(L-446) Wireless neuromodulation at submillimeter precision via a microwave split-ring resonator

Electromagnetic waves have long been used to non-invasively modulate various biological systems. For example, transcranial magnetic stimulation (TMS) has successfully reached the deep brain to treat Parkinson’s Disease, depression, and epilepsy. However, due to the long wavelength of the electromagnetic waves employed, TMS offers poor spatial resolution of a few centimeters. Photons have sub-micron wavelength and provide single-cell modulation, but the strong tissue scattering prevents photons from noninvasively reaching deep tissue. Microwaves fill the gap between optical waves and magnetic waves yet have rarely been explored for neuromodulation. Microwaves have been known to provide >50 mm penetration depth into the human brain noninvasively, while maintaining more than 50% of their energy. Microwave wavelengths are also much shorter than those of magnetic waves, promising higher spatial resolution. Reports of non-thermal microwave neuromodulation date back to the 1970s, where a reversible reduction in the firing rate of aplysia pacemaker neurons was observed under low intensity microwave. Here, we report minimally invasive microwave neuromodulation at an unprecedented spatial resolution by taking advantage of an implantable split-ring resonator (SRR) design. The SRR has a perimeter of approximately one half of the resonant microwave wavelength, thus acting as a resonant antenna. It couples the microwave wirelessly and concentrates it at the gap, producing a localized electrical field. The device allows for neuromodulation with a submillimeter spatial resolution, beyond the microwave diffraction limit, while using a total dosage of ~500 J/kg, over 7 times lower than the threshold for safe microwave exposure.

Rat primary cortical neurons were cultured in vitro and dyed with Oregon green dye for calcium imaging. Experiments were carried out at 27-30°C. To confirm nonthermal microwave inhibition in mammalian neurons, microwave was applied directly to the cells for 3s at a power density of 2 W/cm². The temperature change in the medium was measured using an optical fiber thermal probe. The temperature change was recreated in the absence of microwave using a laser shone on a carbon absorber to heat cells. To demonstrate more efficient inhibition with the SRR, the SRR was placed perpendicularly over the cells with its gap ~55 um from the cells. Microwave was applied to the cells at 0.8 W/cm2 for 1s. To prolong the treatment but maintain the dosage, the microwave was pulsed at 10 Hz repetition rate with 10 ms pulse width for 10s. This was repeated at power densities ranging from 0.1-2 W/cm2 to demonstrate the dependence of inhibition efficiency on power density. Finally, to demonstrate application in an in vivo epilepsy model, seizure was induced in a mouse via cortical injection of picrotoxin. The SRR was placed flat on the cortex at the injection site and electrical measurements were taken near the gap. Pulsed microwave was applied for 10s at a power density of 0.3 W/cm2 to suppress the seizure. The experiment was repeated in the absence of the SRR as a control.

Application of 2 W/cm² microwave for 3s to neurons resulted in a 37% reduction in activity immediately following treatment. Reduced activity was observed for up to 20s after microwave treatment, after which activity returned to baseline levels. A temperature change of 0.35 C was measured. When the temperature change was reproduced on cells independently of microwaves, no significant change in activity was observed. These results suggest that microwave can inhibit neuronal activity via a nonthermal mechanism.

When the SRR was added above the cells and microwave applied at 2 W/cm² for 1s, no significant inhibition was observed. However, when the treatment was prolonged by pulsing the microwave to achieve an equivalent dosage to the 1s continuous wave, activity was reduced by 73%. The microwave dosage is notably 3 times smaller than the dosage required for microwave inhibition without the SRR. Cells more than 1 mm from the SRR gap were not significantly affected. Pulsed microwave treatment with the SRR was repeated at power densities of 2 W/cm², 1.3 W/cm², 0.8 W/cm², 0.5 W/cm², 0.3 W/cm², and 0.1 W/cm². As power density decreased, inhibition efficiency decreased, with a lower threshold of 0.1 W/cm². These results suggest that pulsing the microwave in conjunction with using the SRR drastically increases inhibition efficiency. The efficiency of microwave inhibition is dependent on power density. The microwave SRR also achieves submillimeter spatial precision.

To demonstrate effectiveness of the microwave SRR in vivo, the SRR was placed on the cortex of a mouse brain in which seizure had been induced. Microwave treatment consisted of 0.3 W/cm² pulsed microwave for 10s. The periods 1 min before and 1 min after microwave treatment were analyzed for changes in spike amplitude and inter-spike interval (ISI). On average, spike amplitude was reduced by 16%. ISI was increased by 50%. Histological assessment of H&E-stained brain slices show that the microwave used did not cause damage to the tissue. These results indicate that the microwave SRR can significantly inhibit seizure activity in an in vivo model of epilepsy by decreasing spike amplitude and frequency.