Transparent, flexible graphene-ITO-based neural microelectrodes for simultaneous electrophysiology recording and calcium imaging of intracortical neural activity in freely moving mice

doi:10.21203/rs.3.rs-4948066/v1

Download PDF

Article

Transparent, flexible graphene-ITO-based neural microelectrodes for simultaneous electrophysiology recording and calcium imaging of intracortical neural activity in freely moving mice

https://doi.org/10.21203/rs.3.rs-4948066/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

To understand of the complex dynamics of neural activity in the brain across various temporal and spatial scales, it is crucial to record intracortical multimodal neural activity by combining electrophysiological recording and calcium imaging techniques. This poses significant constraints on the geometrical, mechanical and optical properties of the electrodes. Here, transparent flexible graphene-ITO-based neural microelectrodes with small feature sizes are developed and validated for simultaneous electrophysiology recording and calcium imaging in the hippocampus of freely moving mice. A micro-etching technique and an oxygen plasma pre-treating method are introduced to facilitate large-area graphene transfer and establish stable low-impedance contacts between graphene and metals, leading to the batch production of high-quality microelectrodes with interconnect widths of 10 µm and recording sites diameters of 20 µm. These electrodes exhibit appropriate impedance and sufficient transparency in the field pf view, enabling simultaneous recording of intracortical local field potentials and even action potentials along with calcium imaging in freely moving mice. Both types of electrophysiological signals are found to correlate with calcium activity. This proof-of-concept work demonstrates that transparent flexible graphene-ITO-based neural microelectrodes are promising tools for multimodal neuroscience research.

Neural activity recordings have significantly advanced in understanding brain functions, elucidating neurological diseases mechanisms, and developing effective treatments^1–3. Electrophysiological techniques allow for high-temporal resolution recording of neural electrical activity and can detect signals across various spatial scales by optimizing electrode feature-size, even for a single neuron^4–7. However, these techniques face challenges in targeting to specific cell types, resulting in lack of spatial information. Conversely, calcium imaging provides cell-type specificity via genetically or chemically engineered calcium indicators. Yet, the temporal resolution is constrained by the kinetics of calcium indicators binding to calcium ions⁸. Therefore, developing a unified tool integrating both electrophysiology and calcium imaging would be invaluable, combining the strengths of each to achieve high spatiotemporal resolution in neural activity recording.

Conventional metal-based flexible neural electrodes are suboptimal for merging electrophysiology and calcium imaging due to the inherent opacity, which obstructs both light delivery and photon collection, thus impeding the field of view (FOV)^9,10. Although metals designed with nano-network, nano-mesh, or nano-grid structures can achieve transparency^11,12, serious electrophysiological signal distortion occurs when metals were exposed to light, triggering the photoelectric effect¹³. Indium tin oxide (ITO) ¹⁴ is the most commonly used material for transparent flexible neural microelectrodes (TFNMEs), offering good conductivity¹⁵, transparency¹⁶, and biocompatibility¹⁷. Nonetheless, its mechanical stability is compromised by the brittleness^18,19. Other advanced transparent conductive materials, such as conductive hydrogels²⁰, conductive polymers^21,22, and carbon-based nanomaterials^23,24 have been recently reported to overcome issues related to obscuration and light-induced artifacts. However, their inherent properties and fabrication methods typically result in recording sites of 50 µm or larger(several times the size of a neuron)²⁵, substantially restricting the spatiotemporal resolution of neural activity recording. Consequently, miniaturizing the electrode’s feature-size to that of a single neuron is essential for capturing high-frequency electrical activity during calcium imaging.

Graphene emerges as a promising candidate for TFNMEs owing to its transparency, flexibility, artifact-free, and biocompatibility^25–27. Nonetheless, scaling down electrode feature-size entails narrower interconnects and smaller recording sites, which typically increases electrode impedance, attenuates signal quality, and lowers defects tolerance. These challenges are exacerbated by difficulties in large-scale graphene transfer and instability of graphene-metal contacts, heightening the likelihood of open-circuits failures²⁸. Ramezani et al.²⁹ developed TFNMEs featuring 20 µm recording sites by interlayer-doping double-layer graphene (id-DLG) and electrochemically depositing platinum nanoparticles (PtNPs), though the electrochemical deposition method was unsuitable for batch production. Despite the small feature-size TFNMEs achieved simultaneous multiunit activity recording and two-photon calcium imaging, they were limited to a fixed mice cortical surface. Xue et al.³⁰ proposed a tightly integrated neuronal imaging microscope (TINIscope) system that recorded local field potentials (LFPs) and calcium signals simultaneously in the hippocampus of mice using platinum-iridium electrodes assembled on the gradient refractive index (GRIN) lens. However, since the rigid electrodes were attached on the side of the lens, it remains uncertain whether the LFPs originated from the FOV, and the correlation between these potentials and the calcium signals has yet to be determined.

