Sleep is an important physiological process that is indispensable to life activities and plays a core role in energy recovery, mood regulation, neuroplasticity, memory consolidation, immune regulation and growth and development1,2. However, sleep-wakefulness involves complex dynamic regulation at multiple levels, from molecules and cells to neural networks. Current research suggests that the regulation of the sleep-wake system primarily involves interactions between multiple brain regions and nuclei, including the midbrain, brainstem, thalamus, hypothalamus, basal forebrain3,4, and cortex5. However, its underlying neural mechanisms remain largely unexplored.

Among the many regulatory centers, the ventral tegmental area (VTA) has long been the focus of behavioral research on reward and motivation6,7, but increasing evidence shows that its core role in sleep-wake transitions cannot be ignored8,9. However, due to the limitations of deep brain recording methods, there is currently a lack of high temporal and spatial resolution research tools to study the refined coding mechanisms of VTA neurons during the sleep-wake cycle.

As a tool that can realize real-time monitoring of electrical activity at the cellular level10, the development of implantable neural micro-nanoelectrodes has brought revolutionary changes to the research in the field of neuroscience11,12. To enhance electrode performance and durability of electrodes, conductive polymers have been extensively used as surface functionalization coatings for neural electrodes13,14. Among them, the most representative poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) (PEDOT:PSS) is commonly utilized in bioelectronics owing to its superior conductivity, film-forming features, low cost and easy preparation15. In recent years, PEDOT:PSS-based materials have been further optimized for application in microelectrode array platforms, targeting areas such as neural signal recording16 and organ-on-a-chip systems17. However, conductive polymers such as PEDOT:PSS are prone to delamination and shedding during long-term use in vivo18, leading to signal attenuation and even device failure. This problem mainly stems from its weak adhesion to the substrate material and the lack of a strong interfacial bond19,20, which limits its practical application in neural interfaces with high requirements for long-term stability21. To this end, Various studies have focused on increasing the adhesion of conductive polymer coatings through methods such as chemical grafting22,23, self-assembly24, or electrode surface roughening25,26. Although these methods have achieved positive results, they generally suffer from complex processes and poor material adaptability. Therefore, an immediate demand exists for designing a more direct, efficient and material-adaptable interface modification strategy to meet the high stability requirements of the neural electrode interface under long-term in vivo implantation conditions.

Notably, the biomimetic material polydopamine (PDA) has recently attracted attention as a potential option for enhancing the stability of neural interfaces owing to its excellent wet adhesion and biocompatibility27,28,29,30. Lee et al. first reported the use of PDA films as coating materials, prepared by dip-coating an alkaline dopamine solution31. Compared to traditional dip-coating methods, electrochemical deposition allows for faster and more uniform PDA film formation32,33. Furthermore, it allows for precise control of film thickness34.

Unlike the conventional multi-step modification strategies that sequentially deposit a PDA layer followed by a conductive layer35,36, this study innovatively employed a mixed electrochemical co-deposition strategy of EDOT:PSS and dopamine (DA) solution, achieving synergistic interface modification in a single-step deposition process. This method retains the superior electrical performance of PEDOT:PSS while considerably improving the coating’s mechanical stability and biocompatibility, potentially becoming a universal solution for constructing high-performance neural electrode interfaces.

In this study, we fabricated a multi-channel microelectrode array (MEA) and significantly improved the electrode interface performance by using a PEDOT:PSS/PDA hybrid coating. This electrode was implanted into the VTA of mice, enabling simultaneous monitoring of local field potentials (LFPs) and single-neuron spikes. In parallel, electroencephalogram (EEG) and electromyogram (EMG) signals were collected, together providing an integrated dataset across the sleep–wake cycle. In summary, this study not only provides a theoretical basis and technical path for the development of long-lasting, stable, and efficient neural electrode interface materials, but also provides new neural circuit evidence for the involvement of the VTA in sleep-wake regulation, potentially promoting further development of neuroscience research and clinical neuromodulation technologies for sleep disorders.

Materials and methods

Design of neural MEA

In this study, a MEA with a “sandwich” structure was fabricated on a silicon-on-insulator wafer37. The three-layer architecture of the MEA consists, from top to bottom, of an electrically insulating layer composed of SiO₂ and Si₃N₄, a conductive layer (Ti/Pt), and a substrate layer comprising the Si substrate and a thermally grown SiO₂ layer (Fig. 1a).

