMed] [Google Scholar]
  • 14.Gibson BC, Sanguinetti JL, Badran BW, Yu AB, Klein EP, Abbott CC, et al. Increased Excitability Induced in the Primary Motor Cortex by Transcranial Ultrasound Stimulation. Front Neurol. 2018;9:1007-. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Blackmore J, Shrivastava S, Sallet J, Butler CR, Cleveland RO. Ultrasound Neuromodulation: A Review of Results, Mechanisms and Safety. Ultrasound in Medicine & Biology. 2019;45(7):1509–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Plaksin M, Shoham S, Kimmel E. Intramembrane Cavitation as a Predictive Bio-Piezoelectric Mechanism for Ultrasonic Brain Stimulation. Physical Review X. 2014;4(1):011004. [Google Scholar]
  • 17.Oh S-J, Lee JM, Kim H-B, Lee J, Han S, Bae JY, et al. Ultrasonic Neuromodulation via Astrocytic TRPA1. Current Biology. 2019;29(20):3386–401.e8. [DOI] [PubMed] [Google Scholar]
  • 18.Van Overwalle F, Baetens K, Mariën P, Vandekerckhove M. Social cognition and the cerebellum: A meta-analysis of over 350 fMRI studies. Neuroimage. 2014;86:554–72. [DOI] [PubMed] [Google Scholar]
  • 19.Block H, Celnik P. Stimulating the cerebellum affects visuomotor adaptation but not intermanual transfer of learning. Cerebellum. 2013;12(6):781–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Schmahmann JD, Caplan D. Cognition, emotion and the cerebellum. Brain. 2006;129(2):290–2. [DOI] [PubMed] [Google Scholar]
  • 21.King RL, Brown JR, Newsome WT, Pauly KB. Effective Parameters for Ultrasound-Induced In Vivo Neurostimulation. Ultrasound in Medicine and Biology. 2013;39(2):312–31. [DOI] [PubMed] [Google Scholar]
  • 22.Baek H, Sariev A, Lee S, Dong SY, Royer S, Kim H. Deep Cerebellar Low-Intensity Focused Ultrasound Stimulation Restores Interhemispheric Balance after Ischemic Stroke in Mice. IEEE Transactions on Neural Systems and Rehabilitation Engineering. 2020;28(9):2073–9. [DOI] [PubMed] [Google Scholar]
  • 23.Baek H, Pahk KJ, Kim M-J, Youn I, Kim H. Modulation of Cerebellar Cortical Plasticity Using Low-Intensity Focused Ultrasound for Poststroke Sensorimotor Function Recovery. Neurorehabilitation and Neural Repair. 2018;32(9):777–87. [DOI] [PubMed] [Google Scholar]
  • 24.Baek H, Sariev A, Dong S, Royer S, Kim H. Cerebellar low-intensity focused ultrasound stimulation can normalize asymmetrical hemispheric delta power after mouse ischemic stroke. Brain Stimulation: Basic, Translational, and Clinical Research in Neuromodulation. 2019;12(2):487. [Google Scholar]
  • 25.Baek H, Sariev A, Kim MJ, Lee H, Kim J, Kim H, editors. A neuroprotective brain stimulation for vulnerable cerebellar Purkinje cell after ischemic stroke: a study with low-intensity focused ultrasound. 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); 2018 18–21 July 2018. [DOI] [PubMed] [Google Scholar]
  • 26.Van Dijck G, Van Hulle MM, Heiney SA, Blazquez PM, Meng H, Angelaki DE, et al. Probabilistic Identification of Cerebellar Cortical Neurones across Species. PLOS ONE. 2013;8(3):e57669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Asan AS, Lang EJ, Sahin M. Entrainment of cerebellar purkinje cells with directional AC electric fields in anesthetized rats. Brain Stimul. 2020;13(6):1548–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Sato T, Shapiro MG, Tsao DY. Ultrasonic Neuromodulation Causes Widespread Cortical Activation via an Indirect Auditory Mechanism. Neuron. 2018;98(5):1031–41.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Guo H, Hamilton M II, Offutt SJ, Gloeckner CD, Li T, Kim Y, et al. Ultrasound Produces Extensive Brain Activation via a Cochlear Pathway. Neuron. 2018;98(5):1020–30.e4. [DOI] [PubMed] [Google Scholar]
  • 30.Clarke RL, ter Haar GR. Temperature rise recorded during lesion formation by high-intensity focused ultrasound. Ultrasound in Medicine and Biology. 1997;23(2):299–306. [DOI] [PubMed] [Google Scholar]
