Contact Lenses That Treat Depression Perform as Well as Prozac in Mouse Trial

South Korean scientists have developed a smart contact lens capable of delivering electrical brain stimulation through the retina. In preclinical testing, it performed on par with a leading antidepressant drug. The implications for eye care professionals could be significant.

A research team from Yonsei University in Seoul has published findings in the Cell Press journal Cell Reports Physical Science describing a wearable contact lens platform that delivers targeted electrical stimulation to mood-regulating brain regions via the eye's own neural circuitry and does so without surgery, pharmaceuticals, or bulky equipment.

The technology, called temporal-interference-based transcorneal electrical stimulation (TI-TES), represents a significant departure from how both smart contact lenses and depression treatments have traditionally been conceived. While previous smart lens research has focused on ocular diagnostics measuring intraocular pressure, tear glucose, and similar biomarkers. This is believed to be the first time a contact lens has been used to treat a brain disorder.

How it works

The lens uses a principle called temporal interference, in which two high-frequency electrical signals (in this study, 2,000 Hz and 2,020 Hz) are delivered simultaneously from electrodes embedded in the lens. Each signal alone passes through ocular tissue with minimal neural effect, but where the two intersect, precisely at the retina, they generate a low-frequency envelope wave (20 Hz) capable of activating retinal ganglion cells (RGCs) and intrinsically photosensitive RGCs (ipRGCs).

From there, the stimulation travels along well-established retinofugal pathways to brain regions directly implicated in depression, including the hippocampus (HPC) and prefrontal cortex (PFC). Senior author Jang-Ung Park described the principle in accessible terms: "Think of two flashlights: each beam alone is dim, but where they overlap, a bright spot appears, and that bright spot can be created far away from the flashlights themselves."

To confirm the pathway was being engaged, the team used in vivo epiretinal recordings showing clear spike activity in RGCs during stimulation, and visual evoked potential (VEP) recordings demonstrating that retinal activation propagated to the visual cortex. Critically, when retinal ganglion cells were ablated using intravitreal NMDA injection, the antidepressant effect of the lens was abolished entirely providing strong causal evidence that the therapeutic effect depends on intact RGC pathways.

Electrode design: transparency and performance

For eyecare professionals, the lens construction itself is noteworthy. The electrodes are built from an ultrathin (3 nm) gallium oxide (GaOx) layer with 1 nm of platinum deposited on top, then further modified with platinum nanoclusters referred to as platinum black (PtB) on the active stimulation sites. The resulting PtB-Pt-GaOx structure maintained over 80% optical transmittance across the visible spectrum, both before and after lens moulding into a soft silicone elastomer.

Electrochemical performance was markedly improved by the nanocluster modification: impedance at 1 kHz dropped from 20.3 kΩ to 2.8 kΩ, and charge storage capacity more than doubled from 7.17 to 15.60 mC cm⁻². The charge injection capacity was calculated at 254.6 μC cm⁻², well within accepted safe limits for neural stimulation electrodes. Accelerated aging tests simulating 12 months of wear showed only minimal impedance change, supporting the platform's durability for sustained use.

Ocular safety profile

The team conducted extensive ocular safety assessments using C57BL/6 mice with intact retinal anatomy. Histological analysis (H&E staining) showed no structural changes to corneal or retinal tissue following three weeks of daily 30-minute stimulation sessions. Immunofluorescence analyses for TUNEL (apoptosis marker), CD45 (inflammatory immune cells), and Iba1 (microglial activation) were all negative. Intraocular pressure remained stable throughout, and a cell viability assay using human corneal epithelial cells (HCE-2) returned 99.25% viability statistically indistinguishable from controls.

Preclinical results: behaviour, neuroscience, and biology

The therapeutic efficacy experiments were conducted in rd1 mice (which carry photoreceptor degeneration, eliminating visual input as a confounding variable) using a corticosterone-induced depression model. Mice were divided into sham, depressed (Dep), TI-TES-treated, and fluoxetine-treated groups, with TI-TES delivered at the optimised parameters of 20 Hz, 200 mV peak-to-peak, for 30 minutes daily over three weeks.

Behavioural outcomes were striking. Compared to the untreated depressed group, TI-TES animals showed a 76% increase in locomotor distance and a 132% increase in time spent in the centre zone during open-field testing. Immobility in tail-suspension and forced-swim tests, measures of behavioural despair, fell by 48% and 45% respectively. Social interaction scores, measured by a three-chamber preference paradigm, recovered to near-sham levels. Across all three behavioural assays, TI-TES outcomes were directly comparable to fluoxetine (Prozac).

At the neural level, the treated group showed markedly restored HPC-PFC functional connectivity. Phase-locking values in the theta band increased by 163.6% (HPC-PFC) compared to depressed controls, and phase coherence was 77.8% higher, consistent with prior evidence that theta-band synchrony is the primary neural signature disrupted in depression.

Biological analyses confirmed corresponding improvements: hippocampal BDNF levels recovered by 131.1% relative to untreated depressed mice; IL-6 and TNF-α mRNA expression in the hippocampus fell by 65.8% and 56.8% respectively; plasma corticosterone dropped 48.1%; and serotonin levels increased by 46.9%.

A machine-learning pipeline integrating all three data streams behavioural, electrophysiological, and biological using a support vector machine (SVM) classifier consistently grouped TI-TES-treated animals with the sham (non-depressed) cohort, clearly separated from untreated depressed controls. The model achieved a precision-recall AUC of 1.00 under cross-validation.

Implications for eyecare practitioners

The research positions the eye as a therapeutic interface in an entirely new clinical context. While the platform remains preclinical and requires substantial translational work. The current prototype uses a wired configuration, and human variability in corneal curvature, tear film, and tissue impedance will need to be addressed. The safety data will be of particular interest to optometrists and ophthalmologists likely to be involved in any future clinical pathway.

The authors acknowledge that moving to human use will require full wireless integration, personalised stimulation protocols, long-term biocompatibility studies in larger animal models, and rigorous clinical trials. They also note that the corticosterone mouse model recapitulates features of depression rather than representing the full complexity of human major depressive disorder.

"Our work opens up an entirely new frontier of treating brain disorders through the eye," said Park. "We believe this wearable, drug-free approach holds tremendous promise for transforming how depression and other brain conditions are treated, including anxiety, drug addiction, and cognitive decline."