Transcranial ultrasonic stimulation modulates single-neuron discharge in macaques performing an antisaccade task

HAL Id: hal-01675303

https://hal.sorbonne-universite.fr/hal-01675303

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires

Transcranial ultrasonic stimulation modulates

single-neuron discharge in macaques performing an

antisaccade task

Nicolas Wattiez, Charlotte Constans, Thomas Deffieux, Pierre Daye,

Jean-François Aubry, Pierre Pouget

To cite this version:

Nicolas Wattiez, Charlotte Constans, Thomas Deffieux, Pierre Daye, Jean-François Aubry, et al.. Transcranial ultrasonic stimulation modulates single-neuron discharge in macaques performing an antisaccade task. Brain Stimulation, Elsevier, 2017, 10 (6), pp.1024-1031. �10.1016/j.brs.2017.07.007�. �hal-01675303�

Transcranial ultrasonic stimulation modulates single-neuron

discharge in macaques performing an antisaccade task

2

3

Nicolas Wattiez1_{, Charlotte Constans}2_{, Thomas Deffieux}2_{, Pierre M. Daye}1_{, Mickael}

4

Tanter2, Jean-François Aubry2§ and Pierre Pouget1§.

5

1_{Institut du Cerveau et de la Moelle épinière, UMRS 975 INSERM, CNRS 7225, UMPC, Paris, France}

6

2_{Institut Langevin Ondes et Images, ESPCI ParisTech, CNRS 7587, UMRS 979 INSERM, Paris, France}

7 8 9

§ _{Both authors are co-responding authors and contributed equally to the work: [email protected] and}

10

[email protected]

11

ABSTRACT

Background

Low intensity transcranial ultrasonic stimulation (TUS) has been demonstrated to

non-15

invasively and transiently stimulate the nervous system. Although US neuromodulation

16

has appeared robust in rodent studies, the effects of US in large mammals and humans

17

have been modest at best. In addition, there is a lack of direct recordings from the

18

stimulated neurons in response to US. Our study investigates the magnitude of the US

19

effects on neuronal discharge in awake behaving monkeys and thus fills the void on both

20

fronts.

21

Objective/Hypothesis

In this study, we demonstrate the feasibility of recording action potentials in the

23

supplementary eye field (SEF) as TUS is applied simultaneously to the frontal eye field

24

(FEF) in macaques performing an antisaccade task.

25

Results

We show that compared to a control stimulation in the visual cortex, SEF activity is

27

significantly modulated shortly after TUS onset. Among all cell types 40% of neurons

28

significantly changed their activity after TUS. Half of the neurons showed a transient

29

increase of activity induced by TUS.

30

Conclusion

Our study demonstrates that the neuromodulatory effects of non-invasive focused

ultrasound can be assessed in real time in awake behaving monkeys by recording

33

discharge activity from a brain region reciprocally connected with the stimulated region.

34

The study opens the door for further parametric studies for fine-tuning the ultrasonic

35

parameters. The ultrasonic effect could indeed be quantified based on the direct

36

measurement of the intensity of the modulation induced on a single neuron in a freely

37

performing animal. The technique should be readily reproducible in other primate

38

laboratories studying brain function, both for exploratory and therapeutic purposes and

39

to facilitate the development of future clinical TUS devices.

INTRODUCTION

41

Focused ultrasound provides a new method to non-invasively stimulate the nervous

42

system using an external focused ultrasonic transducer to induce transient local

43

stimulation. Fry and colleagues demonstrated that US can suppress visually evoked

44

responses(1) in the primary visual cortex of anesthetized cats using focused ultrasound.

45

Gavrilov and collaborators elicited thermal, tactile, and pain sensations when they

46

applied focused ultrasound to the human and(2). The application of transcranial

47

ultrasonic stimulation (TUS) to the brain gained momentum in the new millenium, with

48

low frequency, low intensity ultrasound being applied the primary motor cortex in rodent,

49

generating motor responses without observing any adverse effect on the stimulated

50

tissue(3–6). Thermal effects are unlikely to play a role in the process when compared

51

with the mechanical effects such as radiation pressure and cavitation(7,8). However, the

52

exact mechanism for neural ultrasonic stimulation remains unknown(9,10). Several

53

models have been developed to explain the neuronal activation observed with acoustic

54

pressure: the Soliton model (11), the bilayer sonophore (BLS) model ( Intramembrane

55

cavitation as a unifying mechanism for ultrasound-induced bioeffects, (12) and the

56

neuronal intramembrane cavitation excitation (NICE) theory (13) combining the BLS

57

model and the Hodgkin-Huxley spiking model (which describes how action potentials are

58

initiated and propagate in neurons). No model has yet been evaluated using single

59

neuron recording during TUS in an alert animal performing a task. An improved

60

understanding of the physiological action of TUS would enable fine tuning of protocols

61

based on a known effect of the stimulation, which will improve the effectiveness of the

technique, and thus accelerate clinical transfer. Here we report the feasibility of

63

simultaneous recording of single neuron activity during TUS in an alert macaque

64

performing an oculomotor task. A focused ultrasound transducer was positioned using a

65

neuronavigation system (Rogue Research) to apply brief pulses of ultrasound in the

66

frontal eye field (FEF) while neurophysiological recordings were performed in the

67

supplementary eye field (SEF).

68

Two discrete areas in frontal cortex are involved in generating saccadic eye movements

69

in primates—the frontal eye field (FEF) and the supplementary eye field (SEF).

