Self-Tracking Energy Transfer for Neural Stimulation in Untethered Mice

Optical or electrical stimulation of neural circuits in mice during natural behavior is an important paradigm for studying brain function. Conventional systems for optogenetics and electrical microstimulation require tethers or large head-mounted devices that disrupt animal behavior. We report a method for wireless powering of small-scale implanted devices based on the strong localization of energy that occurs during resonant interaction between a radio-frequency cavity and intrinsic modes in mice. The system features self-tracking over a wide (16-cm diameter) operational area, and is used to demonstrate wireless activation of cortical neurons with miniaturized stimulators ($10{\mathrm{mm}}^3$, 20 mg) fully implanted under the skin.

Figure 1

Self-tracking energy transfer. (a) Schematic illustrating the operation of the system. The mouse moves freely above the lattice while the resonant cavity is excited with a continuous-wave input. (b) Magnetic-field distribution illustrating coupling between the cavity resonance and an intrinsic mode in the mouse. (b) Magnetic-field distribution 1.5 cm above the surface at 1.5 GHz. The diameter of the cavity is 21 cm.

Figure 2

Characterization of modes under plane-wave illumination. (a),(b) Energy stored in the object as a function of frequency when the incident electric field is parallel to (a) the major axis and (b) the minor axis of the object for different object sizes: major axis length 4, 5, and 6 cm (uniform scaling). (c),(d) Contour plot of the magnetic-field distribution corresponding to the peaks indicated by the arrows in (a) and (b), respectively. The field intensity is normalized to the incident field intensity $|H_0|$. Scale bar is 5 cm.

Figure 3

Demonstration of self-tracking energy transfer to dielectric object. (a) Contour plot of the field profile above the lattice for a centered and off-centered dielectric sphere (diameter of 2.8 cm). (b) Calculated $|\kappa |$ as a function of distance from the center. (c) Measured magnetic-field profile (dots) and simulation fit (solid lines) along the center axis. The magnetic field is normalized to $|H_0|$, for the field recorded in absence of the object.

Figure 4

Energy transfer to a power measurement device. (a) Schematic of the device consisting of a 2-mm-diameter coil, capacitors, diodes, an integrated circuit, and a LED. (b) Simulated magnetic-field distribution for a device implanted in the back of a mouse. Scale bar is 2 cm. (c) Photograph of mouse with device implanted subcutaneously in the back. When the ambient light is off, the LED is visible through the skin. (d) Two paths of motion over 30 s of free movement tracked by video. (e) Power harvested by the device over the two paths. The efficiency of the energy transfer is $\eta =0.75\times {10}^{-3}$ ($\pm 0.14\times {10}^{-3}$ standard deviation shaded).

Figure 5

Activation of wirelessly powered cranial electrodes induces robust c-fos expression in the regions immediately around the electrode. (a) Photograph of mouse with implanted stimulator before the skin is sutured over the device. (b) Photograph of the device. Scale bar is 2 mm. (c) Intense c-fos activation (red) in the region around electrode [labeled $e$ in the enlargement of 5], inferior to the electrode track. Scale bar is $200\mu \mathrm{m}$. Inset region at $63\times$ showing c-fos activation in more detail. Scale bar is $100\mu \mathrm{m}$. (d) c-fos activation in region with electrode-related scarring, but without electrode track. Scale bar is $200\mu \mathrm{m}$. (e) Sham electrode implantation and stimulation does not induce c-fos expression near the electrode insertion site or in (f) regions with electrode-related scarring, but no electrode track. Scale bar is $200\mu \mathrm{m}$.