This work reports the development and validation of small feature-size TFNMEs based on graphene-ITO (Gr-ITO-TFNMEs) for high spatiotemporal resolution in intracortical neural activity recording. Gr-ITO-TFNMEs employ ITO as the interface material for recording sites, with metals serving as the interconnects and graphene acting as a bridge between the ITO and metal interconnects. The configuration ensures that the FOV is composed entirely of transparent materials and incorporates graphene's flexibility, ITO's biocompatibility, and superior conductivity, decreasing bending caused fatigue or damage, enhancing mechanical stability and eliminating curling. A micro-etching technique is proposed to enable crack-free, wrinkle-free, and large-area graphene transfer through multiple alignments and transfers of small-area graphene. In addition, an oxygen plasma pre-treating the graphene method facilitates the stable, low-impedance contacts between graphene and metals. These technological advancements promote the batch production and high-quality fabrication of Gr-ITO-TFNMEs with small feature-size (interconnect widths of 10 µm and recording sites diameters with 20 µm). The resulting Gr-ITO-TFNMEs exhibit an electrochemical impedance of 1.45 MΩ at 1 kHz and demonstrate an optical transmittance greater than 80% at 520 nm. Finally, the ‘watch-style’ Gr-ITO-TFNMEs are integrated onto the GRIN lens of the TINIscope system, simultaneously recording electrophysiological signals and performing calcium imaging in the hippocampus of freely moving mice. The integrated Gr-ITO-TFNMEs can stably and reliably record LFP and action potentials. The power spectral density of LFP and the number of action potentials showed correlations with calcium activity.

2.1 Design of the Gr-ITO-TFNMEs

Given the poor interfacial behavior of graphene-based TFNMEs (Gr-TFNMEs) when scaling down electrode feature-size to single-cell size and the fundamental requirement for transparency and flexibility of TFNMEs, several material adaptations was made: ITO was utilized as the interface material for the electrodes, replacing graphene; metal interconnects (Cr/Au/Cr) were employed in the backend to reduce the trace impedance; and graphene was solely used as a conductive bridge between ITO and Cr/Au/Cr. The bridging method overcomes the brittleness of ITO and optimizes conductivity without compromising transparency and flexibility. The detailed configuration is depicted in the dashed box in Fig. 1(a). The introduction of ITO allowed for Polyimide (PI) 2600 series (DuPont, HD Microsystems) to serve as both the bottom and top insulation layers of the electrodes, compatible with dry etching and offering excellent mechanical properties and transparency. The layout of Gr-ITO-TFNMEs is illustrated in Fig. 1(a). To stably integrate Gr-ITO-TFNMEs with the objective lens (GRIN lens) of the TINIscope system, the electrodes were designed like a ‘watch-style’ structure, as shown in Fig. 1(b). The ‘watch dial’ represents the FOV and was aligned with the end face of the lens, serving as the transparent window for calcium imaging. Six recording sites, each with of a diameter of 20 µm, were centrally distributed on the FOV to record neural electrical activity. The ‘watch strap’, with a width of 180 µm, was designed to bend at 90°and attach to the side of the lens to reinforce the integration. The bent area contained no ITO; instead, a PI-graphene/Au-PI configuration was employed. The graphene/Au traces were designed with a width of 10 µm. The method of bending Gr-ITO-TFNMEs to integrate with the lens is illustrated in Fig. 1(c). The outstanding flexibility of Gr-ITO-TFNMEs enables them to be seamlessly and non-destructively integrated onto GRIN lens with diameters of 1 mm or even 500 µm, as displayed in Fig. 1(d).

2.2 Fabrication of the Gr-ITO-TFNMEs

During the fabrication of TFNMEs containing graphene, two major challenges were encountered: (1) The surface tension of water complicated large-area graphene transfer. It is difficult to spread the large-area graphene out smoothly on the substrate without overlaps or wrinkles; (2) The hydrophobic nature of graphene surface hindered the formation of stable, low-impedance contacts with metals. We proposed a micro-etching technique and pre-treating the graphene with oxygen plasma to address these challenges and achieve the batch production and high-quality fabrication of Gr-ITO-TFNMEs. The micro-etching technique utilized a photoresist mask for oxygen plasma treatment to create alignment contours on the PI, enabling large-area, cracks-free, and wrinkle-free graphene transfer through multiple alignments and transfers of small-area graphene. Pre-treating the graphene with oxygen plasma hydrophilized the surface, allowing metals to be uniformly deposited and form low-impedance contacts. These treatments significantly enhanced the yield of electrode fabrication. The fabrication of Gr-ITO-TFNMEs contained the following 9 steps, as shown in Fig. 2. (1) PI was spin-coated and cured to form the 3 µm bottom insulation layer on the 4-inch silicon wafer. (2) Alignment contours (~ 100 nm) were formed on the bottom PI by oxygen plasma treatment under a positive photoresist (MICROPOST S1805 G2) mask. (3) Suspended self-help transfer few-layer graphene (XFNANO, 800 ~ 1000 Ω/□) was aligned face-to-face with the alignment contours and sequentially transferred to the target areas on the surface of deionized water. After baking at 150 ℃ for reducing wrinkles, as demonstrated in Fig. S1, the graphene was soaked in acetone to remove polymethyl methacrylate (PMMA). (4) Cr/Au/Cr with thickness of 10/100/10 nm were thermally evaporated to form metal interconnects through a lift-off process after treated by oxygen plasma. As shown in Fig. S2, this treatment facilitated the uniform and stable deposition of metals. Additionally, low-impedance contacts were created between graphene and metals, validated by measuring the impedance of short-circuited metal interconnects to graphene, found to be approximately 1 kΩ. (5) The graphene was patterned by reactive ion etching (RIE) with photoresist mask (AZ-6130). (6) A 140 nm of ITO was sputtered to form the interface of recording sites with a lift-off process. (7) A 3 µm layer of PI was spin-coated and cured as the top insulation layer. (8) The outline of Gr-ITO-TFNMEs, connection pads and openings of recording sites were defined by RIE with photoresist mask (AZ-4620). (9) The entire silicon wafer was submerged in deionized water to release the electrodes. The detailed fabrication processes were described in the Materials and methods below.