70

Anatomically the FEF is reciprocally and bilaterally connected with the SEF(14,15) and

71

these connections appear to vary the least between primate species, suggesting that the

72

most conserved functions of these regions may serve as a link for facilitating, planning of

73

eye movements(16). In both regions, eye movements can be evoked by direct electrical

74

microstimulation with relatively low current(17). Invasive intra-cortical recordings and/or

75

stimulation techniques have been widely used to better apprehend the role of these two

76

frontal areas, in combination with ocular motor measurements and mapping techniques,

77

such as electrophysiological and fMRI recordings in non-human primates. Those

78

approaches have provided causal evidence about the role of the FEF, with direct

79

bearing on ocular motor and attentional function(18,19). More recently, the causal

80

involvement of FEF in attentional orienting has been demonstrated(20,21) as well as its

81

ability to modulate different aspects of visual perception(22,23) while indirect control and

82

monitoring processes have been assigned to SEF functions(24,25). In order to examine

83

the possible distant role of neuromodulation caused by FUS applied to a cortical brain

84

region, we decided to target a cortical circuit controlling voluntary eye movement.

Our short-term goal was to demonstrate the feasibility of combining real time focused

86

ultrasound with single unit (SU) recordings in an alert macaque performing an

87

antisaccade task. The combined approach of TUS and single unit recording has the

88

potential to elucidate the mechanisms underlying the effects of focused ultrasound on

89

the central nervous system. TUS and single unit recording could also be combined with

90

3D functional ultrasound imaging of brain activity, as recently demonstrated in alert

91

animals in conjunction with EEG recordings(26,27). This will serve the causal

non-92

invasive exploration of cognition in non-human primate models.

93 94

MATERIALS AND METHODS

95

Focused ultrasound

96

A single element ultrasound transducer (H115, diameter 64 mm, Sonic Concept, Bothell,

97

WA, USA), geometrically focused to 64 mm, was used with a coupling cone (C103,

98

Sonic Concepts, Bothell, WA, USA) filled with degassed water.

99

The transducer has four frequency resonances: 200 kHz, 320 kHz, 850 kHz and 1380

100

kHz. We chose the same main parameters (frequency and burst length) as our previous

101

study(6) which had shown successful neuromodulation with the same transducer in

102

monkeys: the ultrasound wave frequency was set to 320kHz and 100ms bursts of

103

ultrasound were generated with a digital function generator, with rise and fall times set to

104

5ms at the beginning and the end of the pulse, respectively (Handyscope HS5, TiePie

105

engineering, Sneek, The Netherlands). A 75-Watt amplifier (75A250A, Amplifier

106

Research, Souderton, PA) was used to deliver the required power to the transducer and

107

the input voltage and current of the transducer were recorded using a voltage probe

(HP9250, TiePie engineering, Sneek, The Netherlands) and a current probe (P6021,

109

Tektronix, Beaverton, Oregon) connected to the HS5 card inputs.

110

The recorded peak to peak voltage (VPP) was 139±26 V for monkey Y and 56±15 V for

111

monkey L. The calibration of the transducer in water with a custom built heterodyne

112

interferometer (28) led to a linear relationship between VPP and peak negative pressure

113

(PNP) in water of 5.68 kPa/V. The peak pressure in the brain of the monkeys was

114

estimated by using 3D simulations of the ultrasonic field (see details below), taking into

115

account the experimental geometry (0.41±0.08 MPa for monkey Y and 0.24±0.07 MPa

116

for monkey L). For comparison, previous studies in human and non-human primates

117

estimated peak pressures of 0.35 MPa(6) and 0.43 MPa(7).

118

The ultrasound was directly applied to previously shaved skin. A coupling cone filled with

119

water ensured ultrasonic coupling between the transducer and the animal’s head. TUS

120

was performed once every 5 trials with an unpredictable minimum duration of 3 seconds

121

between trials. There was therefore a pause of at least 10 seconds between ultrasound

122

pulses.

123

The corresponding overall spatial peak pulse average intensity (ISPPA) was thus

124

estimated to be 5.6 ± 0.2 W/cm2 and 1.9 ± 0.2 W/cm2 for monkey Y and L respectively

125

(ISPPA=_#$∗&!² with and c approximated to '_()*+, and -_()*+,). For comparison, the Food

126

and Drug Administration (FDA) ISPPA limit for ultrasound imaging safety guidelines is

127

190 W/cm2 for adult transcranial imaging [Food, U. and D. Administration, Information

128

for manufacturers seeking marketing clearance of diagnostic ultrasound systems and

129

transducers. Rockville, MD: Center for Devices and Radiological Health, US Food and

Drug Administration, 1997]. We emphasize here that we do not claim to be below FDA

131

requirements as long pulses (100ms, or 32000 periods) are used here, as compared to

132

short pulses for imaging (one period for Bmode imaging, corresponding to 3µs at the

133

frequency used here).

134 135

Experimental setup

136

The antisaccade paradigms and data acquisition were under the control of a computer

137

running a Rexeno real-time data acquisition system(6).

138 139

Behavioral Task

140

Two captive-born macaques (Maccaca mulatta, “Y”, and “L”, one male and one female

141

11 and 8 years old and 17 and 7 kg respectively) participated in this study. The monkeys

142

were paired-housed and handled in strict accordance with the recommendations of the

143

Weatherall Report on good animal practice. Our experiments were approved by the

144

Animal Health and Veterinary Medication Division of the Department of Public Veterinary

145

Health, Nutrition and Food Safety of the French Ministry of Health (current renewal N°

146

DTPP 2010-424). Monkeys had access to toys, mirrors and swings and also had visual,

147

auditory and olfactory contact with other animals, and, when appropriate, could touch

148

and groom each other. An institutional veterinary doctor constantly monitored their

well-149

being and health conditions. Prior to participating in the study, both animals were

150

periodically chaired, head-posted and trained to perform a series of tasks for a period of

151

6-12 months, until they became regular and proficient performers.

