Here, we present a method for the long-term performance and safety assessment of soft subdural electrode arrays in a minipig model, describing surgical method and tools, postoperative magnetic resonance imaging, electrophysiology of the auditory cortex, electrochemical properties of the implant, and postmortem immunochemistry.
Neurological impairments and diseases can be diagnosed or treated using electrocorticography (ECoG) arrays. In drug-resistant epilepsy, these help delineate the epileptic region to resect. In long-term applications such as brain-computer interfaces, these epicortical electrodes are used to record the movement intention of the brain, to control the robotic limbs of paralyzed patients. However, current stiff electrode grids do not answer the need for high-resolution brain recordings and long-term biointegration. Recently, conformable electrode arrays have been proposed to achieve long-term implant stability with high performance. However, preclinical studies for these new implant technologies are needed to validate their long-term functionality and safety profile for their translation to human patients. In this context, porcine models are routinely employed in developing medical devices due to their large organ sizes and easy animal handling. However, only a few brain applications are described in the literature, mostly due to surgery limitations and integration of the implant system on a living animal.
Here, we report the method for long-term implantation (6 months) and evaluation of soft ECoG arrays in the minipig model. The study first presents the implant system, consisting of a soft microfabricated electrode array integrated with a magnetic resonance imaging (MRI)-compatible polymeric transdermal port that houses instrumentation connectors for electrophysiology recordings. Then, the study describes the surgical procedure, from subdural implantation to animal recovery. We focus on the auditory cortex as an example target area where evoked potentials are induced by acoustic stimulation. We finally describe a data acquisition sequence that includes MRI of the whole brain, implant electrochemical characterization, intraoperative and freely moving electrophysiology, and immunohistochemistry staining of the extracted brains.
This model can be used to investigate the safety and function of novel design of cortical prostheses; mandatory preclinical study to envision translation to human patients.
Neurological impairments and diseases can be diagnosed or treated using electrocorticography (ECoG) arrays. These electrode grids are implanted at the surface of the brain and allow for recording or stimulation of the human cortex1. In the case of drug-resistant epilepsy, for example, they help delineate the epileptic region to resect2. In long-term applications such as brain-computer interfaces, these epicortical electrodes are used to record the movement intention of the brain, to control the robotic limbs of paralyzed patients3. However, current electrode grids are made from stiff metallic blocks embedded in rigid polymeric substrates and do not answer the need for high-resolution brain recordings and long-term subdural biointegration (>30 days). Rather, they create local tissue reactions that lead to fibrotic encapsulation of the implanted device, leading to worse performance over time. Recently, flexible or stretchable electrode arrays using thin polymeric substrates manufactured by microfabrication techniques have been proposed to achieve high performance in long-term implantations by limiting the tissue reaction4,5. However, preclinical studies for these new implant technologies are needed to validate their long-term functionality and safety profile, so that translation to human patients may be envisioned. In this context, minipig and pig models are routinely employed in the development of devices in other medical contexts (e.g., the cardiovascular, skeletal, or gastric systems) due to their large organ sizes and easy animal handling6,7,8. However, only a few applications targeting the brain for neurophysiology are described in the literature, mostly due to surgical approach limitations and integration of the implant system on a living animal9,10,11,12. These are often not compatible with chronic implantation in living animals, as they would require, for example, the development of complex hardware such as implantable embedded electronics. Additionally, they do not investigate the influence of the implant system on the target tissue, which is crucial for the biosafety aspect in translational studies. The porcine model is close to human anatomy in terms of cortical structure, skull bone, and skin thickness13. Furthermore, their ability to learn behavioral tasks makes them a powerful model for investigating functional rehabilitation strategies or sensory perceptions14.
The translation of new technologies and therapies to humans necessitates the assessment of safety and efficacy, as required by competent medical authorities. These are usually described in technical documents and norms15, however they only require the passing of these tests and do not investigate the actual effect of the device implantation or collection of other useful data in parallel to the safety study. For a complete biosafety and performance study on the brain, we present here a longitudinal and systematic collection of brain imaging data, electrophysiological measurements, assessment of electrochemical properties of the implanted electrodes, and postmortem histology in a porcine model. To achieve this, several aspects need to be considered, in order to create a complete experimental model: (i) minimally invasive surgical access for device implantation together with a mechanically stable transdermal port to connect to the electrodes, (ii) a robust electrophysiological recording paradigm that serves as performance output for the implanted electrodes both under anesthesia and in freely moving conditions, (iii) in vivo imaging (computerized tomography [CT] and/or magnetic resonance imaging [MRI]) at different time points to follow the evolution of the brain and implant, as well as the compatibility of the implanted system with the imaging equipment, and (iv) a tissue preparation pipeline to extract the brain for histological analysis.