The entire process significantly enhanced the quality of electrode for batch fabrication through a micro-etching technique and oxygen plasma pre-treating, achieving the preparation of Gr-ITO-TFNMEs with small feature-size. The similar fabrication processes of Gr-TFNMEs and ITO-based TFNMEs (ITO-TFNMEs) are described in Supplementary information (Note S1 and Note S2). For easy plug-and-play in vitro validation and animal experiments, a custom flexible printed circuit board (FPC) was designed for easy connection to test equipment such as electrochemical workstations or amplifiers for electrophysiology recording. The Gr-ITO-TFNMEs and the FPC were connected via gold ball bonding^31,32, replacing the commonly used anisotropic conductive film (ACF) thermal bonding³³. Under ultrasonic pressure, the gold balls electrically connected through circular vias of the square welding pads on the Gr-ITO-TFNMEs to the pads on the FPC, followed by encapsulation and reinforcement with sealing glue, as depicted in Fig. S3.

2.3 Integration of Gr-ITO-TFNMEs with the GRIN lens

The TINIscope system was a miniaturized integrated fluorescence microscope for brain imaging³⁰. The excitation light was focused through a series of optical components and an objective lens onto the neurons. The resulting calcium fluorescence was then collected by the same lens, routed through additional optical components, and finally focused onto an image sensor. The integration of Gr-ITO-TFNMEs with the TINIscope system involved aligning and adhering the end face of objective lens to the FOV of electrodes, ensuring stable connectivity to the electrophysiological recording system without disrupting optical functionality. To minimize tissue damage during lens implantation, a 1 mm diameter GRIN relay lens was affixed to the objective lens using optical adhesive, as shown in Fig. 3(a). The objective lens was bonded to a customized metal baseplate which had two M1 threads for securing the microscope and leveling the lens. The detailed integration of Gr-ITO-TFNMEs with the GRIN lens was conducted with micromanipulator operations, followed five steps, as depicted in Figs. 3(b-f): (1) Fixing and leveling the lens using a lens holder; (2) Applying optical adhesive to the end face of the lens using a fine-gauge needle; (3) Attaching the electrode in reverse to a glass slide, aligning the electrode ‘dial’ with the end face of the lens using a slide holder, and curing optical adhesive; (4) Using a wire tool to attach the electrode ‘strap’ vertically downward from the cut face of the lens; (5) Winding the unattached part of the electrode around the 1.8 mm diameter GRIN lens and securing the electrode and FPC to the baseplate with flexible silicone. The detailed structures of the customized lens holder, slide holder, and wire tool are provided in Fig. S4.

2.4 Electrochemical characterization

The electrochemical characterization of Gr-ITO-TFNMEs was conducted in 0.1 M phosphate-buffered saline (PBS) with the Gr-ITO-TFNMEs serving as the working electrode, with a platinum electrode functioning as both the counter and reference electrode, as shown in Fig. 4(a). Figure 4(b) presents the cyclic voltammetry (CV) characteristics as a representative sample, which display a complete envelope without redox peaks. The result indicates a uniform electrode surface with capacitive characteristics, consistent with the desired interface properties of neural microelectrodes. Figures 4(c-e) show the electrochemical impedance at 1 kHz distribution across 6 recording sites and the averaged electrochemical impedance spectroscopy (EIS) over a frequency range from 100 Hz to 100 kHz. The improved fabrication process results in an acceptable electrochemical impedance, approximately 1.45 MΩ, within the range acceptable for single-unit recording. The impedance is slightly higher than that of same-sized ITO-TFNMEs (as shown in Fig. S5), attributed to the high sheet resistance caused by the two-dimensional atomic structure and grain boundaries of graphene. To evaluate the capability for sustained recording, an accelerated aging test was carried out by soaking the electrodes in PBS at 67°C for one week, comparing three different types of electrodes: Gr-ITO-TFNMEs, Gr-TFNMEs, Au-based flexible microelectrodes (Au-FNMEs, as control). As shown in Fig. 4(f), the Gr-ITO-TFNMEs exhibit the same impedance change trend as the Au-FNMEs, indicating comparable stability. In contrast, the Gr-TFNMEs show a continuous decrease in impedance, suspected to result from delamination between the top insulating layer and the graphene.