Fig. 1: Fabrication of MEA and directional modification of PEDOT:PSS/PDA.
Fig. 1: Fabrication of MEA and directional modification of PEDOT:PSS/PDA.
Full size image

a Schematic of the “sandwich” structure of the MEA. b Enlarged optical microscopy image of the MEA tip. c Photograph of the fully assembled device used for in vivo recording of sleep–wake neural signal. d MEMS fabrication process flow of the MEA: (1) Thoroughly clean the SOI wafer; (2) Thermally grow a SiO₂ insulating layer on the substrate; (3) Define the Ti/Pt conductive layer pattern on the wafer by photolithography and sputtering; (4) Remove the unwanted metal using a standard lift-off process; (5) Deposit SiO₂/Si₃N₄ bilayer insulation films by plasma-enhanced chemical vapor deposition; (6) Form electrode sites and contact pads by photolithography and plasma etching; (7) Pattern the silicon shank structure by photolithography and deep reactive ion etching; (8) Protect the electrode areas with photoresist and sealing wax prior to release; (9) Release the MEA by wet etching of the silicon handle layer in KOH solution. e Directed modification strategy of PEDOT:PSS/PDA using a three-electrode system to co-deposit PDA complexes and PEDOT:PSS chains onto the electrode surface, forming an interpenetrating network structure

To enable precise recording of neural activity in deep brain regions during sleep–wake states, we custom-developed 16-channel MEA, each with a 20 µm-sized recording site, suitable for the dimensions of single neurons. In addition, a large rectangular grounding site (250 µm × 25 µm) was designed to provide a stable potential reference for accurate neural signal acquisition (Fig. 1b). The geometry and dimensions of the shank were tailored according to the size and stereotaxic coordinates of the VTA in the mouse coronal brain atlas: the total shank length was 6 mm with a width of only 380 µm (Fig. S1), matching the anatomical features of the VTA. Furthermore, the MEA thickness was limited to 25 µm, structurally minimizing tissue damage and inflammatory responses associated with in vivo implantation. Finally, the MEA, EEG leads, EMG leads, and custom Printed Circuit Board (PCB) were integrated and encapsulated into a unified device for performance testing and in vivo implantation, aimed at accurately monitoring neural activity during sleep and wakefulness (Fig. 1c). The overall process flow for preparing MEA using microelectromechanical systems (MEMS) technology is shown in Fig. 1d. A detailed description of each fabrication step is provided in the Supplementary Methods section entitled “Fabrication Procedure of Neural MEAs”, where sufficient technical details are included.

Targeted modification of PEDOT:PSS/PDA

20 mM EDOT, 0.0142 g PSS, and 1 mg/mL DA were mixed in 5 mM Tris-HCl buffer (pH 7.4, sterile) to prepare a 10 mL mixed plating solution. The solution was then sonicated for 30 min to ensure thorough mixing (Fig. 1e). 3,4-Ethoxylene dioxythiophene (EDOT, CAS 126213-50-1, 97%), Polystyrene sulfonic acid (PSS, CAS 25704-18-1, average Mw ≈ 70,000), Dopamine hydrochloride (DA, CAS 62-31-7, 98%), and Tris-hydrochloride buffer (CAS 77-86-1) were used as received.

Prior to formal modification, the electroplating solution was deoxygenated by bubbling N₂ for 5 minutes to prevent possible oxidative self-polymerization of dopamine. Simultaneously, the microelectrode surface was cleaned using oxygen plasma (vacuum ≤ 2 Pa; power: 300 W; oxygen flow rate: 30 L/min; time: 3 min). In this study, two different substrates were used for electrodeposition experiments: MEA sites and indium tin oxide (ITO) conductive glass. The ITO conductive glass was specifically used for hydrophilicity testing and in vitro cytotoxicity assessment, while the MEA sites served as the primary functional interface for the remaining performance characterizations. PEDOT:PSS/PDA was deposited electrochemically on the different substrates. PEDOT:PSS was also electroplated on the substrate using the same method for subsequent performance comparison. The electroplating process was carried out on a VersaSTAT 4 electrochemical workstation produced by AMETEK Princeton Applied Research, USA, using a three-electrode system, in which a microelectrode array or ITO conductive glass served as the working electrode, with a Pt electrode as the counter and an Ag/AgCl electrode as the reference.

The electrochemical deposition parameters and cyclic voltammetry (CV) curves for PEDOT:PSS and PEDOT:PSS/PDA coatings on different substrates were optimized through extensive exploratory experiments, as detailed in Fig. S2.

Electrical performance and mechanical stability testing

Electrochemical impedance spectroscopy (EIS) was conducted to evaluate the impedance and phase characteristics of the microelectrode site. The charge storage capacity (CSC) was quantitatively analyzed by CV at a scan rate of 100 mV/s. The potential window was within the electrochemical stability window of aqueous solution to avoid interference from the water decomposition reaction. The CSC value was determined by integrating the response current over the scanned voltage range38, with the detailed calculation method and equation provided in the methods section of supplementary materials.