Here, we report on the method for long-term implantation (6 months) and evaluation of soft ECoG arrays in the minipig model (shown schematically in Figure 1). The soft electrode arrays were presented in our previous reports and are made from thin silicone membranes embedding elastic gold thin films used as electrical tracks16,17. The contact with the tissue is made through a mix of platinum nanoparticles embedded in a silicone matrix for a soft and efficient electrochemical interface to the brain tissue18. The implants are connected through a flexible cable tunneled subdurally through the skull and the skin to a transdermal port that houses the connectors on the head of the animal. The size and shape of the implant can be customized according to the target and the needs of the study. The current electrode strips in this study mirrors the real size of the clinical strips. Clinically available subdural strips and grids were used as comparators using the same approach. The polymeric MRI compatible transdermal port is placed on the skull using a footplate system that anchors it firmly to the skull. Here, we describe in detail the surgical procedure, from subdural implantation of both hemispheres to recovery of the animal. We focus on the auditory cortex as an example target area, where evoked potentials are induced by acoustic stimulation both in anesthetized and freely-moving conditions. At different time points, the animal's brain is imaged in MRI (or CT for the clinical electrodes) under anesthesia and the electrochemical properties of the electrodes are measured. Electrode characterization methods are used to follow the evolution of the implant and the electrode-tissue interface (see Schiavone et al.19 for more details). These include chronoamperometry to probe the stimulation abilities of the electrode contact, electrochemical impedance spectroscopy (EIS) that can indicate the evolution of the resistive and capacitive components of the electrode, and inter-channel resistance measurements to probe for hermetic encapsulation failures. Finally, we have developed a tissue extraction pipeline to perfuse the brain after euthanasia, explant it with the electrodes in place, section it, and perform histological analysis using different inflammation markers. Overall, this method will allow preclinical studies with robust multimodal data collection for future clinical translation of new technologies and therapies on the brain.
Surgical and behavioral procedures were approved by the local ethical committee in accordance with the guidelines for the Care and Use of Laboratory Animals and approved by local (Canton of Geneva) and federal (Swiss) veterinary authorities with authorization number GE11120A. Female Göttingen minipigs (n = 7) at 2-6 months of age (5-8 kg) were used in this study.
1. Presurgical planning
2. Surgical implantation of soft ECoG arrays
3. In vivo characterization of the soft implant
4. Electrophysiological recording
5. In vivo imaging
6. Freely moving recording
7. Perfusion and tissue preparation
8. Histology
In order to confirm the placement (Figure 3A) and functionality of the devices, electrophysiological recordings are performed intraoperatively after pedestal placement. The baseline signal is first acquired over 2 min with no stimuli as the control of basal activity. Secondly, the animal is acoustically stimulated with a tone burst at different frequencies (500-20,000 Hz), and the raw data is averaged over the stimulus period to map auditory evoked potentials across the array (e.g., at 800 Hz compared to baseline; Figure 3B). The data shown here are unprocessed, but if too much noise is present, notch and bandpass filters can be applied. Typical sources of noise in the surgical theater include heating pads, plugged drills, and suction or cauterizers (among others) that should be removed prior to acquisition. In awake recordings, large muscle movement around the head, such as chewing, should be avoided for cleaner data sets.
This protocol was applied at every recording time point, and signals for a single channel could be compared over time. One example is illustrated in Figure 3C, showing the robustness and evolution of the response. The recording capacity of each contact over the time course of the experiment can be evaluated by calculating the standard deviation of the baseline signal at every time point (Figure 3D). In this study, the signal-to-noise ratio decreased and settled between day 0 and month 6, despite some variability due to the limited duration of the recording period (i.e., 2 min). This can be further correlated to electrode impedances.
The in vivo imaging is performed postoperatively to assess the brain state and implant positioning. In the first iteration of the protocol, no intraoperative x-ray was performed, resulting in a folded device, as is visible in Figure 4A on a T1-weighted MRI sequence (see in addition Figure 4B). No behavioral change was observed in the animal, but over time, this resulted in a fibrotic encapsulation around the device due to the macroscopic compression of the brain around the implant location (Figure 4C). After this experience, intraoperative x-ray was introduced, as shown in Figure 4D, where the radiopaque markers (black bars visible on the implant in inset Figure 4D) are shown to be well positioned. The surface of the brain is then intact, as can be observed in the postoperative MRI in Figure 4E. Overall, with this implant and pedestal system, whole-brain imaging is possible. Different sequences in the coronal planes enable to see anatomical structures (Figure 4F,G; T1 and T2 MRI sequences) or the presence of liquid and blood around the implant (Figure 4H; TSE-weighted MRI sequence). The pedestal system creates almost no artifacts, except for some small black-contrasted voids around the titanium screws (see Figure 4G). Additionally, clinical electrodes are used as comparators in this study, but cannot be imaged in the MRI due to heating and safety concerns. Therefore, CT scans are acquired on these animals, as shown in Figure 4I. The electrodes are clearly visible, and the pedestal system is not influencing the image quality.
