I. Core Research Summary

• Three-in-one integrated neural interface

The single-probe partitioning platform achieves three core functions: ① site‑specific aerosol printing of the NT‑PEI‑ChR2 non‑viral gene delivery vector, enabling localized neuronal transfection via electroporation; ② simultaneous aerosol printing of GelMA encapsulating UCNP upconversion nanoparticles, with 980 nm near‑infrared light penetrating the skull to convert into blue light for activating light‑sensitive channels; and ③ a porous AuIO three‑dimensional architecture that dramatically increases the electrode’s specific surface area, concurrently recording neuronal action potentials—thus completing a closed loop of transfection–stimulation–recording in a single device.

• Plasmonic Resonance–Enhanced Wireless Optogenetics

The 3D porous gold scaffold generates localized surface plasmon resonance (LSPR), enhancing the near-infrared-to-blue light emission intensity by approximately twofold. Low-power NIR irradiation can effectively activate ChR2, with photothermal heating below 1°C, thereby preventing thermal damage to brain tissue and eliminating the need for intracranial fiber‑optic implantation.

• Targeted and Controllable Non-Viral Gene Delivery

Neuropeptide (NT)-modified PEI carriers enable neuron‑specific targeting; the porous AuIO structure can encapsulate gene complexes and release them in an electric‑pulse‑controlled manner, confining expression of light‑sensitive proteins to the electrode vicinity without widespread off‑target effects, thereby avoiding the immune responses and integration risks associated with viral vectors, with a transfection efficiency as high as 78%.

• High-fidelity in vivo deep-brain recording

The impedance of the three-dimensional porous electrode is significantly reduced, and in vivo experiments in the hippocampal dentate gyrus reliably distinguish neuronal firing patterns across multiple groups. Near-infrared pulse–synchronized stimulation induces phase‑locked neural action potentials with an excellent signal-to-noise ratio.

• High-precision partitioned forming of aerosols

Micrometer‑scale, site‑specific deposition of two functional bioinks enables the simultaneous printing of a gene carrier and an upconversion luminescent layer in separate channels of the same electrode array, eliminating the need for photolithographic masks and allowing rapid customization of neural probes with distinct functional zones.

II. Core Operating Principle of Aerosol Spraying and Its Unique Process Advantages

(1) Complete Multifunctional Electrode Printing Process

• A colloidal crystal PS template is fabricated on a microelectrode substrate, and after electrodeposition of gold, the template is removed to obtain a porous 3D AuIO structure.
• Prepare GelMA‑encapsulated UCNP upconversion hydrogel ink and NT‑PEI‑ChR2 gene‑based bioink.
• The aerosol jet system uses a CNC‑controlled trajectory to define the dispensing channel, precisely depositing UCNP hydrogel onto electrodes #2 and #3, where it is UV‑induced in situ crosslinked and immobilized.
• Switch to bio‑ink and aerosol‑print the gene nanocomposite in channels #1 and #4, with room‑temperature adsorption–based loading.
• After probe encapsulation, the device was stereotactically implanted into the hippocampus of mice, followed by electroporation-mediated transfection, near-infrared–based wireless control in vitro, and simultaneous recording of neural electrical signals.

(II) Five Core Process Advantages (Compared with Lithography, Spin Coating, Drop‑Coating, and Inkjet Printing)

• Zone-specific, site‑directed, precise deposition with no cross‑contamination among multiple biomaterials.

It is possible to print the light‑conversion layer or the gene carrier independently in a designated microelectrode channel, with no ink bleed into adjacent channels, thereby achieving perfect functional partitioning for a single probe. By contrast, drop‑coating and spin‑coating cannot achieve such precise partitioning and result in non‑selective coverage across the entire area.

• Biological deposition at mild, ambient temperatures preserves the integrity of genes and the activity of nanoparticles.

The entire process involves no high-temperature sintering or strong-acid etching, ensuring that the hydrogel, plasmid DNA, and upconversion nanoparticles retain their structural integrity and full biological activity, making them compatible with a wide range of medical‑grade bioinks.

• Micrometer-scale patterning compatible with miniature neural electrodes

The nozzle precisely focuses on 15–30 μm microelectrode sites, compatible with brain‑implantable microprobes, and forms a thin, conformal coating on a porous AuIO scaffold to maximize payload capacity.