To simulate mechanical stresses such as micromotion, shear forces, and fluid disturbances that electrodes may encounter during actual implantation and long-term in vivo embedding, and to evaluate their mechanical stability under dynamic physiological conditions, ultrasound-based mechanical stability tests were conducted in this study. The electrodes were fixed inside cylindrical containers prefilled with phosphate-buffered saline, ensuring that the electrode tips were fully immersed in the solution. The entire setup was then placed in a water bath ultrasonic cleaner (ultrasound power: 180 W; frequency: 40 kHz) for ultrasonic treatment durations of 30 minutes and 1 h respectively (Fig. S3a). Electrochemical performance evaluations, including EIS and CV measurements, were immediately performed after each ultrasonic treatment stage. Changes in the electrochemical properties before and after ultrasonic treatment were compared to comprehensively assess the adhesion stability of the electrode coatings and the interface reliability under mechanical perturbations.

Hydrophilicity and biological toxicity testing

In the hydrophilicity test, a deionized water droplet was pipetted onto the modified ITO surface. After the droplet stabilizes, the liquid-solid interface between the droplet and the sample surface was captured. The droplet profile is then fitted and the contact angle is calculated to characterize the hydrophilicity of the material surface.

In biocompatibility testing, live/dead staining and CCK8 cytotoxicity assays were performed using in vitro cultured mouse neural stem cells (NSCs, C17.2 cell line, iCell Bioscience Inc., Shanghai). Unmodified ITO glass served as a control. The specific operation process and relative cell viability calculation method are detailed in the methods section of supplementary materials.

MEA implantation and histological verification

A total of six adult C57BL/6 N mice, weighing 30 ± 5 g and aged 9 weeks, were used in this study. Mice were allowed to acclimate for one week in a controlled environment (24 °C, 40% humidity, 12 h light and dark cycle) prior to the surgical procedure. Anesthesia was induced and maintained with isoflurane inhalation. All surgical instruments were sterilized with alcohol. Mice were positioned in a stereotaxic frame, with bilateral ear holders adjusted to keep the head stable. The scalp was shaved, and a central cut was created to reveal important cranial reference points. The bregma was designated as the stereotaxic zero point, and the head was leveled using the stereotaxic probe, with an error controlled within ±0.03 mm. A craniotomy was carried out at the VTA target site (AP: –2.91 mm; ML: 0.75 mm; DV: 4.75 mm) to enable MEA implantation. (Fig. S4a). Four additional small holes were drilled to accommodate two EEG cranial screws and two ground screws; the two EEG screws were positioned over the prefrontal cortex region of the mouse (Fig. S4b). After implanting the MEA, EEG and ground leads were connected to the respective cranial screws, and the EMG electrode was implanted into the neck muscles (Fig. S4c). The entire assembly was secured with dental cement (Fig. S4d). A postoperative recovery period of at least one week was provided (Fig. S4e) to ensure stable health conditions and improve the reliability of recorded data.

After sleep-wake neural information monitoring, we performed histological sectioning on the animals to verify that the MEA were accurately implanted in the target brain region. Detailed procedures are described in the methods section of supplementary materials.

Sleep-wake neural information processing and analysis

After the mice recovered from surgery, neural signals were continuously recorded for 12 h in an open area using a proprietary multi-channel neural signal acquisition system39 to explore the characteristics of neural activity during different brain functional states of sleep and wakefulness(Fig. S4f). During this period, the mice were free to move and had ample food and water.

Subsequent signal processing and statistical analysis were performed using NeuroExplorer 4 (Nex Technologies, Colorado Springs, USA), Offline Sorter v3 (Plexon Inc., Dallas, Texas, USA), MATLAB R2018b (The MathWorks, Inc., Natick, Massachusetts, USA), and Origin 2021 (OriginLab Corporation, Northampton, Massachusetts, USA). The Lunion Stage Sleep Analysis System (LunionDate, Shanghai, China) was used to categorize the sleep-wake phase of mice into wakefulness, rapid eye movement (REM) and non-rapid eye movement (NREM) sleep according to EEG and EMG signals. The detailed description of the neuronal classification procedure is provided in the Supplementary Methods. In addition, electrophysiological signals were further separated by oscillation frequency into δ (0.5–4 Hz), θ (4–8 Hz), and other bands40.

The performance of the electrodes in acquiring in vivo electrophysiological signals was quantified by the signal-to-noise ratio (SNR)41, calculated as detailed in the methods section of the supplementary materials.