2.5 Optical Characterization

2.5.1 Transmittance measurements

The transmittance of Gr-ITO-TFNMEs was measured using a fiber optic spectrometer, as described in our previous work¹⁰.The schematic diagram of the setup is shown in Fig. 5(a). To characterize the impact of different materials on transmittance in the FOV, various combinations was measured: PI-PI, PI-Graphene-PI, PI-Graphene-ITO, PI-Graphene-ITO-PI, and PI-ITO-PI, corresponding to different imaging regions as shown in Fig. 5(b). Additionally, to compare with traditional metal electrodes, the transmittance of PI-Au-PI and PI-Au combinations was also measured. The optical transmittance of different regions of Gr-ITO-TFNMEs and Au-FNMEs is shown in Fig. 5(c). The Au-FNMEs completely obstructed light propagation due to the inherent opacity of metal. There are notable differences in transmittance among recording sites (PI-Graphene-ITO, PI-ITO-PI, PI-Graphene-ITO-PI), interconnects (PI-Graphene-PI), and other non-functional area of FOV (PI-PI). The overall light transmittance at the wavelength of 520 nm is above 80%. These differences in transmittance across different regions can aid in specifically marking electrodes during calcium imaging. The weighted average transmittance across areas of different material combinations for Au-FNMEs and Gr-ITO-TFNMEs is shown in Fig. 5(d), where the overall transparency of Gr-ITO-TFNMEs is significantly improved compared to that of Au-FNMEs. Additionally, an oscillation in transmittance throughout the entire FOV was observed, attributed to thin-film interference, which is consistent with previous work¹⁰.

2.5.2 Imaging of fluorescent polystyrene microbeads

To characterize the impact of Gr-ITO-TFNMEs on calcium imaging, 5 µm diameter fluorescent polystyrene microbeads were utilized to simulate neurons. Au-FNMEs and Gr-ITO-TFNMEs were positioned on polydimethylsiloxane (PDMS, Sylgard 184, Dow Corning) containing a low concentration of fluorescent microbeads to compare their effects on imaging, as shown in Fig. 5(e). The detailed fabrication process for the PDMS containing a low concentration of fluorescent microbeads is provided in Materials and methods below. Figures 5(f-h) shows the imaging of 5 µm diameter fluorescent polystyrene microbeads through Au-FNMEs and Gr-ITO-TFNMEs, including areas within and outside the FOV. The Au-FNMEs fully block the imaging of microbeads at recording sites and interconnects, significantly impairing imaging. In contrast, imaging focused on the surface (Fig. 5(g)) and below the surface (Fig. 5(h)) of Gr-ITO-TFNMEs clearly shows the microbeads, indicating that the electrodes have little effect on fluorescence imaging. Higher magnification imaging is presented in Fig. S6. Fig. S6(a) distinctly shows the fluorescent microbeads under recording sites and interconnects, while Fig. S6(b) shows areas both within and outside the FOV of Gr-ITO-TFNMEs, demonstrating that although the light intensity of the microbeads was still relatively weakened, the fluorescent microbeads can be imaging clearly through recording sites.

2.6 Electrophysiological recording and calcium imaging in vivo

To validate the ability of simultaneous electrophysiology recording and calcium imaging, the GRIN lens integrated with Gr-ITO-TFNMEs was implanted to the surface of the CA1, as shown in, as shown in Figs. 6(a) and (b). Fig. S7 shows the expression of GCaMP7s in the hippocampus. After integrating Gr-ITO-TFNMEs, the TINIscope system maintains high-quality identification of individual neurons (Figs. 6(c) and (d) and Supplementary Video S1), confirming that the introduction of Gr-ITO-TFNMEs has minimal impact on calcium imaging. Importantly, no light-induced artifacts were observed when the excitation light was turned on (Fig. S8). Representative power spectral density of the LFPs for recording site 4 is demonstrated in Fig. 6(e). The average calcium intensity across the FOV changes following the LFPs power spectral density, as shown in Fig. 6(f). The correlation between the calcium intensity and the LFPs power spectral density extracted from the entire recording reaches 0.7398 and is shown in Fig. 6(g). The power spectral densities of the LFPs from the other five recording sites and their correlations with the average calcium intensity are presented in Fig. S9.

Simultaneously, the Gr-ITO-TFNMEs recorded single neuron spikes in the hippocampus. It was the first instance of simultaneous single neuron activity recording and calcium imaging intracortically with TFNMEs, as shown in Figs. 6(h) and (i). The signal-to-noise ratio (SNR) of the spikes reached up to 16, exemplifying the exceptional capability of Gr-ITO-TFNMEs in recording electrophysiological signals. The auto-correlogram of the detected neuron (Fig. S10(a)) displayed clear symmetry and a rapid decrease to baseline at the central peak, and the interspike interval histogram (Fig. S10(b)) showed an exponential decay, indicating a single neuron. Although changes in calcium signal intensity do not directly correspond to changes in action potentials, there is a significant correlation between calcium peaks and the number of spikes within the calcium peak duration⁸. Figure 6(j) shows the representative calcium trace (top) and the spike raster plot (bottom) that exhibit maximum correlation (r = 0.7846), which was obtained by traversing all neurons' calcium peaks and the corresponding number of spikes. It can be observed that fluctuations in calcium trace generally correspond to intense spike activity within brief periods, although not all instances of intense spike activity result in high peaks in the calcium trace. This discrepancy is speculated to be related to the location of calcium proteins, with spikes possibly occurring in regions lacking calcium proteins.