• Maskless Digital Rapid Iteration

CAD’s free‑form adjustment allows users to define the print area and deposition thickness, while a versatile probe array can be prototyped in just a few hours, eliminating the need for photolithography mask fabrication and mold‑making lead times.

• Multi-type bioink compatibility switching

Two types of biomaterials with vastly different viscosities—hydrogels and gene–nanocomposites—enable rapid replacement of the atomization chamber, allowing a single system to perform multifunctional probe modification without the need for multiple coating platforms.

III. Five Core Technological Innovations

Innovation 1: Aerosol-based dual-function bio‑ink–enabled integrated neural probe with partitioned printing

Innovation 2: A 3D AuIO plasmonic-enhanced near-infrared wireless optogenetic system

Innovative 3: A Neural-Targeted, Controllable Electrical-Release Non-Viral Delivery Platform

Innovation 4: Synergistic Enhancement via Porous Three-Dimensional Electrodes Combined with Aerosol-Based Biological Modification

Innovative Five-Component Extracorporeal Near-Infrared Non-Invasive Closed-Loop Modulation Scheme for Deep Brain Stimulation

IV. Existing Process Challenges and Corresponding Solutions for Overcoming Them

Challenge 1: Drop‑coating and spin‑coating cannot achieve full‑area coverage, nor can they enable probe‑based functionalization of distinct regions.

Challenge 2: High temperature and strong chemical treatment degrade plasmid DNA and upconversion nanoparticles.

Challenge 3: Conventional optogenetics requires the implantation of optical fibers, leading to long-term damage to brain tissue and glial scarring.

Challenge 4: Viral vectors pose immune risks and exhibit widespread off-target effects, limiting their clinical translation.

Challenge 5: Planar electrodes have a low specific surface area, resulting in a poor signal-to-noise ratio in neural recordings.

V. Future Diversified Industrial Application Scenarios

1. Closed-loop deep brain stimulation device for mental disorders

2. Basic Research Platform for Neuroscience in Higher Education Institutions

3. Minimally Invasive Neurointerventional Medical Devices for the Human Brain

4. High-throughput screening chip for brain organoids and brain slices

5. Flexible Peripheral Nerve Modulation Implantable Device

Scheme 1. Overview of a multifunctional three-dimensional AuIO electrode array with in vivo closed-loop neural modulation capabilities. This system integrates three core functionalities: (1) targeted gene delivery; (2) optogenetic stimulation; and (3) neural signal recording, all consolidated on a single platform. (1) The three-dimensional AuIO architecture is fabricated via template-assisted electrodeposition, leveraging its high specific surface area to immobilize NT‑PEI complexes carrying the ChR2 plasmid on an AAV scaffold. (2) UCNPs are embedded within a GelMA hydrogel matrix and deposited onto the electrode surface using inkjet printing, enabling the conversion of near-infrared excitation into visible light for remote activation of ChR2. (3) Neural signals can be recorded in real time through the same electrode array, facilitating the implementation of a closed-loop interface. Furthermore, the AuIO nanostructure enhances the optical output of the UCNPs via localized surface plasmon resonance, thereby improving the efficiency of optogenetic stimulation. Collectively, these components form the foundational functional core of next-generation closed-loop neural modulation systems.

