Detection of electrophysiological activities in the F7 area of the rhesus macaque cortex during recovery from prolonged anesthesia using implantable microelectrode array Open Access

Anesthesia plays a crucial role in modern medicine, providing effective blocking of sensation during surgery and ensuring that patients or animals remain pain-free, while also offering a stable physiological environment for the procedure.¹ However, the process of anesthesia involves complex physiological and neurobiological changes that can alter brain function. These changes are typically characterized by a gradual transition from conscious awareness to unconsciousness, followed by recovery once the anesthetic agent is withdrawn.

Among the various brain regions affected during anesthesia, the prefrontal cortex (PFC) is particularly sensitive.² The PFC shows pronounced vulnerability to anesthesia-induced disturbances. Notably, stimulating cholinergic neurons in the PFC³ or activating the central thalamus, which in turn activates the PFC,⁴ can induce awakening in anesthetized animals, highlighting the critical role of the PFC in maintaining consciousness. One specific region of the PFC, the F7 area, located in the premotor cortex, is implicated in motor planning, coordination, and sensorimotor integration. F7 neurons receive input from multiple brain areas, including the medial intraparietal (MIP) area, the dorsal part of the pre-cuneal cortex (V6A), the posterior cingulate cortex,^5–7 and others, positioning F7 as an integral hub in sensorimotor processing.

Understanding the neural mechanisms of anesthesia recovery, especially the effects on neurons within the F7 region, is critical for advancing our knowledge of the impact of prolonged anesthetic exposure on cognitive function. Furthermore, extended exposure to anesthesia may have neurotoxic effects, particularly when high concentrations are used, prolonged durations are involved, or when animals are still in early developmental stages.^8–11 These effects can have long-lasting consequences on brain function and cognition, making it important to study the neural recovery processes following anesthesia.

Although much research has focused on anesthetic effects in general,^12,13 there is a lack of detailed studies on the neural activity in the F7 region during anesthesia recovery. Current research in this area is limited, despite the fact that understanding the recovery of the PFC, particularly the F7 area, can provide insight into both basic neurobiology and clinical practice, particularly for surgeries involving long anesthesia durations. Moreover, it can also offer new perspectives for research on sleep arousal and the resting state, potentially unearthing novel connections between anesthetic-induced changes in the F7 area and these fundamental physiological states.

In terms of electrophysiological monitoring, traditional electrode techniques such as microwires and tungsten electrodes^14,15 have notable limitations for channel density and spatial resolution. As the field advances, there is an increasing need for multichannel high-spatiotemporal-resolution recordings to study cellular-level activity during anesthesia recovery. This is particularly relevant when exploring complex neural processes such as those in the F7 area, which involve intricate dynamics at the level of individual neurons.

In this study, we design and fabricate a novel 16-channel microelectrode array (MEA) modified with platinum black nanoparticles (PtNPs), which significantly improve its electrical performance. This enhanced MEA is used to record cellular-level electrophysiological signals during the recovery phase of anesthesia in the rhesus macaque. We focus on the F7 area, examining changes in neuronal firing rates and local field potentials (LFPs) during the process from anesthesia to recovery. Our results provide new insights into the activation of the F7 area, with increased firing rates corresponding to the recovery process, and show specific changes in LFP power across different frequency bands. These findings contribute to our understanding of how the F7 area and its neural activity recover after anesthesia, setting the stage for future studies of the brain’s recovery mechanisms following prolonged anesthetic exposure.

Chloroplatinic acid (H₂PtCl₆) and lead acetate [(CH₃COO)₂Pb] were obtained from Leyan Corporation (Shanghai, China). Saline was obtained from Double-Crane Pharmaceutical Company (Beijing, China). Phosphate-buffered saline (PBS) was obtained from Sigma-Aldrich (Shanghai, China).

Electrochemical modification and characterization were performed using Gamry Reference 600 (Gamry Instruments, USA). Physiological signals of the macaque during experiment, including heart rate, body temperature and respiratory rate, were monitored using a PM12B-CB monitor (Chengyitong Corporation, China). Other apparatus included an S-4800 scanning electron microscope (Hitachi, Japan), an M205C stereo microscope (Leica Biosystems, IL, USA), and a BX51TRF optical microscope (Olympus, Japan).