After the implantation period, the animal is perfused, and the brain extracted. In this study, the analysis of the inflammatory response is performed on each hemisphere independently. Cutting the brain in half is easier for tissue preparation before sectioning, and has the advantage that sections can be mounted on standard microscopy slides. One example of a brain sample is shown before (Figure 5A) and after (Figure 5B) cutting in blocks. The outline of the implant is clearly visible and has created a small dent in the brain. By cutting in parallel planes, the tissue is then already aligned to the cryostat, and sections can readily be cut without tissue loss for trimming (Figure 5C). After staining, the whole tissue section is imaged (Figure 5D), where for example, the neuron layer is clearly visible in detail (see NeuN marker). Whole sections are fragile and can sometimes lead to some loss of tissue (see the bottom of Figure 5D), but the area of interest is intact. On a closer view, enabled by confocal microscopy imaging at 40x, the cells are clearly defined and enable fine investigation of inflammatory markers, for example (Figure 5E). Further quantifying analysis can be performed to compare inflammation between control and implanted hemispheres. Figure 6 shows the electrochemical characterization of the implanted electrodes. The In vitro electrochemical impedance spectroscopy of the soft electrode array with impedance modulus and phase is shown in Figure 6A and the impedance modulus at 1 kHz over 6 months of implantation is shown in Figure 6B.
Figure 1: Schematic of the experiment. Please click here to view a larger version of this figure.
Figure 2: Minimally invasive implantation of soft ECoG onto the brain. (A) Surgical access to the skull, with indication of bregma. (B) Bilateral craniotomy with visible dura mater. (C) Slit durotomy on the first hemisphere. (D) Subdural implantation of soft ECoG and dura mater closure. (E) Slit durotomy on the second hemisphere. Bone flap fixation on the first hemisphere using titanium bridges. (F) Implantation of soft ECoG on the second hemisphere and dura mater closure. (G) Bone flap fixation on the second hemisphere. (H) Footplate positioning on the skull. (I) Pedestal fixation onto the footplate. (J) Skin closure around the pedestal base. Please click here to view a larger version of this figure.
Figure 3: Recording of auditory evoked potentials. (A) Schematic of electrode placement at the surface of the temporal lobe. (B) Representative mapping of baseline activity (gray traces) and auditory evoked potentials in response to an 800 Hz tone burst stimulation (purple trace). Each average corresponds to one channel on the soft ECoG array. The averaging is triggered on the analog input signal from the sound stimulation. "ON" and "OFF" acoustic stimulation periods are noted on one channel in the bottom left. (C) Evolution over time (day 0, month 2, and month 5) of a single channel response after acoustic stimulus, compared to baseline signal when no stimulus is presented (gray). The averaging is triggered on the analog input signal from the sound stimulation. The "ON" and "OFF" stimulation periods are noted at the bottom. The evoked potential of the "ON" stimulation is marked with arrows. (D) Standard deviation per channel (colored dots) per time point of the baseline recording. Median values are represented in bold blue. Please click here to view a larger version of this figure.
Figure 4: In vivo imaging of the brain and implanted electrodes. (A) Postoperative T1-weighted MRI in the coronal plane. An arrow indicates a folded implant. (B) Magnified portion of A, where the folding of the implant creates a dent in the brain. (C) T1-weighted MRI at 1 month implantation, showing compression of the brain due to the fibrotic encapsulation of the brain at the same location as C. (D) Intraoperative plane x-ray verifying implant placement and no folding, as observed by the radiopaque marker placement. Inset: Photograph of implant with radiopaque marker visible. (E) Postoperative T1-weighted MRI in the coronal plane with optimal implant placement. (F) T1-weighted MRI at 1 month implantation. (G) T2-weighted MRI at 1 month implantation. An arrow shows the imaging artifact from the titanium screws holding the footplate in place on the skull. (H) TSE-weighted MRI at 1 month implantation. (I) CT scan of the animal implanted with the clinical electrodes. Please click here to view a larger version of this figure.
Figure 5: Histology analysis of the brain after long-term implantation. (A) Photograph of an explanted and perfused brain-left hemisphere. (B) Perfused brain cut in blocks prior to the freezing step. (C) Picture of whole block sectioning setup on the cryostat; the entire "pre-cut blocks" can be sectioned. (D) Immunostaining imaging of the whole hemisphere (slide scanner, 20x objective) and(E) zooming on the first layers of the cortex (confocal imaging, 40x objective) showing glial cells, astrocytes, and neurons. Please click here to view a larger version of this figure.