We developed a powerful tool for multimodal neural activity recording and addressed the challenges in the application of TFNMEs from two main perspectives.

On one hand, in electrode fabrication: Extensive research on TFNMEs have demonstrated their capability for multimodal neural activity recording without negative effects^{20–23,26,27,29,34}, as detailed in Table 1. However, due to limitations in fabrication methods and material properties, most electrodes feature relatively large recording sites or are constructed of different insulation materials, resulting in heterogeneous interfaces that restrict the effectiveness and stability of neural electrical activity recording. The combination of ITO and graphene was proposed for TFNMEs to overcome the mechanical deficiencies of ITO-TFNMEs while ensuring overall flexibility and making the electrodes integratable with other implantation tools. Furthermore, this combination harmonizes the material used for both the top and bottom insulation layers, which are compatible with dry etching, thereby eliminating issues such as curling and delamination. A micro-etching technique and pre-treating the graphene with oxygen plasma were proposed to address the challenges in fabricating TFNMEs containing graphene. These advancements have successfully reduced electrode feature-size to the single-cell level (recording sites size of 20 µm and interconnects width of 10 µm), enabling the recording of single neuron activity. The micro-etching technique employs multiple transfers of small-area graphene instead of a single transfer of large-area graphene, significantly reducing wrinkles and cracks during the transfer process and achieving high-quality large-area graphene transfers. Pre-treating the graphene with oxygen plasma specifically targets the process sequence where graphene is transferred before metal deposition. This approach addresses the poor step coverage of graphene and ensures low-impedance contacts between graphene and metals. The impedance of short-circuited metal interconnects to graphene is approximately 1 kΩ when graphene is transferred before metal deposition, compared to 30 kΩ when metal is deposited first. Notably, the electrochemical impedance of Gr-TFNMEs fabricated using both techniques is 1.8 MΩ at 1kHz, which is lower than other Gr-TFNMEs²³, demonstrating the effectiveness of these techniques for fabricating high-quality electrodes containing graphene.

Table 1

Comparison with other transparent flexible neural microelectrodes.
Reference	Function materials	Insulation layer	Electrode size	Electrophysiology	Optophysiology	Animal status
This study	ITO-Graphene	Polyimide/Polyimide	314 µm² (d = 20 µm)	Spikes & LFP (intracortical)	Calcium imaging	Freely moving
Ledochowitsch¹⁴	ITO	Parylene C/Parylene C	7850 µm²(d = 100 µm)	ECoG	Optogenetics	Head-fixed
Cho²¹	PEDOT:PSS-EG	PET-SU8	90000 µm²(w = 300 µm)	LFP & ECoG (cortical)	Optogenetics	Head-fixed
Dijk²²	PEDOT:PSS	Parylene C/Parylene C	7850 µm²(d = 100 µm)	LFP (cortical)	Calcium imaging	Head-fixed
Zhang²³	CNT	PDMS-SU8	10000 µm²(w = 100 µm)	ECoG	Calcium imaging	Head-fixed
Thunemann²⁷	Graphene	PET-SU8	10000 µm²(w = 100 µm)	LFP (cortical)	Calcium imaging & Optogenetics	Head-fixed
Ramezani²⁹	id-DLG with PtNPs	Parylene C/Parylene C	314 µm² (d = 20 µm)	Spikes (cortical)	Calcium imaging	Head-fixed

On the other hand, in the multimodal recording of intracortical neural activity: Previous research^{20–23,26,27,34} predominantly focused on application on the cortical surface, with fewer investigations on intracortical neural activity. Moreover, individual neuronal action potentials (also known as spikes) were rarely captured by TFNMEs due to the size of electrodes. Although a recent study²⁹ utilizing Gr-TFNMEs adjusted the recording sites size to 20 µm through id-DLG and electroplating with PtNPs, successfully recording spikes, it was restricted to recording cortical neural activity in head-fixed mice due to imaging limitations. We integrated TFNMEs with the GRIN lens to successfully record neural activity in freely moving mice and captured intracortical spikes with TFNMEs for the first time. However, the recordings were limited to spontaneous neural activity, demonstrating the potential of Gr-ITO-TFNMEs as a versatile tool for multimodal neuroscience research. Future work should aim to expand electrode density to record a broader range of high spatiotemporal resolution neural activities and incorporate behavioral testing to capture nonlinear neural dynamics.

In conclusion, this study has proposed a new type of transparent flexible neural microelectrodes that can record intracortical multimodal neural activity in freely moving mice. As an effective tool, the electrodes pave the way for further advancements in understanding the complex dynamics of brain with high spatiotemporal resolution for multifunctional neuroscience research.