The length of the probe shank is 11 mm, allowing it to detect cortical brain regions. To minimize damage to brain tissue, the probe wire width is limited to 4 μm, resulting in a probe width of only 280 μm [Fig. 1(a)]. The probe thickness is ∼27 μm. The MEA has a total of 16 recording sites, each with a diameter of 16 μm and a spacing of 60 μm between adjacent sites. All sites are located at the edge of the probe for optimal detection of neuronal signals.

The electrode uses silicon-on-insulator (SOI, 25 μm Si/1 μm SiO₂/350 μm Si, Suzhou Yancai Corporation, China) as the substrate and is fabricated using microelectromechanical systems (MEMS) technology. Silicon has a much higher Young’s modulus than brain tissue, providing sufficient mechanical strength.¹⁶ Silicon-based electrodes have been validated for long-term electrophysiological signal recording.^17,18 First, the SOI substrate was cleaned with concentrated sulfuric acid, followed by thermal oxidation to form a 500 nm-thick silicon dioxide layer as the insulating layer. Subsequently, photolithography was applied, followed by sputtering of Cr to a thickness of 30 nm and Pt to a thickness of 250 nm. After liftoff, a conductive layer, including the sites and wires, was formed. Plasma-enhanced chemical vapor deposition (PECVD) was then used to deposit a 300 nm-thick silicon dioxide layer and a 500 nm-thick silicon nitride layer as insulating layers. Silicon nitride, as a passivation layer, effectively isolated the probe from the complex ionic environment of the brain, preventing degradation and ensuring sustained performance.^19–21 Next, photolithography and reactive ion etching (RIE) were applied to expose the electrode sites and pads for packaging. Photolithography and deep silicon etching were then used to remove the top silicon layer of the SOI, resulting in the electrode structure. The front of the SOI was protected with black adhesive, and then the silicon layer on the back of the SOI was removed by wet etching with 30% KOH at 80 °C. After dissolution of the black adhesive, the MEAs were released. Finally, backside dry etching was applied to the MEAs to etch the buried oxide layer and release residual stress, resulting in low-curvature MEAs.²² 

The impedance of the electrode sites can affect the baseline noise in electrophysiological recordings.²³ To reduce the electrode impedance, PtNPs were modified on the surface of the electrode sites. First, 1 g of chloroplatinic acid was dissolved in 40 ml of PBS solution, and 64 mg of lead acetate was dissolved in 40 ml of PBS solution. The two solutions were then mixed and left to stand for 12 h, followed by filtration. PtNPs were deposited by chronoamperometry (−1 V, 60 s). The surface morphology of the modified electrode sites is shown in Figs. 1(b) and 1(c).

One male rhesus macaque (11 years old, weighing 10.4 kg) was used in this study. The rhesus macaque was housed in an individual primate cage (Lingfu Biotechnology Company, Beijing, China). All animal procedures followed the guidelines set by the local animal welfare committee.

Electrode implantation was performed under stereotaxic coordinates guided by magnetic resonance imaging (MRI). MRI images were obtained using a 3.0 T uMR790 imaging system (United Imaging, China).

Anesthesia was induced in the rhesus macaque using isoflurane gas (1%–3%). The animal was then positioned in an appropriate stereotaxic frame. A semi-elliptical incision was made along the midline of the scalp, followed by retraction of the skin, fascia, and muscle to expose the skull. The periosteum was scraped off, and a craniotomy was performed at an approximate coordinate reference to the supraorbital ridge. On the basis of the MRI images, a small cranial window (4 × 4 mm²) was created over the F7 area of the prefrontal cortex by carefully removing a section of the skull. The dura mater was incised and reflected to expose the cortical surface.