Results and discussion

Modification and surface morphology of PEDOT:PSS/PDA coating

During the initial or early cycles of CV scanning, DA molecules are first oxidized under the applied anodic potential (approximately 0.2–0.7 V vs. Ag/AgCl, as shown in the CV curves in Fig. S2b) to dopaminequinone. This is followed by intramolecular cyclization to form leucodopaminochrome, which is further oxidized to dopaminochrome. After intramolecular rearrangement, 5,6-dihydroxyindole (DHI) is generated35. DHI serves as a key monomer in the polymerization of PDA. DHI molecules themselves can also be oxidized to DHI quinone.

During the same anodic potential scan (approximately >0.7 V vs. Ag/AgCl), EDOT monomers also begin to oxidize to form cationic radicals, which serve as active centers for EDOT polymerization. These EDOT cationic radicals continuously combine with each other to form positively charged PEDOT chains. PSS, as a polyanion, acts as a charge-balancing dopant during polymerization, tightly binding to the PEDOT chains through electrostatic interactions to form a PEDOT:PSS complex19,42.

In subsequent CV cycles, oxidized DHI species (DHI quinone or radicals) react with neighboring reduced DHI or DA molecules, leading to the formation of oligomers that eventually grow into a PDA polymer network rich in catechol/quinone, amino, and imino functional groups36. The reaction mechanism of PDA electropolymerization is illustrated in Fig. S5. Near the electrode surface, functional groups of PDA, such as phenolic hydroxyls and amines, may interact with sulfonate groups of PSS and cationic sites on PEDOT via weak interactions including hydrogen bonding, electrostatic attraction, and π–π interactions. Driven by applied potential and ongoing polymerization, PDA complexes and PEDOT:PSS chains gradually lose solubility and co-deposit onto the electrode surface.

The entire deposition process is not a simple physical mixing and stacking, but is accompanied by the cooperative assembly of molecules. PDA has extremely strong adhesion, and its early deposition layer or oligomer can serve as a “binder” and nucleation site for the subsequent PEDOT:PSS deposition, promoting the firm adhesion of PEDOT:PSS to the electrode surface. Conversely, the first deposited PEDOT:PSS conductive network may also provide an electron transfer pathway for the oxidative polymerization of DA/DHI. Ultimately, the polymer network of PDA and the conductive network of PEDOT:PSS intertwine and penetrate each other during the deposition process to form an interconnected interpenetrating network structure (Fig. 1e).

The surface structure of neural electrodes directly influences their interfacial properties with tissue and the efficiency of electrical signal transmission. Especially at the microscale of microelectrodes, coating uniformity, roughness, and adhesion state can significantly affect their subsequent electrical performance and in vivo stability.

Optical microscopy observations (Fig. 2a) revealed that the bare electrodes exhibited a metallic luster with a smooth and flat surface. PEDOT:PSS-modified electrodes showed a uniform, flat, light-blue film, indicating homogeneous deposition of the conductive polymer on the electrode surface. In contrast, electrodes coated with the PEDOT:PSS/PDA composite displayed a uniformly dark black appearance, likely due to the introduction of PDA altering the polymer film morphology during co-deposition.

Fig. 2: Surface morphology of different coating.
Fig. 2: Surface morphology of different coating.
Full size image

a Comparison of optical microscopic images of different coatings. Both PEDOT:PSS and PEDOT:PSS/PDA coatings have obvious color characteristics. b, c Comparison of SEM images of different coatings. The PEDOT:PSS/PDA coating exhibits an obvious rough structure

Scanning electron microscopy (SEM) images further elucidated changes in surface microstructure (Fig. 2b, c). The bare electrode surface was smooth, while PEDOT:PSS modification resulted in a dense and uniform film with a very flat morphology. The PEDOT:PSS/PDA co-deposited surface, however, exhibited a notably roughened morphology, with localized particle aggregation and irregular deposition structures. This increase in surface roughness may arise from PDA disrupting the ordered stacking of PEDOT:PSS during deposition. On one hand, PDA’s hydrogen bonding network and non-covalent interactions promote aggregation and crosslinking of PEDOT chains; on the other hand, PDA itself forms particulate structures during electropolymerization. The synergistic effect of these factors leads to a more hierarchically structured surface morphology.

To characterize the elemental composition of the different modification layers and further verify their effective deposition on the electrode surface, energy-dispersive spectroscopy (EDS) analysis was performed. As shown in Fig. S6, PEDOT:PSS modification on bare electrodes resulted in significant increases in the atomic percentages (At%) of carbon and sulfur, reflecting the polymer backbone and dopant ions. Following PEDOT:PSS/PDA co-deposition, nitrogen—a characteristic element of PDA—and an increased proportion of oxygen were additionally detected, confirming the successful incorporation of amino and phenolic hydroxyl functional groups.