Fabrication processes of Gr-ITO-TFNMEs

(1) A 3 µm layer of PI was spin-coated at 3000 rpm and cured to form the bottom insulation layer after cleaning the 4-inch silicon wafer. (2) Alignment contours (~ 100 nm) were formed on the bottom PI by oxygen plasma treatment under a positive photoresist (MICROPOST S1805 G2) mask. It should be noted that the areas treated with oxygen plasma must be those intended for subsequent graphene transfer. These treated areas became hydrophilic, facilitating the transfer of graphene to the target areas and decreasing the possibility of formation of excess bubbles that could damage the graphene. (3) Suspended self-help transfer few-layer graphene (XFNANO, 800 ~ 1000 Ω/□) was aligned face-to-face with the alignment contours and sequentially transferred to the target areas after being released onto the surface of deionized water in a beaker. The graphene on PI was then baked at 150 ℃ for 15 minutes to dry residual moisture and partially melt polymethyl methacrylate (PMMA), enhancing contact between the graphene and the bottom PI, reducing wrinkles, and improving adhesion, as demonstrated in Fig. S1. The effect of wrinkles reduction was notably better after baking at 150°C compared to 100°C. After baking, the graphene was soaked in acetone for about 3 hours to remove PMMA. (4) Metal interconnects were patterned through a lift-off process using AR-N4340 negative photoresist and thermally evaporated Cr/Au/Cr with thickness of 10/100/10 nm, respectively. Prior to thermal evaporation, the entire silicon wafer was treated by oxygen plasma to enhance the hydrophilicity of graphene. As shown in Fig. S2, this treatment facilitated the uniform and stable deposition of metals. Additionally, low-impedance contacts were formed between graphene and metals, validated by measuring the impedance of short-circuited metal interconnects to graphene, found to be approximately 1 kΩ. (5) The graphene was patterned by reactive ion etching (RIE) with AZ-6130 photoresist as the etch mask. (6) A 140 nm of ITO was sputtered to form the interface of recording sites with a lift-off process. (7) A 3 µm layer of PI was spin-coated and cured as the top insulation layer. (8) The outline of Gr-ITO-TFNMEs, connection pads and openings of recording sites were defined by RIE with AZ-4620 photoresist as the etch mask. (9) The entire silicon wafer was submerged in deionized water to release the electrodes.

Electrochemical Characterization

The CV and EIS were measured on a two-electrode system (CHI660E, Chenhua Inc., China) in 1x PBS solution (Cytiva HyClone). The experimental devices served as the working electrode, with a platinum electrode functioning as both the counter and reference electrode. CV characteristics were obtained by scanning from an initial voltage of -0.5 V to a final voltage of 0.5 V at a scan rate of 0.05 V/s, and then scanning back to the initial voltage at the same rate. EIS was performed by applying a sinusoidal voltage with an amplitude of 0.05 V over a frequency range from 100 Hz to 100 kHz. The impedance at 1 kHz was used to represent the electrochemical impedance. An accelerated aging test was carried out to evaluate the capability for sustained recording by soaking the electrodes in PBS at 67°C for one week (equivalent to eight weeks at 37°C, with an acceleration factor of 8³⁵). The impedance was measured every 24 hours, comparing three different types of electrodes: Gr-ITO-TFNMEs, Gr-TFNMEs, Au-based flexible microelectrodes (Au-FNMEs, as control).

Fabrication process for the PDMS containing a low concentration of fluorescent microbeads

The microbeads were first thoroughly dispersed in the PDMS cross-linker at a1:500 ratio with ultrasonic treatment for 30 minutes. This microbead-containing crosslinker was then mixed with the PDMS prepolymer at a ratio of 1:10. After stirring and degassing to remove bubbles, the mixture was spin-coated onto pre-cured PDMS blocks at 2000 rpm. The excitation wavelength of the microbeads was 488 nm, and the emission wavelength was 520 nm.

Animals

Male C57BL/6J mice were housed in a controlled environment with a 12-hour light/dark cycle and ad libitum access to food and water. All animal experiments were conducted following protocols approved by the IACUC (Institutional Animal Care and Use Committee) of Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences (SIAT-IACUC-20221124-NS-NXXZX-ZDK-A2028-02).

Surgery

Mice, approximately 8 weeks old, were anesthetized with isoflurane vapor (2% for induction and 1% − 1.5% for maintenance) mixed with O2 and fixed on a stereotaxic apparatus. Body temperature was maintained at 37.5°C using a heating pad during surgery and anesthesia recovery. AAV2/9-hSyn-GCaMP7s virus (Shanghai Taitool Bioscience) was injected into the dorsal hippocampus (dHP) at AP: -2.1 mm, ML: +1.7 mm, DV: -1.4 mm from the surface of the skull with a 9-degree angle. A total of 400 nl of virus was injected at a rate of 15 nl/min. After 3 weeks, mice were remounted on a stereotaxic apparatus for implantation of the imaging lens. A 1.2 mm-diameter hole was drilled in the skull at virus injection site, and the cortex above the CA1 at virus injection site was removed (9-degree angle) using a negative pressure pump. Then, the GRIN lens integrated with Gr-ITO-TFNMEs was implanted to the surface of the CA1 (DV: -1.15 mm below the surface of the skull) and secured to the skull using dental cement. After four weeks recovery and adaptation, electrophysiology recording and calcium imaging were simultaneously conducted in freely moving mice.