After 9 h of anesthesia (other procedures conducted during the 9 h anesthesia period did not affect this study), the MEA was slowly implanted into the F7 area. A platinum wire was placed between the skull and cortex to serve as a ground. After recording neural signals under anesthesia for a designated period, the isoflurane gas was discontinued, and the monkey was allowed to breathe room air. To prevent electrode breakage and potential injury due to sudden movements, anesthesia was reintroduced, recording was stopped, and the electrode was withdrawn upon detecting slight movements. At the end of the experiment, a gelatin sponge was placed in the cranial cavity, and the remaining exposed skin was sutured to close the wound.

Throughout the procedure, body temperature was maintained using a heating pad, while heart rate, body temperature, respiratory rate, and oxygen saturation were continuously monitored. After surgery, the monkey was administered ampicillin at a dose of 40 000 units per kg of body weight as an antibiotic treatment.

All surgeries were conducted under sterile conditions in a certified animal surgical facility. All procedures were approved by the Lingfu Biotechnology Company (LFSW-DWPZ-20240416-01). All the procedures described above adhered to the guidelines of the State Scientific and Technological Commission for the care and use of laboratory animals.

The grounding platinum wire was placed between the skull and cortex of the rhesus monkey to minimize bioelectrical noise, and the differential method between the recording site and the reference electrode [the rectangular strip in Fig. 1(a)] helped to filter out interference from respiratory rhythms, movement, and other bioelectrical signals. Electrophysiological signals were initially amplified, filtered, and digitized using a custom-designed 32-channel headstage based on the RHD2132. The data were then transmitted to a computer via an FPGA Spartan 6-based data acquisition system (RHD USB Interface Board, Intan, USA).

Electrophysiological signals were recorded at a sampling rate of 30 kHz (with 50 Hz power line interference filtered out). The obtained signals were low-pass filtered at 200 Hz, and LFP signals were obtained through down-sampling to 1 kHz. Spike signals were extracted from the high-pass filtered signals at 200 Hz using a thresholding method within a 3 ms time window and were aligned to the local minima.

Spike clustering analysis was performed using Offline Sorter software, power spectral density (PSD) analysis was conducted with Neuro Explorer software, and significance testing was carried out using Origin software. Other data analyses were performed using Python.

After the fabrication of the MEA, the electrode surface was further modified with nanomaterials to reduce the electrode impedance. Lowering the electrode impedance reduces thermal noise and improves the signal-to-noise ratio (SNR), thereby enhancing the ability to detect weak neural signals.²⁴ Additionally, reduced impedance facilitates better charge transfer, leading to more stable and reliable electrophysiological recordings.²⁵ As shown in Figs. 1(b) and 1(c), after the electrode site was modified with PtNPs, they became rough and porous. This structure significantly increased the specific surface area of the electrode sites, leading to a reduction in electrode impedance.

Electrochemical impedance spectroscopy (EIS) measurements were performed in PBS solution, and the results are shown in Fig. 2. The power of neuronal action potentials is mainly concentrated around 1000 Hz. At this typical frequency, after modification with PtNPs, the electrode impedance decreased from 590.9 ± 77.5 to 22.5 ± 15.2 kΩ [Fig. 2(c)], and the phase shifted from −81° ± 0.8° to −33.5° ± 4.1° [Fig. 2(d)], resulting in a significant enhancement of the electrical performance of the electrodes.

After acquisition of coronal MRI images, these were compared with the rhesus macaque brain atlas to determine the target brain area for implantation. The probe implantation site is shown in Fig. 3(a). During the prolonged anesthesia, the macaque’s body temperature maintained constant, its heart rate remained normal, and its respiratory rate was at a relatively low level. Subsequently, the electrode was slowly implanted into the macaque’s F7 area. After recording for a designated period, anesthesia was discontinued, and the macaque was allowed to breathe air. The macaque’s movements were closely monitored, and if any slight movement was detected, anesthesia was reintroduced and recording was stopped. After removal of the MEA, the wound was sutured, and the experiment was concluded.

The macaque’s respiratory rate during the anesthesia-to-recovery process is shown in Fig. 4(a). During the 0–240 s period, the monkey breathed anesthetic gas, and its respiratory rate remained at a low level. Afterward, the monkey began to breathe air. In the 240–545 s period, the respiratory rate gradually and steadily increased, termed recovery 1. From 545 to 660 s, the respiratory rate rapidly increased and fluctuated significantly, termed recovery 2. At 660 s, slight movements of the monkey were observed, and anesthesia was immediately reinstated, halting electrophysiological signal recording.