Prior to the first experimental session, animals were specifically trained in an

153

antisaccade (AS) task. Between 500-1000 milliseconds after initial fixation onset on a

154

brown central fixation stimulus and simultaneously to its disappearance (no gap), a red

155

square appeared for 1000ms at a 16° of visual angle to a rightward or leftward location

156

(direction randomly selected). Monkeys were trained not to look at this peripheral target

157

but instead, initiate as soon as possible a saccade in the opposite direction. After the

158

saccade, the monkey received a reward if the saccade fell within a 10° x 10° window

159

centered at the mirror location of the visual target. Failure to trigger a saccade within

160

1000ms of target onset cancelled the trial. The AS task was chosen since prior human

161

and monkey TUS experiments have revealed prosaccade paradigms to be much less

162

sensitive to single pulse TUS interference than antisaccades(6). Trials were aligned on

163

the appearance of the peripheral target.

164 165

Data Analysis

166

Eye movements were recorded with an infrared video eyetracker (Eyelink 1k,

SR-167

Research, Ontario, Canada). Eye position was digitized and sampled at 1 kHz and

168

stored for off-line analysis. All the analyses were performed on a personal computer with

169

custom software written in the Python programming language. Eye position signals were

170

low-pass filtered using a third order Savitzky–Golay filter (window: 41ms). Eye velocity

171

and eye acceleration were extracted from the Savitzky-Golay filter. Saccades were

172

detected with a 30°/s velocity threshold.

173

For each cortical site tested with TUS, we used an online windows discriminator to

174

isolate one neuron at a time for each recording. Offline cluster cutting (PCA) techniques

were used systematically to confirm that action potential waveforms were produced by

176

individual neuronal elements.

177

Spikes were sorted on-line using a box sorting method as a pair of time-level windows

178

(Rasputin, Plexon Inc.). Offline examination of the sorting quality was systematically

179

realized using a semi-automatic PCA calculation without restriction to a certain number

180

of waveforms (Offline-Sorter Plexon Inc). Following the extraction of the spike time, we

181

evaluated the average firing rate of each neuron (and the confidence interval of the

182

average) using bootstrapping techniques on the computation of the spike density

183

function (27). The principle of bootstrap is to randomly select trials without replacement

184

from the larger dataset (here control trials) and with replacement in the smaller dataset

185

(TUS trials). This random selection is made N times where N = 50 * sample size of the

186

smaller dataset as described by Daye and colleagues(29). The effect of FUS between

187

functional cell types was compared using the chi-squared test.

188 189

Surgical procedure

190

The surgical procedures for the titanium head-post implant were as previously

191

described(6). Each animal was deeply anaesthetized using ketamine hydrochloride (5

192

mg/kg i.m.) for initial sedation before induction with 2-4% isoflurane gas, which was then

193

maintained during surgery. Pain medication was given prior to surgery and routinely

194

given after surgery. Implanted head posts (7.1 mm internal diameter) (Part 6-FHP-X2F,

195

Crist Instrument, Hagerstown, MD, USA) were designed to mate with a head-post holder

196

(Part 6-FHB-S2B, Crist Instrument, Hagerstown, MD, USA). SEF Recording chambers

197

were placed according to stereotactic coordinates (AP: 26; ML: 0) and MRI scans. The

craniotomies were performed under deep anesthesia with ketamine hydrochloride (5

199

mg/kg i.m.) for initial sedation, and induction and maintenance with 2-4% isoflurane gas.

200

Pain medication was given prior the surgery and routinely given after surgery.

201 202

Neurophysiological data acquisition and analysis

203

For single unit recordings, multiple tungsten microelectrodes (FHC, 8-10 MΩ) were

204

introduced through guide tubes and independently positioned with a custom-made

205

stepper-motor micro-drive. The recordings were made on a MAP system (Plexon Inc.,

206

TX, USA), storing spike times with 25μs resolution. Online discrimination based on wave

207

shape parameters was used to select a single neuron on each independent recording

208

electrode channel. Respectively 39 and 46 neurons were recorded in monkey Y and L.

209

All isolated neurons recorded were kept for the analysis. All sets of recordings consisted

210

in more than 386 and 400 valid trials for monkey Y and L respectively.

211 212

FEF Targeting with TUS

213

Targeting of the FEF with the TUS transducer was performed using a neuro-navigator

214

tool (Brainsight, Rogue Research, Montreal, Canada). An MRI dataset (structural T1)

215

was acquired for each monkey with an attached MR fiducial marker array. The FEF

216

target was manually registered once and used for all sessions. Prior to each session, the

217

position of the fiducial marker array was manually registered to a position sensor serving

218

as a reference and fixed to the chair. In a second step, the focal spot of the TUS

219

transducer, materialized using a 3D printed piece, was registered using a tracked pointer

220

to a second position sensor, which was fixed behind the transducer. Before each

session, the transducer, supported by a mechanical arm (Rogue Research, Montreal,

222

Canada) was then manually moved until the focal spot was aligned and positioned over

223

the predetermined FEF target. The arm was then locked for the rest of the session.

224 225

Numerical simulations

226

Acoustic density and velocity maps were reconstructed from CT scans of both monkeys’

227

heads (Philips Brilliance 64 CT Scanner , Voltage: 120kV, Intensity: 350mA ; FilterType:

228

'D'). The brain, skin and muscle tissue was considered as a homogeneous medium and

229

skull was modeled as a heterogeneous model by using the following conversion from

230

Hounsfield units (HU) to density and velocity of the skull1:

231

' = '_*/001++ 34 − 34*/001+

34₆₎₇+ 34_*/001+('6)7− '*/001+)

232

HUmax was determined from each map as the maximal value in the skull and HUtissue as

233

the mean value of HU in the brain, muscle and skin.

