Research ArticleCOGNITIVE NEUROSCIENCE

Remote, brain region–specific control of choice behavior with ultrasonic waves

See allHide authors and affiliations

Science Advances  20 May 2020:
Vol. 6, no. 21, eaaz4193
DOI: 10.1126/sciadv.aaz4193
  • Fig. 1 Choice task and stimulus.

    (A) Task. The subject fixates a central target. One target appears in the left or the right visual hemifield. After a brief random delay, a second target appears in the opposite hemifield. The subject receives a reward for looking at either target (monkey A) or at the first target (monkey B). Ultrasound is applied in blocks of three to six trials, strictly interleaved with no stimulation blocks of the same duration, 100 ms before the appearance of the first target. (B) Functional characterization of the visuomotor system. We delivered the ultrasound noninvasively (intact skull and skin) into the frontal eye fields (FEF). From anatomical and functional studies, it is known that left/right FEF preferentially represents targets in the right/left visual hemifield. The outline of the FEF was rendered using the Calabrese et al. (50) atlas with Paxinos brain regions. (C) Effects in previous pharmacological neuromodulation studies. When a large amount of strong inhibitory/disinhibitory drugs is injected into the left FEF in this task, animals show a strong ipsilateral/contralateral bias in this task [reproduced from (30), with permission]. These results are as expected given the contralateral nature of the visual hemifield representation and are analogous when other nodes of the visuomotor network, such as the parietal area LIP, are perturbed. (D) Stimulus. The ultrasound stimulus (0.6 MPa, 270 kHz, 300 ms duration) was pulsed at 500 Hz with 1-ms tone burst duration. The ultrasound was applied through a coupling cone filled with agar gel. The resulting pressure, measured in free field, is provided along the lateral (1-mm steps) and axial (2-mm steps) dimensions in color.

  • Fig. 2 Effects of ultrasonic stimulation on choice behavior.

    (A) Single-session examples. Mean (± SEM) proportion of choices of the rightward target as a function of the difference in target onset times. Positive difference stands for the cases in which the rightward target appeared first. The black data points reflect choice behavior in the trials in which the animal was not stimulated, whereas the blue data points represent choice behavior in the stimulated trials. The data were fit with a four-parameter sigmoid function (see Materials and Methods). The data are presented separately for left and right FEF stimulation sessions (left and right columns; see illustration on top) and for monkeys A and B (top and bottom rows). (B) Quantification of the effects for all sessions. No stimulation data were fit with the sigmoid function [black curves in (A)]. Using the fit, we identified the time difference on the abscissa for which the animal chose both targets in equal proportion (see Materials and Methods). At that point, we then assessed the proportion of rightward choices during the stimulated trials. This analysis was performed for each individual session (dark bars) and for data pooled over the sessions (light bars). The number of sessions is provided in the inset. Asterisks indicate the cases in which the mean effects statistically differ from equal preference [two-sided t test for individual sessions (dark bars) and randomization test for data pooled across sessions (light bars)]. *P <0.05, **P <0.01, ***P <0.001. The illustration on the bottom summarizes the polarity of the biasing effects.

  • Fig. 3 The effects are stronger for initial stimuli during successive stimulation.

    (A) Mean ± SEM choices of the contralateral target as a function of trial number within a block, separately for nonstimulated (black) and stimulated (blue) blocks of trials. Data were pooled across left and right FEF stimulation sites. The asterisks indicate effect significance (two-sided two-sample proportion test; *P<0.05, **P <0.01, ***P<0.001). (B) Same format as in Fig. 2B using data of the first three trials of each stimulation block.

  • Fig. 4 Biasing nature of the effects.

    (A) The amount of horizontal shift of the decision curves by ultrasound (blue versus black in Fig. 2A). The data are presented separately for left and right FEF stimulation sessions (left and right panels) and separately for each session in each animal (effect histogram). The orange (brown) counts correspond to sessions of monkey A (B) and the counts are cumulative (no overlap of the orange and brown histograms). The asterisks indicate the effect significance (two-sided t test; **P <0.01, ***P <0.001). (B) The table shows the number of sessions (left columns) in which the parameters fitted to each decision curve (rows) changed significantly during ultrasound stimulation. See Materials and Methods for details of the statistical test. The right column shows the average magnitude of the change over the significant sessions. Horizontal position is measured in milliseconds and slope is measured in per millisecond.

  • Fig. 5 Stimulation of motor cortex had no effect on choice behavior.

    Same format as in Figs. 2 and 4, for ultrasound stimulation of left motor cortex (left) and right motor cortex (right).

  • Fig. 6 Decision tasks are sensitive to neuromodulation.

    Decision-making renders a subject’s decision system in a delicate balance between two alternatives. A relatively small perturbation (small filled circle) of the neuronal circuitry representing an alternative has a strong leverage on the subject’s decision for that alternative.

Supplementary Materials

  • Supplementary Materials

    Remote, brain region–specific control of choice behavior with ultrasonic waves

    Jan Kubanek, Julian Brown, Patrick Ye, Kim Butts Pauly, Tirin Moore, William Newsome

    Download Supplement

    This PDF file includes:

    • Figs. S1 to S3

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article