The recorded neuronal firing activity is shown in Fig. 4(b). During anesthesia, neuronal firing activity was relatively low. During the recovery 1 period, neuronal firing activity increased and was concentrated around the time of rapid changes in respiratory rate. During the recovery 2 period, neuronal firing activity was inconsistent, but overall it displayed more intense discharges.

Figure 4(c) shows the LFPs from different channels. During anesthesia, the LFPs exhibited small fluctuations, reflecting a state of neural activity silence after prolonged anesthesia. In the recovery 1 period, the LFP fluctuations increased, mainly showing large low-frequency variations, with relatively consistent fluctuations across channels. In the recovery 2 period, the differences between the LFPs across channels became more pronounced, with high-frequency low-amplitude oscillations appearing in channels 1 and 2, while channel 3 continued to display low-frequency high-amplitude fluctuations.

As shown in Fig. 5(a), the monkey’s respiratory rate increased significantly from anesthesia to recovery 1, and then further increased during recovery 2. To analyze the characteristics of the population of neurons in the F7 area during anesthesia recovery, the average firing rate curve of the population neurons was calculated, as shown in Fig. 5(b). The firing rate of the population neurons exhibited a gradual increase, consistent with the changes in the respiratory rate. This suggests that the neurons in the F7 area are activated during the recovery process from anesthesia. Figure 5(c) shows the average PSD of LFPs from multiple channels. Each column in the figure represents the power spectral distribution from 0 to 100 Hz within a 5 s interval, with red indicating higher power levels. During anesthesia, the power spectrum of LFPs is predominantly concentrated in the low-frequency range. By contrast, during recovery 1 and recovery 2, the power spectrum of LFPs shows an increase in high-frequency activity compared with the anesthesia period. Especially during recovery 2, around 300 and 460 s, the power of LFPs in the high-frequency range increases significantly.

Owing to the significant differences in respiratory rates across the three anesthesia periods, the neuronal firing activity was compared laterally across the different periods. The average firing rate of the population of neurons significantly increased with recovery time [Fig. 6(a)]. Figure 6(b) shows the PSD of LFPs across different periods. Overall, the PSD in the 0–100 Hz range during recovery 1 and recovery 2 is higher than that observed during anesthesia. Between ∼0 and 10 Hz, the PSD in recovery 2 is slightly higher than that in recovery 1, whereas between 10 and 100 Hz, the PSD in recovery 2 is slightly lower than that in recovery 1. To further quantify this difference, the power across different frequency bands was analyzed (δ band 0–4 Hz, θ band 4–8 Hz, α band 8–13 Hz, β band 13–30 Hz, γ band 30–100 Hz). Figure 6(c) shows that the power of LFPs and the δ band increased gradually. The power in the θ band increased during recovery 1 compared with anesthesia and remained similar to recovery 1 during recovery 2. By contrast, the power in the α, β, and γ bands first increased and then decreased. The changes in neuronal firing rate are consistent with the variations in the power of LFPs. The sustained increase in the θ band suggests that this frequency range may play an important role during the early stages of anesthesia recovery.

In this study, a silicon-based microelectrode array was designed and fabricated, and, after modification with PtNPs, successfully recorded neural electrophysiological signals at the cellular level from the monkey cortex. The PtNP-modified electrode sites showed very low electrode impedance (22.5 kΩ), which is better than that achieved with other modification materials (149 kΩ with TiN and 81 kΩ with TiW/Pt),^26,27 demonstrating superior electrical performance. In addition, PtNPs can maintain stable impedance and neural electrophysiological signal recording capabilities for several months.^28,29 Traditional microwire electrodes typically have site diameters of ∼80 μm,³⁰ while tetrode electrodes consist of four closely packed 25 μm-diameter microwires,^31,32 making precise positioning over a larger spatial area challenging. By contrast, while MRI offers a broad detection range, its temporal resolution is around 350 ms, and its spatial resolution is ∼2.5 mm.^33,34 In this study, the MEA comprised 16 channels, with a site diameter of 16 μm and a minimum intersite spacing of 60 μm. The sampling rate was 30 kHz, corresponding to a temporal resolution of 33.3 μs. Compared with traditional microwire electrodes, our MEA provides a broader detection range, and compared with MRI, it offers a significantly higher spatiotemporal resolution with equipment that is more economical and convenient.