  • 31.Izadifar Z, Izadifar Z, Chapman D, Babyn P. An Introduction to High Intensity Focused Ultrasound: Systematic Review on Principles, Devices, and Clinical Applications. J Clin Med. 2020;9(2):460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.ter Haar G Therapeutic ultrasound. European Journal of Ultrasound. 1999;9(1):3–9. [DOI] [PubMed] [Google Scholar]
  • 33.Krasovitski B, Frenkel V, Shoham S, Kimmel E. Intramembrane cavitation as a unifying mechanism for ultrasound-induced bioeffects. Proc Natl Acad Sci U S A. 2011;108(8):3258–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Yoo S, Mittelstein DR, Hurt R, Lacroix J, Shapiro MG. Focused ultrasound excites neurons via mechanosensitive calcium accumulation and ion channel amplification. bioRxiv. 2020:2020.05.19.101196. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Min B-K, Yang PS, Bohlke M, Park S, R.Vago D, Maher TJ, et al. Focused ultrasound modulates the level of cortical neurotransmitters: Potential as a new functional brain mapping technique. International Journal of Imaging Systems and Technology. 2011;21(2):232–40. [Google Scholar]
  • 36.Yang PS, Kim H, Lee W, Bohlke M, Park S, Maher TJ, et al. Transcranial focused ultrasound to the thalamus is associated with reduced extracellular GABA levels in rats. Neuropsychobiology. 2012;65(3):153–60. [DOI] [PubMed] [Google Scholar]
  • 37.Plaksin M, Kimmel E, Shoham S. Cell-Type-Selective Effects of Intramembrane Cavitation as a Unifying Theoretical Framework for Ultrasonic Neuromodulation. eNeuro. 2016;3(3):ENEURO.0136–15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Kubanek J, Shi J, Marsh J, Chen D, Deng C, Cui J. Ultrasound modulates ion channel currents. Scientific reports. 2016;6:24170-. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Kim H, Park MY, Lee SD, Lee W, Chiu A, Yoo S-S. Suppression of EEG visual-evoked potentials in rats through neuromodulatory focused ultrasound. Neuroreport. 2015;26(4):211–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Deffieux T, Younan Y, Wattiez N, Tanter M, Pouget P, Aubry JF. Low-intensity focused ultrasound modulates monkey visuomotor behavior. Curr Biol. 2013;23(23):2430–3. [DOI] [PubMed] [Google Scholar]
  • 41.Legon W, Sato TF, Opitz A, Mueller J, Barbour A, Williams A, et al. Transcranial focused ultrasound modulates the activity of primary somatosensory cortex in humans. Nat Neurosci. 2014;17(2):322–9. [DOI] [PubMed] [Google Scholar]
  • 42.Menz MD, Oralkan O, Khuri-Yakub PT, Baccus SA. Precise neural stimulation in the retina using focused ultrasound. J Neurosci. 2013;33(10):4550–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Yoon K, Lee W, Lee JE, Xu L, Croce P, Foley L, et al. Effects of sonication parameters on transcranial focused ultrasound brain stimulation in an ovine model. PLoS One. 2019;14(10):e0224311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Kim H, Chiu A, Lee SD, Fischer K, Yoo S-S. Focused Ultrasound-mediated Non-invasive Brain Stimulation: Examination of Sonication Parameters. Brain Stimul. 2014;7(5):748–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Isope P, Murphy TH. Low threshold calcium currents in rat cerebellar Purkinje cell dendritic spines are mediated by T-type calcium channels. J Physiol. 2005;562(Pt 1):257–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Brown AM, Arancillo M, Lin T, Catt DR, Zhou J, Lackey EP, et al. Molecular layer interneurons shape the spike activity of cerebellar Purkinje cells. Scientific reports 2019;9(1):1742-. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Tang T, Blenkinsop TA, Lang EJ. Complex spike synchrony dependent modulation of rat deep cerebellar nuclear activity. Elife. 2019;8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Person AL, Raman IM. Purkinje neuron synchrony elicits time-locked spiking in the cerebellar nuclei. Nature. 2011;481(7382):502–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.De Schutter E, Steuber V. Patterns and pauses in Purkinje cell simple spike trains: experiments, modeling and theory. Neuroscience. 2009;162(3):816–26. [DOI] [PubMed] [Google Scholar]
  • 50.Biorender. Created with Biorender.com 2020 [Google Scholar]

RESOURCES