234

The power law model for attenuation is abs = α ∗ Ф? _{where the porosity Ф is defined by}

235

Ф = $@ABC$

$@ABC$DAEFG in the skull (Aubry et al. 2003 Experimental demonstration of noninvasive

236

transskull adaptive focusing based on prior computed tomography scans). α and I were

237

experimentally determined by transmission measures and simulations to a monkey skull

238

flap : the best match was found with I = 0.5 and α = 8.0 NO/-Q/R3ST.T_{, leading to α =}

239

2.3 db/cm at 320 kHz.

240

The acoustic parameters are summarized in table 2. CT images were acquired with the

241

smallest field of view to obtain the best possible resolution in the transverse plane: 0.358

242

mm and 0.136 mm for monkey Y and L respectively. The acoustic density and velocity

243

3D maps were then subsampled to respectively 0.358 and 0.407 mm isotropic resolution

for monkey Y and monkey L, which gave a ratio of 13.6 and 12.0 pixels per wavelength

245

in tissue at 320 kHz.

246

The pseudo-spectral method-based software k-Wave2 _{was used to perform the linear}

247

acoustic simulation through the water cone, skull and brain with the same geometry

248

used in vivo with the neuro-navigator positioning. The head was modelled as immersed

249

in water, given that, in that setup, tissue/air interfaces are unlikely to play a significant

250

role compared to skull/tissue interfaces. Not considering the air medium, whose

251

impedance is much lower than water, significantly reduced the computation time.

252

To limit the computation time, the total simulation duration was limited to 150 µs, which

253

was sufficient to reach a steady-state. For each simulation, the pressure field was stored

254

in three dimensions and in a full volume including the FEF for all time steps. Time

255

profiles of the acoustic pressure were extracted at the geometric focus and at the point

256

of maximal pressure amplitude over time. Maps of the in-situ time peak pressure at

257

steady-state were also estimated by taking the time peak of the pressure field at each

258

pixel in 3 planes (orthogonal to, or including the transducer direction), centered on the

259

geometric focus. As the muscular tissue surrounding monkey Y’s skull was much thicker

260

than monkey L’s (respectively about 16mm and 5mm), a 1 cm transducer’s pressed into

261

the tissue was taken into account in the simulation. This value was set based on the

262

actual in vivo positioning of the transducer, based on the neuro-navigator guidance for

263

FEF targeting. The calibration of the transducer in water showed a linear relationship

264

between peak to peak voltage (VPP) and peak negative pressure (PNP) in water of 5.68

265

kPa/V. Simulations were first performed in water without the skull, in order to normalize

the simulations based on actual pressure measurements in water. Table 1 summarizes

267

the results of the simulations.

268 269

RESULTS

270 271

The goal of our study is to show that TUS and single unit recordings can be combined to

272

probe the interactions between two reciprocally-connected distal bilateral areas. First,

273

we present how TUS applied to FEF affects the neural activity of a SEF neuron. Then

274

we extend these analyses to the whole neuronal population that has been recorded to

275

investigate the overall effect of TUS on a distal connected area.

276 277

Simulation of ultrasonic pressure field in the brain

278

The voltage applied to the ultrasonic transducer was respectively equal to 139±26 V and

279

56±15 V for monkey Y and L. As shown in figure 2, female monkey L had a smaller head

280

than male monkey Y. 3D numerical simulations of the ultrasound propagation were

281

conducted in order to investigate the ultrasound field in the two different monkey brains

282

while taking into account head geometry, and the reflection, refraction and absorption

283

associated with the full skull cavity and head geometry, (table 1). A linear simulation

284

model was used (see supplementary materials) in which the acoustic pressure was

285

proportional to the applied voltage (VPP). In water, the peak pressure was found to be

286

2.5 times higher with monkey Y’s settings than monkey L’s. Nevertheless, the peak

287

pressure in the FEF is only 1.7 times higher for monkey Y. This is most probably due to

anatomical differences such as skull thickness, skull density, skull geometry and frontal

289

muscular tissue thickness (table 1 and figure 1). Owing to the different acoustic

290

pressures used between the two animals, as well as the compounding effects of

291

differences in head size, skull thickness, the recorded neuron data from each animal

292

was analyzed separately and is quantified in figure 4 (top panel for monkey Y and

293

middle panel for monkey L). To minimize the difference between peak pressure in the

294

brain between the two animals the pressure was slightly reduced for monkey Y

295

(compared to Deffieux et al. 2013). Behaviorally, the latencies were significantly

296

modulated by TUS for monkey L across sessions (n=46, p<0.01) and marginally

297

modulated for monkey Y (n=39, p=0.08).

298 299

Typical trials

300

Figure 3: Recordings of neuronal spikes activated by single pulse TUS. (a) Example raw

301

data from 41 trials with TUS (red) and 179 without stimulation (blue). (b) Bootstrap

302

difference of the control vs. TUS condition at the same site (p<0.01). These data were

303

selected for illustration because the TUS elicited activation of diverse neuronal

304

elements. The triangles represent latencies for each trial respectively for TUS (red) or no

305

stimulation (blue) trials. The average firing activity across trials are shown during control

306

trials (blue line) and following TUS (red line) (figure 3a,b). Red and blue shaded areas

307

represent the 99% confidence interval respectively for average activity. The time course

308

of the difference between these two conditions (control and TUS) is shown in figure 3c,d.