Electrophysiology recording and calcium imaging

The signals were sampled at 30 kHz and transmitted to the data acquisition equipment (NeuroStudio System, Jiangsu Brain Medical Technology Co.). The electrical cables were connected to slip rings to avoid entanglement with the TINIscope system. Additionally, a shielding mesh (100 cm ×80 cm ×120 cm) covered the whole system to minimize interference with electrophysiological signals. Calcium imaging was acquired at 12.5 frames per second by the TINIscope system. Electrophysiological signals and calcium imaging were synchronized by sending TTL pulses to Neurostudio and controlling the LED signal lights within the behavioral field of view.

Histology

After recording, mice were sacrificed and transcranial perfused with saline followed by 4% paraformaldehyde (Sigma) in PBS. The brains were kept in 4% paraformaldehyde at 4°C for 24 hours and subsequently immersed in 30% sucrose for 72 hours before being sliced into 40-µm coronal sections and imaged under a fluorescence 247 microscope (Olympus MVX10).

Data Analysis

Calcium imaging was processed with motion correction³⁶ and constrained non-negative matrix factorization³⁷, which facilitate the identification of individual neuron and the acquisition of neuron traces. Electrophysiological signals were analyzed and processed utilizing the Offline Sorter and NeuroExplorer software.

Acknowledgements

The authors gratefully acknowledge the support of the National Key R&D Program of China (2022YFF1202303 and 2023YFF1203702); NSFC-GuangdongJointFund (U20A6005); the National Natural Science Foundation of China (62071447); STI2030-Major Projects (2021ZD0200100), the Science and Technology Innovation Committee of Shenzhen Municipality (JCYJ20220818101611024).

Conflict of interest

The authors declare no competing interests.

Author Contributions

Miao Yuan: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data Curation, Writing - Original draft, Visualization. Fei Li: Conceptualization, Validation, Investigation, Data Curation, Writing - Original draft. Feng Xue: Methodology, Software, Data Curation, Writing - Original draft. Yang Wang: Methodology, Formal analysis, Writing - Original draft. Rongyu Tang: Validation, Resources. Yijun Wang: Resources, Project administration, Funding acquisition. Guoqiang Bi: Conceptualization, Resources, Supervision, Project administration, Funding acquisition. Weihua Pei: Conceptualization, Methodology, Resources, Writing - Original draft, Supervision, Project administration, Funding acquisition,

