Transcranial focused ultrasound modulates cortical and thalamic motor activity in awake sheep

Transcranial application of pulsed low-intensity focused ultrasound (FUS) modulates the excitability of region-specific brain areas, and anesthetic confounders on brain activity warrant the evaluation of the technique in awake animals. We examined the neuromodulatory effects of FUS in unanesthetized sheep by developing a custom-fit headgear capable of reproducibly placing an acoustic focus on the unilateral motor cortex (M1) and corresponding thalamic area. The efferent responses to sonication, based on the acoustic parameters previously identified in anesthetized sheep, were measured using electromyography (EMG) from both hind limbs across three experimental conditions: on-target sonication, off-target sonication, and without sonication. Excitatory sonication yielded greater amplitude of EMG signals obtained from the hind limb contralateral to sonication than that from the ipsilateral limb. Spurious appearance of motion-related EMG signals limited the amount of analyzed data (~ 10% selection of acquired data) during excitatory sonication, and the averaged EMG response rates elicited by the M1 and thalamic stimulations were 7.5 ± 1.4% and 6.7 ± 1.5%, respectively. Suppressive sonication, while sheep walked on the treadmill, temporarily reduced the EMG amplitude from the limb contralateral to sonication. No significant change was found in the EMG amplitudes during the off-target sonication. Behavioral observation throughout the study and histological analysis showed no sign of brain tissue damage caused by the acoustic stimulation. Marginal response rates observed during excitatory sonication call for technical refinement to reduce motion artifacts during EMG acquisitions as well as acoustic aberration correction schemes to improve spatial accuracy of sonication. Yet, our results indicate that low-intensity FUS modulated the excitability of regional brain tissues reversibly and safely in awake sheep, supporting its potential in theragnostic applications.


Evaluation of thermal effects
We examined the potential thermal effects from sonication by estimating the temperature increase at the sonicated M1 and adjacent skull by sequentially solving the Khokhlov-Zabolotskaya-Kuznetsov (KZK) equation and bio-heat transfer equation through an open-source high intensity FUS (HIFU) simulator based on MATLAB scripts 32 . The simulation was performed at resolution of 0.5 mm based on the previous numerical study of FUS propagation through the skull 13 using the sonication parameters and the maximum in situ acoustic intensity for each of excitatory (20.5 W/cm 2 Isppa) and suppressive sonication (13.7 W/cm 2 Isppa). The simulation of temperature rise from the M1 was conducted with a temporal resolution of 0.2 ms using the acoustic properties (speed of sound of 1482 m/s, density of 1000 kg/m 3 , attenuation coefficient of 0.217 dB/m·MHz -1 ) and thermal properties of the brain (specific heat of 3696 J/kg·K -1 , thermal conductivity of 0.55 W/m·K -1 , perfusion rate of 14.1 kg/m 3 ·s -1 ) 7 . For the simulation of thermal effects in the skull adjacent to the M1, specific heat of 1300 J/kg·K -1 , thermal conductivity of 0.4 W/m·K -1 , and perfusion rate of 0.143 kg/m 3 ·s -1 were used 66-68 .

Comparison of EMG amplitudes between on-target and off-target conditions
The group-averaged amplitudes of the EMG signals from each of experimental conditions (on-target, offtarget and no-FUS) were compared using one-way ANOVA followed by Tukey-Kramer post-hoc analysis.
In the comparison between on-target and off-target condition, the group-averaged amplitudes of the EMG signals from the hind limb contralateral to sonication became significantly greater upon on-target stimulation to the M1 and thalamus, compared to those obtained from the off-target condition (one-way

Comparison of EMG amplitudes between on-/off-target and no-FUS conditions
The EMG signal from the right hind limb contralateral to on-target sonication to the M1 was significantly greater than that from the same leg of the no-FUS condition within time segments of 42.0-108.6, 128.1-143.7, 155.5-186.7, 214.1-218.0, and 714.1-725.8 ms after the FUS onset (one-way ANOVA followed by Tukey-Kramer post-hoc analysis, F(2,27) = 9.1-36.5, P < 0.001 presented by brown dots in Fig. S2a). In the thalamic stimulation, a greater EMG signal from the right hind limb compared to that in the no-
Suppressive sonication of the thalamus selectively reduced the EMG amplitudes from the right hind limb in the F2 (-4.1 ± 3.2%; Tukey-Kramer post-hoc analysis, P < 0.01), compared to those from the off-target and no-FUS conditions (Fig. S4a). No differences were observed in all other time segments. There was no significant difference in the EMG amplitude obtained from the left hind limb (ipsilateral to sonication) across time segments throughout the experimental conditions (Fig. S4b). Table S1. Sheep-specific responsiveness to suppressive sonication applied to either M1 or thalamus. Presence of successful effects was determined if sheep showed at least one responsiveness (i.e., 'Y') in either 'F1' or 'F2' segment (i.e., duration of suppressive sonication). 'N' indicates the absence of responsiveness during suppressive sonication. Figure S1. Exemplar positively-skewed distribution of EMG amplitudes obtained from one of the sheep during pre-FUS segments. The red line indicates a Gaussian distribution fitted to the EMG amplitude profile. The mode value of the EMG amplitude distribution was set to the mean value of the Gaussian distribution. Figure S2. Comparison of time-locked EMG measurements from the excitatory sonication among experimental conditions. Comparisons of group-averaged EMG amplitudes (N = 10) in excitatory FUS to the M1 (a, c) and the thalamus (b, d) with respect to the EMG amplitudes from the off-target and no-FUS conditions. The red and magenta lines represent EMG amplitudes from the hind limb contralateral to sonication while the blue and cyan lines represent EMG amplitudes from the ipsilateral hind limb. Black lines represent EMG obtained during the no-FUS condition. The shaded area indicates the standard errors across all animals. The gray bar is the duration of sonication (200 ms). The green (for on-target FUS vs. off-target FUS) and brown (for on-target FUS vs. no-FUS) dots indicate significant differences (one-way ANOVA followed by Tukey-Kramer post-hoc analysis, P < 0.001) in the EMG amplitudes between the paired conditions.  . The condition-dependent percentage difference in the amplitudes was compared using one-way ANOVA followed by Tukey-Kramer post-hoc analysis ( * P < 0.05, ** P < 0.01). The red, orange, magenta, and gray color bars indicate average percentage values of the contralateral EMG signals in the on-target M1 & thalamus, off-target, and no-FUS conditions, respectively. The blue, purple, cyan, and gray color bars indicate average percentage values of the ipsilateral EMG signals. The error bars indicate standard errors. Figure S5. The relationship between in situ Isppa and response rates or EMG signals. Scatter plots of in situ Isppa and corresponding (a) response rates and (b) maximum EMG amplitude during FUS segment from excitatory sonication, as well as (c) percentage change of EMG signal amplitude from suppressive sonication. R 2 is a regression coefficient. P is a p-value estimated from regression analysis. Figure S6. Estimated temperature rise at the sonicated M1 (brain tissue) and adjacent skull for (a) excitatory sonication given at in situ Isppa = 20.5 W/cm 2 with 50 repetitions every 5 s and (b) suppressive sonication given for 1 min at in situ Isppa = 13.7 W/cm 2 .