During the process from anesthesia to recovery, the monkey’s respiratory rate gradually increased, and neuronal firing activity also increased. The LFPs shifted from low-frequency, small-amplitude oscillations to low-frequency, large-amplitude oscillations, and later to high-frequency, low-amplitude oscillations, with power gradually increasing. These activities corresponded to the changes in the respiratory rate, suggesting that the F7 area was activated during the recovery from anesthesia. This result is consistent with previous studies.² In the recovery 1 period, the spikes and LFP activity across channels showed a degree of consistency, while in recovery 2, differences in spikes and LFP activity between channels were observed. This was in line with the significant fluctuations in respiratory rate during recovery 2 and may suggest the gradual recovery of consciousness.

Analysis of the LFPs across frequency bands showed that from anesthesia to recovery, the δ band steadily increased, reflecting the reduction in anesthetic depth. During recovery 1, there was a significant increase in the power of the θ, α, β, and γ bands, resembling the high-frequency oscillatory states observed in a conscious state.³⁵ However, in recovery 2, the power of the α, β, and γ bands decreased to varying extents, indicating that the recovery of neuronal activity after prolonged anesthesia is not immediate. Since recording was stopped after the detection of slight movements of the monkey, this phase might correspond to the early stages of recovery from anesthesia.

The lack of a significant decrease in the θ band during recovery 2 suggests its important role in the process of consciousness awakening. Additionally, research suggests that prolonged anesthesia may impair the animal’s cognitive function,⁸ and so the decline in high-frequency bands during recovery 2 may also indicate neuronal damage after long-term anesthesia, with the need for more time to restore normal activity.

This study monitored the changes in electrophysiological signals from anesthesia to recovery following prolonged anesthesia. However, owing to the limitations of acute experiments, longer recovery periods could not be observed. Further studies, possibly involving implanted systems to monitor neuronal activity throughout the transition from anesthesia to full wakefulness, are needed to better understand the discharge mechanisms of the prefrontal cortex during recovery from anesthesia.

This work has presented a novel 16-channel silicon-based MEA designed for cortical studies in nonhuman primates. After fabrication of the MEA using MEMS technology and its modification with PtNPs, the electrode impedance at 1000 Hz was significantly reduced from 590.9 ± 77.5 to 22.5 ± 15.2 kΩ, greatly enhancing the electrical performance of the MEA. Electrophysiological recordings in the rhesus macaque validated the electrode’s ability to detect high-quality neural signals, allowing the detection of cellular-level neural activity with high spatial and temporal resolution. Analysis of neural information during the anesthesia recovery process revealed that neurons in the F7 area were activated during recovery. As time progressed, the spikes rate gradually increased, and in the recovery 1 period, the neuronal firing activity synchronized with the rapid increase in respiratory rate. The characteristics of spikes and LFPs were analyzed across different periods, and it was observed that both spikes and LFP activity increased synchronously over time. Additionally, LFPs exhibited higher PSDs in the δ and θ bands during recovery 1 and recovery 2, while the power in the α, β, and γ bands initially increased and then decreased across the three periods. This study provides a new tool for detecting neural information at the cellular level in nonhuman primate cortex and lays the foundation for future research into the neural mechanisms underlying the recovery process of F7 area neurons following prolonged anesthesia.

This work was sponsored by the National Key R&D Program of China (Grant Nos. 2022YFC2402500 and 2022YFB3205602), the National Natural Science Foundation of China (Grant Nos. 62121003, T2293730, T2293731, 62333020, 62171434, and 62471291), the Major Program of Scientific and Technical Innovation 2030 (Grant No. 2021ZD02016030), the Joint Foundation Program of the Chinese Academy of Sciences (Grant No. 8091A170201), and the Scientific Instrument Developing Project of the Chinese Academy of Sciences (Grant No. PTYQ2024BJ0009). We thank Lingfu Biotechnology Company and Yuefeng Li for their support and assistance in the surgery and experiment.

