Ultrasound beams activate neurons in the eye to help restore vision

22 Mar 2022 Tami Freeman

The ultrasound stimulation system: Acoustic waves targeted at the retina excite the retinal neurons, which generate neural signals that are transmitted through the optic nerve to the brain. Brain activity is recorded using a multielectrode array. (Courtesy: CC BY 4.0/BME Frontiers 10.34133/2022/9829316)

Retinal degeneration, in which the light-sensitive photoreceptors in the eye deteriorate and lose function, is a major cause of blindness worldwide. But even though the retinal cells have lost sensitivity to light, the neural circuits connected to the brain are often well preserved. This provides an opportunity to restore vision by bypassing the damaged photoreceptors and directly stimulating the retinal neurons.

Visual prostheses that restore sight via electrical stimulation of retinal neurons have already been developed and used successfully in patients. But such devices are invasive and require complex implantation surgeries. Instead, researchers at the University of Southern California propose the use of non-invasive ultrasound to activate these neurons. Reporting their findings in BME Frontiers, they demonstrate that stimulating a blind rat’s eyes with ultrasound activated small groups of neurons in the animal’s eye.

“This is a step towards a non-invasive retinal prosthesis that works without invasive eye surgeries,” says corresponding author Qifa Zhou in a press statement. “Special glasses with a camera and an ultrasound transducer are intended to give blind and partially sighted people a new view of the world.”The USC researchers (from left to right): Gengxi Lu, Xuejun Qian, Mark Humayun, Qifa Zhou and Biju Thomas. (Courtesy: Department of Ophthalmology, USC)

Proof-of-concept

The rationale for using ultrasound is that the sound waves exert mechanical pressure on neurons in the eye, activating them to send signals to the brain. Zhou and colleagues first examined the impact of ultrasound stimulation in normal-sighted rats, using a 3.1 MHz ultrasound transducer with a focal depth of 10 mm. They recorded neuron activity using a 32-channel electrode array (150 μm spacing) inserted into the animal’s visual cortex or superior colliculus (SC), the brain area connected directly to the optic nerve.

The researchers measured the rats’ responses to light and ultrasound stimulation, observing comparable neuron activities from both stimuli. These findings in normal-sighted rats indicated that ultrasound can provide an alternative way to stimulate the retina.

Next, the team investigated a rat model of retinal degenerative blindness. Neural signals recorded in 16 blind rats showed that ultrasound could stimulate the retinal neurons. As expected, no light-evoked visual activity was achieved. The induced neuron activities in blind rats were generally weaker in amplitude and duration than those observed in normal-sighted rats.

In contrast to light-evoked neuron activity, the response to ultrasound stimulation can be modified by the beam parameters. The researchers assessed the effects of varying the ultrasound intensity (acoustic pressures from 1.29 to 3.37 MPa) and duration (from 1 to 200 ms). They observed that both the amplitude and duration of neuron activity increased with increasing ultrasound intensity. The amplitude did not vary for ultrasound durations of 10 ms or longer, but the response duration did increase with increased ultrasound duration.

Towards clinical transfer

One important requirement for a visual prosthesis is that its user can see sharp images. To quantify the spatial resolution of ultrasound-evoked neuron activity, the researchers reconstructed the response across the SC surface while moving the transducer. Activated SC regions had spatial resolutions ranging from 161 to 299 μm, with this variation likely due to the curvature of the retina. The average spatial resolution was 250 μm – comparable to that of the first FDA-approved retinal prosthesis, the Argus II.

The temporal resolution (frame rate) of ultrasound stimulation is another important factor, as it determines whether a prosthesis can provide smooth vision of moving objects. Consecutive 20 s stimulations delivered at various frame rates revealed that frame rates of up to 5 Hz generated stable neuron activity, while a higher frame rate (10 Hz) could potentially suppress the firing neurons. They note that this suppression was caused by neuron saturation rather than neuron damage.Pattern generation: A C-shaped beam is projected onto the retina (a) and a 56-channel MEA placed over the entire SC surface (b). The hydrophone-measured acoustic field (c) and the MEA-recorded neuron activities (d) show the C-shaped pattern. (Courtesy: CC BY 4.0/BME Frontiers 10.34133/2022/9829316)

Finally, to validate the technique’s ability to generate visual patterns, the researchers designed a 4.4 MHz helical transducer that projects ultrasound onto the retina in the shape of the letter “C”. Using a 56-channel electrode array placed over the entire SC region, they observed the same C-shaped pattern of neuron activities in the SC.

“These results represent a step towards non-invasive retinal prosthesis development using ultrasound,” conclude co-first authors Xuejun Qian and Gengxi Lu. “The in vivo demonstration of visual restoration in blind rats suggested that ultrasound opens a new avenue for the development of a novel non-invasive retinal prosthesis.”READ MORE

Connecting the dots to artificially restore vision

“For the next step, we are working on several deeper investigations,” Lu tells Physics World. These include ultrasound stimulation with a higher centre frequency to provide better spatial resolution and lower stimulation threshold; use of a 2D ultrasound array to dynamically generate different stimulation patterns; and behaviour tests to show how an awake animal responds to the ultrasound stimulation, in addition to the neuron recording.

Texas-based Nanoscope Technologies plans to license this patent-pending ultrasound stimulation technique and provide support for future experiments. If these studies are successful, the team predicts that the technology could be translated into human clinical trials within the next three to five years.

Tami Freeman is an online editor for Physics World

FROM PHYSICSWORLD.COM   4/4/2022