Cho, Y., Park, S., Lee, J. & Yu, K. J. Emerging Materials and Technologies with Applications in Flexible Neural Implants: A Comprehensive Review of Current Issues with Neural Devices. Advanced Materials 33, 2005786 (2021).
Cho, Y. U., Lim, S. L., Hong, J.-H. & Yu, K. J. Transparent neural implantable devices: a comprehensive review of challenges and progress. npj Flex Electron 6, 53 (2022).
Wellman, S. M. et al. A Materials Roadmap to Functional Neural Interface Design. Adv. Funct. Mater. 28, 1701269 (2018).
Buzsáki, G., Anastassiou, C. A. & Koch, C. The origin of extracellular fields and currents — EEG, ECoG, LFP and spikes. Nat Rev Neurosci 13, 407–420 (2012).
Fattahi, P., Yang, G., Kim, G. & Abidian, M. R. A Review of Organic and Inorganic Biomaterials for Neural Interfaces. Adv. Mater. 26, 1846–1885 (2014).
Huang, S., Liu, Y., Zhao, Y., Ren, Z. & Guo, C. F. Flexible Electronics: Stretchable Electrodes and Their Future. Advanced Functional Materials 29, 1805924 (2019).
Viswam, V., Obien, M. E. J., Franke, F., Frey, U. & Hierlemann, A. Optimal Electrode Size for Multi-Scale Extracellular-Potential Recording From Neuronal Assemblies. Front. Neurosci. 13, (2019).
Chen, T.-W. et al. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 499, 295–300 (2013).
Obaid, S. N. et al. Multifunctional Flexible Biointerfaces for Simultaneous Colocalized Optophysiology and Electrophysiology. Advanced Functional Materials 30, 1910027 (2020).
Wu, X. et al. A Modified Miniscope System for Simultaneous Electrophysiology and Calcium Imaging in vivo. Front. Integr. Neurosci. 15, 682019 (2021).
Koschinski, L. et al. Validation of transparent and flexible neural implants for simultaneous electrophysiology, functional imaging, and optogenetics. J. Mater. Chem. B (2023) doi:10.1039/D3TB01191G.
Chen, Z. et al. Flexible and Transparent Metal Oxide/Metal Grid Hybrid Interfaces for Electrophysiology and Optogenetics. Advanced Materials Technologies 5, 2000322 (2020).
Mikulovic, S. et al. On the photovoltaic effect in local field potential recordings. Neurophotonics 3, 015002 (2016).
Ledochowitsch, P. et al. Strategies for optical control and simultaneous electrical readout of extended cortical circuits. Journal of Neuroscience Methods 256, 220–231 (2015).
Ledochowitsch, P. et al. Strategies for optical control and simultaneous electrical readout of extended cortical circuits. Journal of Neuroscience Methods 256, 220–231 (2015).
Ahmed, N. M. et al. The effect of post annealing temperature on grain size of indium-tin-oxide for optical and electrical properties improvement. Results in Physics 13, 102159 (2019).
Lyau, J.-B., Wu, H.-C. & Chen, H. 21 - Biological studies with tin oxide materials. in Tin Oxide Materials (ed. Orlandi, M. O.) 599–613 (Elsevier, 2020). doi:10.1016/B978-0-12-815924-8.00021-9.
Mittnacht, A., Alt, M. T., Stieglitz, T. & Eickenscheidt, M. Stability of polyimide integrated ITO electrodes. in 2019 9th International IEEE/EMBS Conference on Neural Engineering (NER) 742–745 (2019). doi:10.1109/NER.2019.8717179.
Zátonyi, A. et al. Transparent, low-autofluorescence microECoG device for simultaneous Ca ²⁺ imaging and cortical electrophysiology in vivo. J. Neural Eng. 17, 016062 (2020).
Nishimura, A. et al. Totally transparent hydrogel-based subdural electrode with patterned salt bridge. Biomed Microdevices 22, 57 (2020).
Cho, Y. U. et al. Ultra-Low Cost, Facile Fabrication of Transparent Neural Electrode Array for Electrocorticography with Photoelectric Artifact-Free Optogenetics. Adv. Funct. Mater. 32, 2105568 (2022).
Dijk, G. Fabrication and in vivo 2-photon microscopy validation of transparent PEDOT:PSS microelectrode arrays. Microsystems & Nanoengineering 8 (2022) doi:10.1038/s41378-022-00434-7.
Zhang, J. et al. Stretchable Transparent Electrode Arrays for Simultaneous Electrical and Optical Interrogation of Neural Circuits in Vivo. Nano Lett. 18, 2903–2911 (2018).
Park, D.-W. et al. Fabrication and utility of a transparent graphene neural electrode array for electrophysiology, in vivo imaging, and optogenetics. Nat Protoc 11, 2201–2222 (2016).
Lu, L. Recent Progress on Transparent Microelectrode-Based Soft Bioelectronic Devices for Neuroscience and Cardiac Research. ACS Appl. Bio Mater. 6, 1701–1719 (2023).
Kuzum, D. et al. Transparent and flexible low noise graphene electrodes for simultaneous electrophysiology and neuroimaging. Nat Commun 5, 5259 (2014).
Thunemann, M. et al. Deep 2-photon imaging and artifact-free optogenetics through transparent graphene microelectrode arrays. Nat Commun 9, 2035 (2018).
Jia, S. et al. Graphene oxide/graphene vertical heterostructure electrodes for highly efficient and flexible organic light emitting diodes. Nanoscale 8, 10714–10723 (2016).
Ramezani, M. et al. High-density transparent graphene arrays for predicting cellular calcium activity at depth from surface potential recordings. Nat. Nanotechnol. 1–10 (2024) doi:10.1038/s41565-023-01576-z.
Xue, F. et al. Multi-region calcium imaging in freely behaving mice with ultra-compact head-mounted fluorescence microscopes. National Science Review 11, nwad294 (2024).
Steins, H. et al. A flexible protruding microelectrode array for neural interfacing in bioelectronic medicine. Microsyst Nanoeng 8, 1–15 (2022).
Wei, C. et al. Distributed implantation of a flexible microelectrode array for neural recording. Microsyst Nanoeng 8, 1–8 (2022).
Guan, S. et al. Self-assembled ultraflexible probes for long-term neural recordings and neuromodulation. Nat Protoc 18, 1712–1744 (2023).
Liang, Q. et al. Electron Conductive and Transparent Hydrogels for Recording Brain Neural Signals and Neuromodulation. Advanced Materials 35, 2211159 (2023).
Hukins, D. W. L., Mahomed, A. & Kukureka, S. N. Accelerated aging for testing polymeric biomaterials and medical devices. Medical Engineering & Physics 30, 1270–1274 (2008).
Pnevmatikakis E. A. & Giovannucci A. NoRMCorre: An online algorithm for piecewise rigid motion correction of calcium imaging data. J. Neurosci. Methods 291, 83–94 (2017).
Zhou, P. et al. Efficient and accurate extraction of in vivo calcium signals from microendoscopic video data. eLife 7, e28728 (2018).

(Not answered)

Supplementaryinformation.docx
Supplementaryvideo.mp4
Supplementary Video S1

Download PDF

Reviewer #2 agreed at journal
05 Nov, 2024
Review #1 received at journal
25 Sep, 2024
Reviewer #1 agreed at journal
11 Sep, 2024
Reviewers invited by journal
26 Aug, 2024
Submission checks completed at journal
21 Aug, 2024
Editor assigned by journal
20 Aug, 2024
First submitted to journal
20 Aug, 2024

You are reading this latest preprint version

Transparent, flexible graphene-ITO-based neural microelectrodes for simultaneous electrophysiology recording and calcium imaging of intracortical neural activity in freely moving mice

Status:

Version 1

Abstract

Figures

Introduction

Results

2.1 Design of the Gr-ITO-TFNMEs

2.2 Fabrication of the Gr-ITO-TFNMEs

2.3 Integration of Gr-ITO-TFNMEs with the GRIN lens

2.4 Electrochemical characterization

2.5 Optical Characterization

2.5.1 Transmittance measurements

2.5.2 Imaging of fluorescent polystyrene microbeads

2.6 Electrophysiological recording and calcium imaging in vivo

Discussion

Materials and methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1