The authors have no conflicts to disclose.

All procedures were approved by the Lingfu Biotechnology Company (LFSW-DWPZ-20240416-01). All the study methods were carried out in accordance with the guidelines of the State Scientific and Technological Commission for the care and use of laboratory animals.

W.X. designed the experiments, fabricated the devices, analyzed the data, and drafted the manuscript. J.L. and J.S. contributed to the study concept. Y.L. and S.L. assisted with the data analysis. M.L., P.J., and S.Z. assisted with the device fabrication. L.J., Z.X., D.Z., and M.W. assisted with the experiments. Y.Y., Y.S., and X.C. guided the theoretical analysis, reviewed the manuscript, and checked it for final submission. All of the authors have approved the submitted version.

Wei Xu: Conceptualization (lead); Data curation (equal); Formal analysis (equal); Investigation (lead); Methodology (lead); Software (equal); Visualization (lead); Writing – original draft (lead); Writing – review & editing (lead). Jinping Luo: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal). Jin Shan: Conceptualization (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Resources (equal); Validation (equal). Yaoyao Liu: Conceptualization (equal); Formal analysis (equal); Resources (equal). Shiya Lv: Data curation (equal); Formal analysis (supporting); Funding acquisition (supporting). Ming Li: Data curation (supporting); Investigation (supporting); Methodology (equal). Peiyao Jiao: Data curation (supporting); Formal analysis (supporting); Methodology (supporting). Siyu Zhang: Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Methodology (supporting). Luyi Jing: Data curation (supporting); Methodology (supporting). Zhaojie Xu: Data curation (supporting); Formal analysis (supporting); Investigation (supporting). Di Zang: Methodology (supporting). Mixia Wang: Conceptualization (supporting); Data curation (equal); Formal analysis (supporting); Methodology (supporting). Yanbing Yu: Funding acquisition (equal); Resources (equal). Yilin Song: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Methodology (equal); Project administration (equal). Xinxia Cai: Conceptualization (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Project administration (lead); Resources (lead); Supervision (lead); Validation (equal).

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Wei Xu received a B.E. degree from the University of Science and Technology Beijing in 2020 and is currently pursuing a Ph.D. degree in the State Key Laboratory of Transducer Technology, Aerospace Information Research Institute, Chinese Academy of Sciences. His research interests are mainly in microelectrode arrays for in vivo brain–machine interface.

Yanbing Yu earned his Master’s degree from Peking Union Medical College in 1999. He currently serves as Director of Neurosurgery at the China–Japan Friendship Hospital and is a Ph.D. supervisor. A pioneer in minimally invasive cranial nerve therapies, he has performed over 30 000 microvascular decompression (MVD) procedures and has developed internationally recognized techniques. He has also pioneered selective peripheral neurotomy for cerebral palsy, with over 10 000 cases. Leading more than ten national research projects and authoring over two hundred publications, Professor Yu has received numerous honors, including the National Outstanding Physician Award and the State Council Special Allowance, and contributes to medical aid in Tibet and Xinjiang.

Yilin Song achieved her Ph.D. degree in Bioelectronics from the Institute of Electronics, Chinese Academy of Sciences in 2011. She is now an Associate Professor in the State Key Laboratory of Transducer Technology, Aerospace Information Research Institute, Chinese Academy of Sciences. Her research interests include micro/nano biosensors, brain–machine interface, and neural information detection technology.

Xinxia Cai received a Ph.D. degree from the University of Glasgow in 2001. She is a Distinguished Professor at the Aerospace Information Research Institute of the Chinese Academy of Sciences, a winner of the National Outstanding Youth Fund, and the academic leader of the Micro/Nano Sensing Technology Innovation Research Group of the National Natural Science Foundation of China (NSFC). For a long time, she has focused mainly on micro/nano biosensors and microsystems such as brain–computer interface neural microelectrode arrays, and she is the Principal Investigator of the major project of the Interdisciplinary Department of the NSFC.