309

The black line and shaded areas represent the 99% confidence interval for activity

difference when comparing TUS and control trials. The blue vertical rectangles represent

311

the time interval during TUS stimulation (100ms) where time t=0 corresponds to target

312

presentation. We consider the difference between TUS and control conditions activity as

313

statistically different from zero when p < 0.01 and the duration of the difference is

314

greater than or equal to 5ms.

315

The green (increase in activity) or red (decrease in activity) segments in fig. 3c,d

316

highlights these significant time periods. In the stimulation condition, figure 3a shows

317

that the activity started to increase 67 ms after TUS onset and became statistically

318

significantly different from the control condition 78 to 111 ms and from 113 to 125 ms

319

after TUS onset (shown by the first significant segment in fig. 3b). The neuron reached a

320

peak activity of 10.1Hz (with 99% CI of 15.48 and 21.61Hz) 99 ms after stimulation

321

onset compared to 2.1Hz (with 99% CI of 5.45 and 9.53Hz) during control trials. For this

322

neuron, this level of activity was maintained during the stimulation. After stimulation

323

ceased, the neuron activity decayed back to the activity observed during the control

324

condition. As shown on the right panel (figure 3b,d), a decrease in activity in the SEF

325

was also observed following FUS in the FEF. The activity of this neuron starts to

326

increase before TUS and appeared to be significantly reduced from 24 to 41 ms after

327 TUS onset. 328 329 Population activity 330

Figure 4a,b illustrates changes in the activity of the population of the neurons recorded

331

in the two animals. Significant periods of neuromodulation were observed for 39% and

41% of the recorded neurons respectively for monkey Y and L. A transient increase of

333

activity for 53% and 47% of neurons recorded respectively in monkey Y and L reveal the

334

perturbation induced by TUS. For 47% and 53% of neurons respectively in monkey Y

335

and L the perturbation was revealed by a significant decrease in activity.

336

Among all recorded neurons the effects of TUS parameters are presented

337

chronologically during the series of experiments for the 39 and 46 neurons respectively

338

recorded in monkey Y and L.

339

Each horizontal gray line represents a neuron. As in figure 3, the time origin corresponds

340

to target presentation, and shaded gray areas represent the stimulation period. The

341

colored lines in figure 4 highlight the periods during which the neuronal activity differed

342

significantly between TUS and control trials. The color of each segment represents the

343

p-value of the difference between TUS and control trials. Green (red) colored segments

344

correspond to excitation (inhibition). Figure 4a,b shows that many of the recorded

345

neurons had their activity modulated (increased or decreased) when TUS was applied in

346

both monkeys. Overall, our analysis demonstrates a significant effect on SEF neuronal

347

activity when TUS was applied to the FEF. Comparison between cell types based on

348

their functional properties(25,30) did not reveal any significant susceptibility (respectively

349

χ

χ

350

of the 3 of fixation; 5 of the 11 of movement; 9 of the 18 non-specific; 1 of the 7

post-351

saccadic; and 0 of 0 visual neurons showed a significant neuro-modulation induced by

352

TUS. For monkey L: of 46 neurons, 2 of 2 of fixation; 2 of 9 of movement; 13 of 26

non-353

specific; 1 of 7 post-saccadic; and 1 of 2 of visual neurons showed significant

modulation by TUS. In contrast to TUS stimulation applied to the FEF, TUS applied to

355

control extra-striate visual cortex did not significantly generate changes in activity rates

356

(figure 4c). Only one of the 20 recorded neurons did show a brief significant period of

357

modulation when the control region was stimulated compared with 39 and 41% of

358

significant neurons showing modulation when the FEF was stimulated in the same

359

animal (monkey L). We verified whether the observed change of activity during the

360

stimulation could be explained by a modification of the spikes isolation of the neuron due

361

to the ultrasound stimulation. We did not observe any difference of spike waveform

362

when comparing TUS and control conditions (figure 5), demonstrating that the pattern of

363

activity of our neuron observed during stimulation trials is not an artifact related to a

364

propagation of the ultrasound wave to the recorded region (CI: 99%). At the population

365

level none of the recorded neurons show a significant modulation of waveforms above

366

the confidence interval (Figure 6).

367 368

DISCUSSION

369

Focused ultrasound stimulation (FUS) techniques have been used to produce increased

370

neuronal firing rates in the peripheral and central nervous system in vitro or in

371

anesthetized preparations (4,31–35). Suppression of activity following FUS has also

372

been reported (5,36,37), as well as more subtle changes in excitability(38). However,

373

non-invasive exploitation of the concomitant effects of transcranial ultrasound stimulation

374

(TUS) in alert animals would have many applications in neuroscience. However, and

despite the increasing volume of work done with the emerging of transcranial ultrasound

376

stimulation (TUS), there is limited knowledge concerning its effects on neurons.

377

Investigating neural connectivity—how two connected regions interact by stimulating one

378

area and recording in another—is the most straightforward application. Our approach

379

was to perform the feasibility of such study by studying the interaction within the

380

oculomotor regions of the frontal lobes in awake primates performing an antisaccade

381

task. By combining TUS with traditional electrophysiology, the ability to study the

382

mechanisms of TUS is vastly expanded, and paves the way for research and

383

development needed to usher in the next generation of noninvasive TUS technologies

384

and treatments. In particular, the study opens the door for further parametric studies for

385

fine-tuning the ultrasonic parameters. The ultrasonic effect could indeed be quantified

386

based on the direct measurement of the intensity of the modulation induced on a single

387

neuron in a freely performing animal. In this paper the recorded region (SEF) was

388

different from the stimulated region (FEF) and despite the fact that these two regions are

389

connected, not each FEF neuron projects to an SEF neuron. The fact that TUS can

390

induce an effect in a connected brain region re-enforces the potential uses of this

391

technic to modify network activities. Further studies are required to better understand the

392

mechanisms of action of TUS. Potentially, clinical use of TUS might be similar to

393

transcranial magnetic stimulation (TMS) and could thus be valuable in the treatment of

394

depression(39), movement disorders(40,41), or patients suffering from perceptual

395

disorders(42). The development of ultrasonic neuro- modulation is largely motivated by

396

future therapeutic applications. This is the only neuromodulation technology currently

397

capable of being selective, targeted, reversible and noninvasive with millimeter-scale

precision essentially across the entire human and non-human primate brain(6–8) and

399

peripheral nervous system(43–45).