Electrical features and mechanical stability of PEDOT:PSS/PDA coating

Electrical properties such as impedance and phase delay of microelectrode arrays are crucial for efficient electrophysiological signal recording43. Lower impedance and phase delay substantially reduce signal attenuation and distortion, thereby providing high-quality neural electrophysiological data. Since the primary frequency band of neural electrophysiological signals is concentrated around 1 kHz, we focused on the electrical characteristics of the microelectrode arrays at this frequency.

As shown in Fig. 3a, compared with the bare electrodes, the impedance of electrodes modified with PEDOT:PSS and PEDOT:PSS/PDA exhibits a significant decrease. The specific values at 1 kHz are presented in Fig. S7a, where the impedance of the bare electrode sites was 2085.66 ± 248.87 kΩ. After PEDOT:PSS modification, the impedance significantly decreased to 40.15 ± 3.33 kΩ, and further reduced to 28.03 ± 3.25 kΩ following PEDOT:PSS/PDA coating. This improvement is primarily attributed to the rough microstructure formed on the co-deposited coating surface, which effectively increases the electrode’s specific surface area, enhances interfacial electron/ion transport, and thus improves signal conduction efficiency. Additionally, the amino and phenolic hydroxyl groups in PDA provide extra electron transfer pathways on top of the inherently good conductivity of PEDOT:PSS, further enhancing the overall interfacial conductivity and synergistically reducing electrode impedance.

Fig. 3: Electrical properties and mechanical stability of different coatings.
Fig. 3: Electrical properties and mechanical stability of different coatings.
Full size image

a Impedance spectra of bare Pt electrodes, PEDOT:PSS-modified electrodes, and PEDOT:PSS/PDA-modified electrodes across different frequencies. Error bars represent standard deviation (SD) (n = 5). b Phase spectra of different electrodes across frequencies. Error bars represent SD (n = 5). c CV curves of different electrodes. d Changes in surface morphology of two coating materials before and after ultrasound. e Changes in impedance at 1 kHz before and after ultrasonication for different coating materials. Error bars represent SD (n = 4). *p < 0.05, **p < 0.01, ***p < 0.001

As illustrated in Fig. 3b, in comparison with the bare electrodes, the phase delay of the electrodes modified with different coatings is also significantly improved. Regarding phase delay at 1 kHz (Fig. S7b), the bare electrode exhibited a high phase delay of 83.96 ± 3.28°. After PEDOT:PSS modification, this was reduced to 61.41 ± 3.19°, while the PEDOT:PSS/PDA-coated electrodes showed a significant decrease to 31.45 ± 4.1°. This reduction may be related to the homogenization of capacitance distribution caused by the rough coating interface, which effectively suppresses polarization effects at the electrode interface. Consequently, the current response at 1 kHz becomes more synchronous with the excitation signal, exhibiting lower phase lag characteristics.

In addition to improvements in impedance and phase response, the CSC is another important parameter for evaluating neural electrode performance. A higher CSC not only facilitates more efficient charge exchange between the electrode and neural tissue but also enhances interfacial electrochemical stability, thereby supporting reliable long-term neural signal recording. CSC was characterized using CV scans (Fig. 3c), where it is qualitatively evident that, within the same potential range, electrodes coated with PEDOT:PSS/PDA exhibit higher current densities at the electrode–electrolyte interface. Quantitative CSC values were then calculated based on the integrated area of the CV curves (Fig. S7c), and the results showed that electrochemical deposition markedly improved CSC, increasing from 2.70 ± 0.33 mC/cm² for bare electrodes to 31.59 ± 0.28 mC/cm² for PEDOT:PSS-modified electrodes, and further to 85.16 ± 1.19 mC/cm² for PEDOT:PSS/PDA co-deposited electrodes. Compared with bare electrodes, the CSC of MEA sites modified with PEDOT:PSS/PDA increased by nearly 30-fold, demonstrating outstanding charge storage capability and interfacial modulation potential. This enhancement is closely related to the enlarged specific surface area provided by the co-deposited coating, while the abundant functional groups in PDA may also participate in charge capture and release during electrochemical processes, further contributing to the CSC improvement.