Figure 1. (a) Representative scanning electron microscopy images showing the spherical morphology and uniform distribution of the NT‑PEI‑ChR2 nanocomposite. This composite is formed via self-assembly between NT‑PEI and ChR2 plasmid DNA, with a nanoscale size suitable for cellular uptake. (b) Zeta potential values of NT‑PEI‑ChR2, PEI‑ChR2, and ChR2. (c) Cross-sectional SEM images of electrodes on ITO substrates bearing three-dimensional AuIO structures, before and after loading with the NT‑PEI‑ChR2 nanocomposite: the porous framework (bottom, “before loading”) features interconnected opaline pores (~800 nm) capable of accommodating gene‑delivery complexes; after loading (top, “after loading”), the pore walls are coated with a polymer‑rich layered structure exhibiting a tentacle‑like morphology, indicating successful penetration and attachment of NT‑PEI‑ChR2 within the 3D scaffold. (d) Cumulative release profiles of the NT‑PEI‑ChR2 gene in planar gold films and three-dimensional AuIO structures, with or without continuous electrical stimulation at 100 μA (frequency 1 Hz, pulse width 250 ms): “Planar‑E+” and “Planar‑E−” denote planar gold films under stimulated and unstimulated conditions, respectively; “AuIO‑E+” and “AuIO‑E−” refer to the 3D AuIO structures under stimulated and unstimulated conditions, respectively. (e) Cumulative release curves for the NT‑PEI‑ChR2 gene loaded into the 3D AuIO structure, subjected to continuous electrical stimulation at 1 Hz, 250 ms pulse width, and varying current intensities (0, 50, 100, and 200 μA). (f) Stepwise cumulative release profiles of the NT‑PEI‑ChR2 gene loaded into the 3D AuIO structure under cyclic electrical stimulation. At a frequency of 1 Hz and a pulse width of 250 ms, alternating biphasic currents of 0 μA and 100 μA were applied for 5 minutes, followed by a 15‑minute rest period. The stimulation periods are indicated by yellow bars. The AuIO‑E− group represents the control group that received no electrical stimulation.

Figure 2. Local electroporation–mediated gene delivery using a three-dimensional AuIO‑modified MEA. (a) Photograph and scanning electron microscopy image of the MEA chip, showing the uniform integration of the three-dimensional AuIO nanostructure on the surface of each microelectrode. (b) Left panel: Fluorescence microscopy images confirm site‑specific mCherry expression between paired microelectrodes within the orange‑colored regions; right panel: Representative images of neuronal cells 48 hours after transfection under four distinct transfection conditions (PEIChR2‑E−, NT‑PEI‑ChR2‑E−, PEI‑ChR2‑E+, and NT‑PEI‑ChR2‑E+). (c) Quantitative analysis of transfection efficiency by flow cytometry: The NT‑modified complex, in combination with electrical stimulation, synergistically enhances local gene delivery, with the NT‑PEI‑ChR2‑E+ group exhibiting the highest efficiency (n = 3, mean ± standard deviation).

Figure 3. Upconversion blue emission from GelMA‑based UCNPs printed onto 3D AuIO‑modified microelectrodes under NIR excitation. (a) Composite optical and fluorescence images of the neural electrode array under different NIR illumination conditions: the left panel shows a bright-field image of GelMA‑based UCNPs printed on 3D AuIO microelectrodes via aerosol jet printing with NIR off; no significant upconversion emission is observed. The middle panel depicts pronounced localized blue upconversion emission (400–500 nm) from the UCNPs printed on the 3D AuIO microelectrodes under 50 mW of 980 nm NIR illumination. The right panel presents a fluorescence emission spectrum, which, when compared with the underlying microelectrode structure, confirms that the blue luminescence originates specifically from the UCNPs at the microelectrode sites and retains strong spatial localization. (b) Emission spectra measured with a miniature Raman system coupled to a CCD spectrometer for various microelectrode configurations under 50 mW of 980 nm NIR excitation: UCNP @ 3D AuIO microelectrodes (blue line) exhibit markedly enhanced blue emission (450–500 nm) relative to UCNP @ Au microelectrodes (red line), whereas the bare Au microelectrodes (black line) show virtually no emission in this spectral range. (c) Quantitative analysis of NIR-to-blue emission intensity: the integrated blue emission power density (450–500 nm) for UCNP @ 3D AuIO microelectrodes reaches 5.81 mW/mm², significantly higher than the 2.93 mW/mm² observed for UCNP @ Au microelectrodes. This performance enhancement is attributed to the LSPR effect arising from the three-dimensional nanostructured AuIO architecture. Data are presented as mean ± standard deviation. Note: Bare gold microelectrodes were excluded from the bar graph because their blue emission was negligible under the measurement conditions.