400

Moreover, deep seated clinical sites could be targeted with TUS, as is currently the case

401

in the thalamus with transcranial high intensity focused ultrasound.

402

ACKNOWLEDGMENTS

404

This work has been supported by the Bettencourt Schueller Foundation, the LABEX

405

WIFI (Laboratory of Excellence within the French Program "Investments for the Future")

406

under references ANR-10-LABX-24 and ANR-10-IDEX-0001-02 PSL and by the

407

National Agency for Research under the program "future investments" with the reference

408

ANR-10-EQPX-15. We wish to thank Jeff Schall and Jérôme Sallet for reading a

409

preliminary version of the manuscript and for suggesting useful improvements. We also

410

wish to thank Mathieu Santin for anatomical MRI acquisition.

411 412

AUTHOR CONTRIBUTIONS

413

N.W., T.D., J.F.A., and P.P., designed the experiments; N.W., performed the

414

experiments. C.C., and J.F.A, performed the simulations. N.W., P.M.D., and P.P.,

415

analyzed the data. N.W., T.D., C.C., P.M.D., M.T., J.F.A and P.P wrote the manuscript.

416

REFERENCES

418

419

1. Fry W, Meyers R, Fry F, Schultz D, Dreyer L, Noyes R. In: Trans Am Neurol

420

Assoc 1958; p16.

421

2. Gavrilov LR, Gersuni GV, Ilyinsky OB, Sirotyuk MG, Tsirulnikov EM, Shchekanov EE.

422

The effect of focused ultrasound on the skin and deep nerve structures of man and

423

animal. Prog Brain Res 1976; 279–92.

424

3. Younan Y, Deffieux T, Larrat B, Fink M, Tanter M, Aubry J-F. Influence of the

425

pressure field distribution in transcranial ultrasonic neurostimulation. Med Phys 2013;

426

40(8):82902.

427

4. King RL, Brown JR, Newsome WT, Pauly KB. Effective parameters for

428

ultrasound-induced in vivo neurostimulation. Ultrasound Med Biol 2013; 39(2):312–31.

429

5. Yoo S-S, Bystritsky A, Lee J-H, Zhang Y, Fischer K, Min B-K, et al. Focused

430

ultrasound modulates region-specific brain activity. NeuroImage 2011; 56(3):1267–75.

431

6. Deffieux T, Younan Y, Wattiez N, Tanter M, Pouget P, Aubry J-F. Low-intensity

432

focused ultrasound modulates monkey visuomotor behavior. Curr Biol 2013;

433

23(23):2430–3.

434

7. Legon W, Sato TF, Opitz A, Mueller J, Barbour A, Williams A, et al. Transcranial

435

focused ultrasound modulates the activity of primary somatosensory cortex in humans.

436

Nat Neurosci 2014; 17(2):322–9.

437

8. Lee W, Kim H, Jung Y, Song I-U, Chung YA, Yoo S-S. Image-guided transcranial

438

focused ultrasound stimulates human primary somatosensory cortex. Sci Rep 2015;

439

5:8743.

440

9. Aubry J-F, Tanter M. MR-Guided Transcranial Focused Ultrasound. Adv Exp Med

441

Biol 2016; 880:97–111.

442

10. Hynynen K, Clement G. Clinical applications of focused ultrasound-the brain. Int J

443

Hyperth Off J Eur Soc Hyperthermic Oncol North Am Hyperth Group 2007; 23(2):193–

444

202.

445

11. Heimburg T, Jackson AD. On soliton propagation in biomembranes and nerves.

446

Proc Natl Acad Sci USA 2005; 102(28):9790–5.

447

12. Krasovitski B, Frenkel V, Shoham S, Kimmel E. Intramembrane cavitation as a

448

unifying mechanism for ultrasound-induced bioeffects. Proc Natl Acad Sci USA 2011;

449

108(8):3258–63.

450

13. Plaksin M, Kimmel E, Shoham S. Cell-Type-Selective Effects of Intramembrane

451

Cavitation as a Unifying Theoretical Framework for Ultrasonic Neuromodulation. eNeuro

452

2016; 3(3).

453

14. Huerta MF, Krubitzer LA, Kaas JH. Frontal eye field as defined by intracortical

454

microstimulation in squirrel monkeys, owl monkeys, and macaque monkeys. II. Cortical

455

connections. J Comp Neurol 1987; 265(3):332–61.

456

15. Luppino G, Matelli M, Rizzolatti G. Cortico-cortical connections of two

457

electrophysiologically identified arm representations in the mesial agranular frontal

458

cortex. Exp Brain Res 1990; 82(1):214–8.

459

16. Percheron G, François C, Pouget P. What makes a frontal area of primate brain

460

the frontal eye field? Front Integr Neurosci 2015; 9:33.

17. Schall JD. Visuomotor Areas of the Frontal Lobe. In: Rockland KS, Kaas JH,

462

Peters A, editors. Extrastriate Cortex in Primates. Boston, MA: Springer US; 1997.

463

p527–638.