While ensuring that the electrode has excellent electrical properties, its structural stability for long-term stable operation in the body fluid environment is also crucial.As shown in Fig. 3d, the PEDOT:PSS coating showed signs of delamination and detachment after 30 minutes of ultrasonic agitation. After 60 minutes, the coating had clearly detached, exposing the bare electrode morphology. In contrast, the PEDOT:PSS/PDA coating maintained a well-developed morphology under the same conditions, with little noticeable damage. EIS results further confirmed this observation: the impedance of the PEDOT:PSS-modified electrode increased dramatically from 41.77 ± 2.02 kΩ before ultrasonic agitation to 940.06 ± 66.13 kΩ at 1 kHz, while the impedance of the PEDOT:PSS/PDA-modified electrode remained unchanged before and after ultrasonic treatment (Fig. 3e). Correspondingly, CV curves revealed a significant deterioration in the charge storage capacity of the PEDOT:PSS coating after ultrasonic agitation (Figure S3b), while that of the PEDOT:PSS/PDA coating remained nearly constant after prolonged ultrasonic treatment(Fig. S3c). These results clearly demonstrate that the PEDOT:PSS/PDA coating possesses superior mechanical stability. This improvement can be attributed to the strong interfacial bonding effect of PDA and the formation of a dense interpenetrating network structure during the co-deposition process. The abundant catechol and amine functional groups in PDA establish multiple covalent and noncovalent interactions, including electrostatic interaction, hydrogen bonding and π–π stacking, with both the electrode surface and the PEDOT:PSS matrix35,36, thereby significantly enhancing interfacial adhesion and mechanical robustness.

Importantly, since the metal layer structure used in this study is highly similar to that of commonly employed neural electrodes16,44, this coating scheme holds strong potential as a universal interface strategy for neural electrodes and clinical applications.

Biocompatibility characterization of PEDOT:PSS/PDA coating

Biocompatibility is a key factor in whether neural electrode materials can serve stably and long-term in the body. Good surface hydrophilicity and low cytotoxicity help promote the adhesion, growth and functional maintenance of nerve cells45. As shown in Fig. 4a, the hydrophilicity of the coating surfaces was first evaluated via water contact angle (WCA) analysis. The results showed that the WCA of the ITO (Bare) surface and the PEDOT:PSS coating were 73.1 ± 1.8° and 70.5 ± 0.7°, respectively, indicating that the surfaces were slightly hydrophilic. Interestingly, the WCA of the surface after PEDOT:PSS/PDA co-deposition decreased significantly to 32.0 ± 1.1° (Fig. 4c), indicating that with the incorporation of PDA, the hydrophilicity of the coating interface was greatly enhanced.

Fig. 4: Biocompatibility characterization of different coating.
Fig. 4: Biocompatibility characterization of different coating.
Full size image

a Water contact angle images of interfaces with different coating materials, including bare ITO conductive glass. b Fluorescence staining images of mouse NSCs cultured on different coating surfaces for 1, 3, and 5 days. c Statistical results of water contact angles on different coating surfaces. Error bars represent SD (n = 5). d Relative cell ability on different coating surfaces over various culture durations. Error bars represent SD (n = 3). e Average cell ability over the entire culture period on different coating surfaces. Error bars represent SD (n = 9). *p < 0.05, **p < 0.01, ***p < 0.001

Subsequently, we evaluated the viability of mouse NSCs cultured on different coating materials using the LIVE/DEAD cell staining assay, with unmodified ITO (Bare) serving as the control. Cells were cultured for 1, 3, and 5 days on each surface, and observed under a fluorescence microscope (Fig. 4b). Across all groups, no substantial cell death was observed; however, compared with PEDOT:PSS, the PEDOT:PSS/PDA coating consistently supported a larger number of adherent and well-spread cells at all time points. Quantitative analysis of relative cell viability (Fig. 4d) showed that although cell activity on all surfaces increased steadily over time, the PEDOT:PSS coating exhibited significantly lower viability on day 1 (37.58 ± 0.54%) and day 3 (69.91 ± 0.52%) than the other groups, indicating that early cell growth and proliferation were suppressed. In contrast, the PEDOT:PSS/PDA coating already demonstrated high viability on day 1 (78.94 ± 1.04%) and further increased to 87.86 ± 1.84% by day 3, reflecting excellent biocompatibility. The average cell viability throughout the entire culture period (Fig. 4e) was consistent with these trends. This improvement may be attributed to the abundant phenolic functional groups in PDA, which enhance the hydrophilicity of the coating and promote cell adhesion and proliferation46, while its biomimetic characteristics may mitigate the potential cytotoxicity of the conductive polymer47. In summary, the PEDOT:PSS/PDA co-deposition coating greatly improved the interfacial environment and provided a good biological foundation for the prolonged in vivo application of neural electrodes.

Electrophysiological activity of the VTA during sleep and wakefulness

After examining the excellent electrical performance, mechanical stability, and biocompatibility of MEA modified with the PEDPT:PSS/PDA coating, they were further applied to in vivo monitoring of neural information within the brain. Specifically, the MEA was precisely implanted into the VTA of mice to record LFPs and spike signals, while EEG and EMG signals were simultaneously acquired in parallel. Together, these recordings enabled the integrated monitoring of neural activity across multiple levels during natural sleep and wakefulness. Comparison of brain tissue sections with a mouse brain atlas (Fig. S8) confirmed that the MEA were successfully implanted into the target brain region.