Figure 4. Ex vivo neuronal activation in acute hippocampal slices induced by NIR light stimulation on an MEA. (a) Schematic of the experimental setup, illustrating remote optical stimulation of ChR2‑mCherry–expressing regions in the dentate gyrus (DG) of the hippocampus using an MEA platform functionalized with UCNPs@3D AuIO. (b) Combined fluorescence microscopy and DIC imaging of an acute cortical–hippocampal slice mounted on the MEA device. The red dashed box highlights the MEA region corresponding to the area marked in panel (c); the ChR2‑mCherry–expressing tissue is outlined by a yellow dotted line, encompassing five microelectrodes modified with UCNPs@3D AuIO—these electrodes convert NIR light into blue light via the UCNPs to enable remote optogenetic stimulation. (c1–c3) Neuronal activation signals were recorded from the five modified microelectrodes positioned beneath the ChR2‑mCherry–expressing tissue: upon NIR illumination, the UCNPs emit blue light that stimulates ChR2‑expressing hippocampal neurons, thereby eliciting MUA discharges. In each panel, the yellow dotted boxes correspond to the same tissue region shown in (b). The sequence of spike potentials presented in the figure, together with quantitative analysis, confirms that the 3D AuIO‑functionalized UCNPs integrated onto the MEA can achieve efficient light‑induced neuronal stimulation under ex vivo conditions, validating their potential for future in vivo optogenetic applications.

Figure 5. In vivo remote NIR light stimulation can elicit neural activity in the dentate gyrus (DG) of the hippocampus. (a) Schematic diagram of the neural electrode array configuration and its corresponding surgical implantation site, enabling targeted opsin gene delivery followed by NIR‑responsive optogenetic stimulation in the mouse dentate gyrus. This 16‑channel neural electrode array was sequentially functionalized via aerosol jet printing: UCNPs encapsulated within GelMA were deposited onto channels Ch#2 and Ch#3 for optical signal transduction, while PEI−NT−ChR2‑mCherry nanocomposites were locally coated onto channels Ch#1 and Ch#4 to facilitate electroporation‑mediated gene transfection. This design allows spatially confined gene delivery through a single surgical procedure and enables fiberless intracranial optogenetic activation using 980 nm NIR illumination. (b) Confocal fluorescence imaging reveals localized ChR2‑mCherry expression within the hippocampal dentate gyrus seven days after electroporation; immunohistochemical analysis confirms this: DAPI (blue) labels cell nuclei, and mCherry (red) indicates successful opsin expression. Expression is concentrated in the probe‑track region adjacent to channels Ch#1 and Ch#4, with partial signals also observed in the CA1 region, reflecting spatially restricted gene delivery consistent with electric field distribution and the orientation of the neural electrode array. A coronal brain section at AP−2 mm relative to the bregma point verifies the in situ positioning of the implanted neural electrode array. The magnified fluorescence image (yellow dashed box) clearly shows robust, specific ChR2‑mCherry expression in the DG, whereas the wide‑field image demonstrates its broader distribution among hippocampal pyramidal neurons. Although uniform transfection across the entire dentate gyrus was not achieved, this localized expression makes functionally relevant cell‑specific optogenetic stimulation within the hippocampus feasible. (c) During remote optogenetic stimulation with 980 nm NIR light, electrophysiological recordings were acquired using the implanted neural electrode array. (c1) Raw spike train data from channel #1 during eight NIR stimulation trials (each lasting 1 second) reveal a marked increase in neuronal firing activity during each light‑exposure period (highlighted in yellow), demonstrating both the successful implementation of optogenetic activation and the dual utility of channel #1 in electroporation‑mediated gene delivery and subsequent neural spike recording. (c2) The corresponding PSTH plot shows a significant enhancement of multiunit activity (MUA) within the stimulation window, confirming that neurons exhibit time‑locked responses to NIR‑driven blue light emission. (c3) Spike classification analysis identifies six distinct hippocampal neurons contributing to the recorded MUA signals, further validating the specificity and efficacy of the UCNP@AuIO‑based fiberless intracranial optogenetic interface in a living organism.

Figure 6. Schematic diagram of NT‑PEI‑ChR2 synthesis. First, NT is reacted with SPDP to introduce a pyridyl disulfide group; the modified NT is then coupled to PEI using Traut’s reagent to form NT‑PEI, which is verified by monitoring the release of pyridine‑2‑thione (as indicated by absorbance at 343 nm); finally, NT‑PEI is complexed with ChR2 plasmid DNA.