464

18. Bruce CJ, Goldberg ME. Primate frontal eye fields. I. Single neurons discharging

465

before saccades. J Neurophysiol 1985; 53(3):603–35.

466

19. Marrocco RT. Saccades induced by stimulation of the frontal eye fields:

467

interaction with voluntary and reflexive eye movements. Brain Res 1978; 146(1):23–34.

468

20. Moore T, Fallah M. Control of eye movements and spatial attention. Proc Natl

469

Acad Sci USA 2001; 98(3):1273–6.

470

21. Moore T, Armstrong KM. Selective gating of visual signals by microstimulation of

471

frontal cortex. Nature 2003; 421(6921):370–3.

472

22. Ekstrom LB, Roelfsema PR, Arsenault JT, Bonmassar G, Vanduffel W. Bottom-up

473

dependent gating of frontal signals in early visual cortex. Science 2008; 321(5887):414–

474

7.

475

23. Thickbroom GW, Stell R, Mastaglia FL. Transcranial magnetic stimulation of the

476

human frontal eye field. J Neurol Sci 1996; 144(1–2):114–8.

477

24. Pouget P. The cortex is in overall control of “voluntary” eye movement. Eye 2015;

478

29(2):241–5.

479

25. Pouget P, Murthy A, Stuphorn V. Cortical control and performance monitoring of

480

interrupting and redirecting movements. Philos Trans R Soc B Biol Sci 2017;

481

372(1718):20160201.

482

26. Sieu L-A, Bergel A, Tiran E, Deffieux T, Pernot M, Gennisson J-L, et al. EEG and

483

functional ultrasound imaging in mobile rats. Nat Methods 2015; 12(9):831–4.

484

27. Macé E, Montaldo G, Cohen I, Baulac M, Fink M, Tanter M. Functional ultrasound

485

imaging of the brain. Nat Methods 2011; 8(8):662–4.

486

28. Jacob X, Barriere C, Takatsu R, Montaldo G, Royer D. Optical measurement of

487

transient ultrasonic shock waves. IEEE 2004; p52–5.

488

29. Daye PM, Roberts DC, Zee DS, Optican LM. Vestibulo-ocular reflex suppression

489

during head-fixed saccades reveals gaze feedback control. J Neurosci Off J Soc

490

Neurosci 2015; 35(3):1192–8.

491

30. Stuphorn V, Schall JD. Executive control of countermanding saccades by the

492

supplementary eye field. Nat Neurosci 2006; 9(7):925–31.

493

31. Tyler WJ, Tufail Y, Finsterwald M, Tauchmann ML, Olson EJ, Majestic C. Remote

494

Excitation of Neuronal Circuits Using Low-Intensity, Low-Frequency Ultrasound.

495

Tanimoto H, editor. PLoS ONE 2008; 3(10):e3511.

496

32. Tufail Y, Matyushov A, Baldwin N, Tauchmann ML, Georges J, Yoshihiro A, et al.

497

Transcranial Pulsed Ultrasound Stimulates Intact Brain Circuits. Neuron 2010;

498

66(5):681–94.

499

33. Kim H, Chiu A, Lee SD, Fischer K, Yoo S-S. Focused Ultrasound-mediated

Non-500

invasive Brain Stimulation: Examination of Sonication Parameters. Brain Stimulat 2014;

501

7(5):748–56.

502

34. Naor O, Hertzberg Y, Zemel E, Kimmel E, Shoham S. Towards multifocal

503

ultrasonic neural stimulation II: design considerations for an acoustic retinal prosthesis. J

504

Neural Eng 2012; 9(2):26006.

505

35. 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.

507

36. Min B-K, Bystritsky A, Jung K-I, Fischer K, Zhang Y, Maeng L-S, et al. Focused

508

ultrasound-mediated suppression of chemically-induced acute epileptic EEG activity.

509

BMC Neurosci 2011; 12(1):23.

510

37. Kim H, Park MY, Lee SD, Lee W, Chiu A, Yoo S-S. Suppression of EEG

visual-511

evoked potentials in rats through neuromodulatory focused ultrasound: NeuroReport

512

2015; 26(4):211–5.

513

38. Moore JH, Gieck JH, Saliba EN, Perrin DH, Ball DW, McCue FC. The biophysical

514

effects of ultrasound on median nerve distal latencies. Electromyogr Clin Neurophysiol

515

2000; 40(3):169–80.

516

39. George MS, Taylor JJ, Short EB. The expanding evidence base for rTMS

517

treatment of depression. Curr Opin Psychiatry 2013; 26(1):13–8.

518

40. Lefaucheur J-P, André-Obadia N, Antal A, Ayache SS, Baeken C, Benninger DH,

519

et al. Evidence-based guidelines on the therapeutic use of repetitive transcranial

520

magnetic stimulation (rTMS). Clin Neurophysiol Off J Int Fed Clin Neurophysiol 2014;

521

125(11):2150–206.

522

41. Hallett M. Transcranial magnetic stimulation and the human brain. Nature 2000;

523

406(6792):147–50.

524

42. Leocani L, Cohen LG, Wassermann EM, Ikoma K, Hallett M. Human corticospinal

525

excitability evaluated with transcranial magnetic stimulation during different reaction time

526

paradigms. Brain J Neurol 2000; 123:1161–73.

527

43. Tsui P-H, Wang S-H, Huang C-C. In vitro effects of ultrasound with different

528

energies on the conduction properties of neural tissue. Ultrasonics 2005; 43(7):560–5.

529

44. Colucci V, Strichartz G, Jolesz F, Vykhodtseva N, Hynynen K. Focused

530

Ultrasound Effects on Nerve Action Potential in vitro. Ultrasound Med Biol 2009;

531

35(10):1737–47.