To assess the influence of different coating materials on the signal recording performance of neural probes, we conducted a 21-day in vivo electrophysiological monitoring. As shown in the representative raw neural waveforms in Fig. 5a, electrodes modified with the PEDOT:PSS/PDA composite coating exhibited action potentials with higher amplitudes—and consequently higher signal-to-noise ratios (SNRs)—compared with PEDOT:PSS-coated electrodes. The temporal evolution of SNR for each coating is presented in Fig. 5b. From the first to the second week post-implantation, both groups showed an increase in SNR, which may be associated with the gradual attenuation of postoperative inflammatory responses. By the third week, a decline in SNR was observed for the PEDOT:PSS electrodes, whereas the PEDOT:PSS/PDA-coated electrodes maintained a significantly higher SNR. This sustained performance is likely due to the enhanced mechanical robustness of the PEDOT:PSS/PDA coating, which helps mitigate delamination and degradation in the complex in vivo environment. Furthermore, Fig. 5c demonstrates that electrodes with PEDOT:PSS/PDA coatings were able to stably record spike waveforms with consistent features over extended implantation periods, providing further evidence of their excellent in vivo stability and signal fidelity in practical brain tissue environments.

Fig. 5: In vivo recording performance of electrodes modified with different coatings.
Fig. 5: In vivo recording performance of electrodes modified with different coatings.
Full size image

a Representative raw neural signal waveforms with the signal-to-noise ratio during recording for each coating modified electrode. b Changes in signal-to-noise ratio of different coating interfaces at different time points after implantation. c Tracking of representative channel spike waveforms of electrodes modified with PEDOT:PSS/PDA composite coatings after multiple days of implantation

As shown in Fig. 6a, behavioral states in mice were classified into wakefulness, NREM/REM sleep based on EEG and EMG signals. Neural activity was simultaneously recorded across these states. Representative spike timestamps and corresponding LFP waveforms from selected channels are displayed in Fig. 6b. Notably, certain channels exhibited distinct state-dependent firing patterns, suggesting that these neurons may be regulated by, or contribute to, transitions between sleep–wake states. For example, spike timestamps from channel 7 revealed a markedly higher firing rate during NREM and REM sleep compared with wakefulness, preliminarily indicating their possible involvement in the maintenance or transition of sleep-related states. The firing characteristics of different neuron types across behavioral states will be analyzed in detail in later sections.

Fig. 6: Neural activity recordings in the VTA across sleep–wake states.
Fig. 6: Neural activity recordings in the VTA across sleep–wake states.
Full size image

a Representative EEG spectrograms and EMG traces in different states. b Representative multichannel neurophysiological recordings in different states, showing spike raster plots and corresponding LFP waveforms

In addition, the LFPs recorded from the VTA displayed distinct patterns across states: during wakefulness, LFPs were characterized by low-amplitude, high-frequency activity; during NREM sleep, they shifted to high-amplitude, low-frequency slow-wave oscillations; and during REM sleep, they reverted to low-amplitude, fast-wave rhythms resembling those in wakefulness. These observations are consistent with previous reports describing state-dependent LFP patterns in the VTA.

State-specific neuronal activity and LFP characteristics across sleep and wakefulness

Investigating the modulation of neuronal firing rates across sleep–wake states is of considerable neurophysiological significance for elucidating the mechanisms underlying the initiation, transition, and maintenance of distinct brain functional states. We first applied the K-means clustering algorithm to spike waveforms recorded from three mice with high signal-to-noise ratios to achieve automatic spike sorting and preliminary neuronal classification48. Subsequently, neurons were further categorized based on their firing rate characteristics across different brain functional states. Across these datasets, a total of 87 single units were identified in the VTA, comprising three distinct neuronal populations with characteristic firing patterns (Fig. 7a–f): wake-responsive neurons (WRNs), sleep-responsive neurons (SRNs), and state-independent neurons (SINs). The proportional distribution of these neuronal populations is illustrated in Figure S9. WRNs exhibited the highest firing rates during wakefulness (Fig. 7g), suggesting a potential role in promoting wake maintenance or facilitating transitions from sleep to wakefulness. In contrast, SRNs exhibited markedly elevated firing rates across NREM and REM sleep relative to wakefulness (Fig. 7h), showing a typical sleep-specific firing pattern. SINs, which were widely distributed in the VTA, exhibited no significant differences in firing rate across brain functional states (Fig. 7i), indicating that their activity remains largely unaffected by sleep–wake transitions.

Fig. 7: Analysis of state-specific neuronal firing activity in the VTA during sleep–wake cycles.
Fig. 7: Analysis of state-specific neuronal firing activity in the VTA during sleep–wake cycles.
Full size image

a–c Representative waveforms of three neuronal populations with distinct firing patterns: wake-responsive neurons (WRNs), sleep-responsive neurons (SRNs), and state-independent neurons (SINs). Error bars represent SD (n = 30). d–f Representative firing autocorrelograms of each neuron type. g–i Average firing rates of each neuron type across different states. Error bars represent SD (n = 5). *p < 0.05, **p < 0.01, ***p < 0.001

Previous studies have shown that VTA glutamatergic neurons can sustain wakefulness, whereas GABAergic neurons exhibit diametrically opposite properties8. Other work has demonstrated that activation of VTA dopaminergic neurons promotes wakefulness and motivation-related behaviors9. The coexistence of multiple state-dependent neuronal populations identified in the present study further underscores the role of the VTA as a critical hub in regulating the dynamics of sleep–wake transitions, reflecting the complexity and diversity of its intrinsic neural network functions. Nevertheless, whether the recorded neuronal populations correspond to specific neurotransmitter-defined cell types remains to be determined, which would provide deeper insights into the fine-tuned regulatory mechanisms of VTA neural circuits in sleep–wake dynamics.

In addition to single-unit spike activity, the low-frequency continuous LFP signals provide valuable clues for deciphering local neural dynamics during the sleep–wake cycle. We compared LFP power in the VTA across different behavioral states. As shown in Fig. S10a, the mean LFP power during wakefulness (3915.05 ± 185.26 µV²) was significantly higher than during NREM sleep (2413.47 ± 82.04 µV²) and REM sleep (2539.25 ± 65.47 µV²). This pattern may reflect the generally higher overall neuronal activity in the VTA during wakefulness, accompanied by more frequent and synchronous synaptic inputs, which in turn generate larger local field fluctuations. Consistent with this interpretation, previous studies have reported enhanced activation of VTA dopaminergic and glutamatergic neurons during wakefulness, which may represent the primary sources of increased LFP power49.

We further analyzed the relative contributions of the δ band (0.5–4 Hz) and θ band (4–8 Hz) to the total power spectral density (PSD) of LFP signals across different behavioral states, and compared these findings with simultaneously recorded EEG signals using the same analytical approach. We observed that both EEG and LFP signals during wakefulness and REM sleep were predominantly dominated by θ-band activity, whereas the NREM sleep stage was characterized by a δ-band–dominant spectral profile (Fig. S10b), consistent with previously reported EEG features of mouse sleep stages50. Interestingly, this state-dependent spectral distinction was even more pronounced in the finer-grained LFP signals (Fig. S10c), with LFPs exhibiting higher θ-band proportions than EEG during wakefulness and REM sleep, and higher δ-band proportions during NREM sleep. This difference likely arises from the distinct origins of EEG and LFP signals: EEG reflects the synchronous activity of large neuronal populations over widespread cortical areas, whereas LFP predominantly represents local synaptic and membrane currents within confined neuronal ensembles. From this perspective, LFP signals not only demonstrate high sensitivity to sleep state transitions but also hold promise as a precise tool for sleep stage classification.

Conclusion

This study addresses the critical bottleneck of lacking high spatiotemporal resolution evidence for neural coding mechanisms underlying sleep–wake regulation in deep brain regions by proposing and validating a PEDOT:PSS/PDA co-electrodeposition strategy. This one-step method fabricates neural electrode interface coatings that simultaneously exhibit excellent electrical performance, mechanical stability, and biocompatibility. Utilizing 16-channel microelectrode arrays modified with this coating, we achieved long-term, high spatiotemporal resolution monitoring of LFPs and spikes activity from the VTA of mice throughout sleep–wake cycles. We identified three distinct neuronal populations within the VTA characterized by state-specific firing patterns—wake-responsive, sleep-responsive, and state-independent neurons—revealing the diverse and fine-grained encoding of sleep–wake transitions by the VTA. Notably, LFP signals demonstrated more pronounced state-dependent spectral specificity than EEG during these transitions, highlighting the superior capacity of mesoscopic LFPs to capture dynamic neural activity details and their potential as a high-precision tool for sleep staging.

In summary, this work provides direct cellular-level evidence for the involvement of the VTA in modulating sleep–wake states and establishes a technical foundation for developing long-term stable, high-performance neural interfaces. These advances hold promise for advancing clinical neuromodulation therapies for sleep disorders and next-generation brain-machine interface technologies.