532

45. Juan EJ, González R, Albors G, Ward MP, Irazoqui P. Vagus nerve modulation

533

using focused pulsed ultrasound: Potential applications and preliminary observations in

534

a rat. Int J Imaging Syst Technol 2014; 24(1):67–71.

535 536 537 538 539 540 541 542

Table 1: Estimations of peak pressure in monkeys' brains 543 544 Monkey Y Monkey L VPP 139±26 V 56±15 V Equivalent PNP in free water 0.76±0.08 MPa 0.31±0.14MPa

Simulated steady state

PNPin FEF 0.41±0.08 MPa 0.24±0.07 MPa

545

Table 2: Acoustic parameters

546 547

a) Water Tissue Max

Velocity (m/s) 1500 1560 3100

Density (kg/m3) 1000 1030 2200 Absorption

(dB/cm/MHz1.1) 0 0.38 8.0

a) Water Tissue Max

Velocity (m/s) 1500 1560 3100 Density (kg/m3) 1000 1030 2200 Absorption (dB/cm/MHz1.1) 0 0.38 8.0 548 Monkey Y Monkey L HUmax 3300 3000 HUtissue 1035 1031 549

FIGURE CAPTIONS:

550

Figure 1: The ultrasonic transducer was attached to a neuro-navigation module that

551

allows co-registration of digitized external coordinate systems with the internal MRI

552

coordinate system. This system makes it possible to direct an attached transducer in

553

real-time to functional and anatomical target regions of the brain (Rogue Research).

554

Using this co-registration, the cigar-shaped focal spot of of FUS was placed over the left

555

FEF (red spot). For each animal the FEF was identified with anatomical landmarks

556

(arcuate sulcus AC, and principal sulcus PS). A coupling cone filled with water ensured

557

ultrasonic coupling between the transducer and the animal’s head. The

558

neurophysiological recordings were performed in the left SEF area (yellow spot) through

559

a recording chamber and grid adaptor.

560 561

Figure 2: Peak negative pressure (MPa) in the brains of monkey Y (left) and L (right)

562

from simulations, taking into account the respective voltage applied to the ultrasonic

563

transducer (139±26 V and 56±15 V).

564 565

Figure 3. Average recorded spikes per seconds with (red lines) or without (blue lines)

566

single pulse TUS. The two columns represent two examples of neurons showing

567

significant change of activity following FUS. Vertical blue shaded area represents the

568

time interval of FUS. Origin of time corresponds to the appearance of the visual

569

stimulus.

(a) Example from a neuron showing an increase of activity with FUS. Triangles

571

represent ranked saccadic latencies for each trial respectively with (red) or without (blue)

572

TUS. (b) Example from a single neuron showing a decrease of activity with FUS while

573

the monkey is preparing a saccade opposite to the target. (c and d) Bootstrap difference

574

of the control vs TUS condition (grey). Significant differences between FUS and control

575

trails are respectively represented by a red (FUS < Control trials) or green (FUS >

576

Control trials) line (p<0.01). These two neurons were selected for illustration because

577

the TUS elicited activation of diverse neuronal elements.

578 579

Figure 4. Modulation while comparing TUS and control trials (p< 0.01) in the population

580

of recorded neurons in the two animals. Significant differences between FUS and control

581

trails are respectively represented by a red (FUS < Control trials) or green (FUS >

582

Control trials) line (p<0.01) for leftward (left panel) or rightward (right panel) direction of

583

eye movement. Vertical gray shaded area represents the time interval of TUS. Origin of

584

time corresponds to the appearance of the visual stimulus. Top panel population of 46

585

for Monkey L. Middle panel: 39 neurons respectively for Monkey Y. Bottom panel of 20

586

neurons while TUS is being applied to the control region. Note that the significant

587

periods of neuro-modulation are drastically reduced (if present at all) during the control

588

condition. All isolated and recorded neurons were analyzed and no selection was made

589

to avoid sampling bias due to cell types.

590 591

Figure 5. Spike waveforms with and without FUS. A) All spike waveforms from a

592

representative session without stimulation (blue thick line —) B) median ± CI 99% (blue

593

thin line —). C) All spike waveforms from a representative session with focused

594

ultrasound stimulation (red thick line —) B) median ± CI 99% (red thin line —).

595 596

Figure 6. Height / Width ratio of spike waveforms with and without FUS for all neurons

597

recorded in monkey L and monkey Y respectively. Note that none of the recording show

598

a significant deviation while comparing FUS and no FUS trials (p >0.6).

599 600

x4

x2

4.7 mm

33 mm

5 mm

10 mm

Coupling cone

Ultrasound

transducer

-3

dB

focal spot

Supporting

arm

AS

PS

Microelectrode

recording sites

Neuronavigator

markers

Upper view of

the Monkey

Head

Monkey Y

Monkey L

a) b)

c) d)

Example recording 1

stim. OFF stim. ON stim. Off

Example recording 2

0.1

0.3

0.5

0

20

Monkey Y

Monkey L

Control

# of N

eur

ons

Time from Stimulation (s)

-0.1

Lateralized fields of movement

0

46

0

39

0.0

0.2

0.4

-0.1 0.0

0.1

0.2

0.3

0.4

0.5

a)

b)

c)

Time (µs)

Time (µs)

V

oltage (µv)

V

oltage (µv)

a)

c)

b)

d)

0 1 2 3 4 5 6 7 8 9 10 11 0 1 2 3 4 5 6 7 8 9 10 11 0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16

Monkey L

Monkey Y

Height/Width ratio without FUS

H ei g h t/ W id th r a ti o u n d er FU S