Figure 6: Electrochemical characterization of the implanted electrodes. (A) In vitro electrochemical impedance spectroscopy of the soft electrode array (small grey lines for each channel, the average in red) with impedance modulus (top) and phase (bottom). (B) Evolution of the impedance modulus at 1 kHz over 6 months of implantation (mean in blue; grey lines are the individual channels; the in vitro measurement is given as reference in red). Please click here to view a larger version of this figure.
Supplementary Figure 1: MRI-compatible pedestal. (A) Chronic MRI-compatible transdermal connection system (pedestal) to access the soft electrode array. (B) Pedestal with electrodes mounted on the footplate for skull anchoring. Inset: Details of the footplate. Please click here to download this File.
Supplementary Figure 2: Surgical access for optimal perfusion of the brain. (A) Skin cut and access to the location of the carotid artery and jugular vein. (B) Dissection of the tissue around the blood vessels. (C,D) Identification and dissection of the tissue around the carotid artery and jugular vein. (E) Isolation of the carotid artery from the tissue beneath. (F) Isolation of the jugular vein from the tissue beneath. (G) Suture wire placement around the carotid artery (suture 1 and suture 2) and the jugular vein (suture 3). (H) Closure of suture 3 at the base of the carotid artery (heart side) to avoid bleeding during opening of the vessel. (I) Clamping of the carotid artery at the opposite side from H. (J) Sectioning of the carotid artery. (K) Inserted catheter in the opening from J. Inset: Primed catheter with saline flushed from a syringe to the catheter tip. (L) Closure of suture 2 to maintain the catheter in place and along the artery. Please click here to download this File.
Supplementary File 1: Parameters for T1- (pages 1-2), T2- (pages (3-4) and TSE-weighted (pages 5-6) MRI sequences, respectively. Please click here to download this File.
Supplementary File 2: Metadata for slide scanner for whole slide imaging of stained brain slices. Please click here to download this File.
Supplementary File 3: Metadata for confocal imaging of magnified section of stained brain slices. Please click here to download this File.
We report here a method for long-term implantation and evaluation of soft ECoG arrays. In this study, we have designed a consistent, minimally invasive surgical approach for bilateral implantation of functional electrode grids over the temporal lobes (here, targeting the auditory cortex). We first evaluated the functionality of the grid by successfully recording evoked potentials over the time course of the study (6 months) and tracking the electrochemical properties of the electrodes (see Figure 6). Secondly, we assessed the biosafety of the grids, in vivo by using MRI and establishing a fully MRI-compatible system, and postmortem by designing a protocol for tissue collection and immunostaining.
To minimize invasiveness, we optimized the size of the craniotomy window. In order to reach the auditory cortex located on the temporal lobe and to avoid resecting the temporal muscle, we have developed a technique to slide the implant under the dura. This technique allows to drastically reduce the surface of the exposed brain and still reach far-away targets. While this type of implantation may seem blind, the implementation of radiopaque markers on the devices that are visualized in the intraoperative plane x-ray allows for verification of positioning, and ensures the array is not folded under the dura mater. The subdural sliding has proven to be safe in most of the repetitions we have performed. Additionally, the durotomy in a slit approach minimizes brain bulging during the time the craniotomy is open and facilitates the closure around the implant without requiring additional material such as artificial dura mater, which could bias the inflammatory response. Finally, the strength of this surgical approach is its ability to be transposed to different cortical regions. Playing with coordinates, the craniotomy position, and the device size, which can all be adjusted, enables this method to target most of the cortex area.
The surgical method presented here, along with functional assessment and investigation of the biointegration over time, is not limited to the soft electrode technology used in this report. Other subdural electrodes that are being developed for human translation could be evaluated with the same protocol. The strength of this method relies on the fact that most of the pieces, such as the cable and pedestal, are modular, personalizable, and can be adapted to the specific device under test. Additionally, intracortical or deep penetrating probes could also be used instead of or in combination with the subdural electrodes, as this only requires adjusting the craniotomy and durotomy geometry. The long-term results can then be compared to their clinical counterparts, as we have done here.
One of the major limitations of the presented method is the presence of skull sinuses in minipigs, which develop over the course of the first year12. In that regard, important aspects to take into account include the age of implantation and also the size of the animal. Performing craniotomies in the adult skull breaks sinuses' integrity and leads to a high risk of major infection in chronic settings. Such sinuses are visible in the plane x-ray and CT scan preoperatively. On the other hand, performing chronic implantation too early, in an animal that is too small, is also not optimal when the skull is undergoing massive growth and remodeling. We hypothesized that these "skull movements" post-surgery could cause the implant to move and fold, which ultimately is detrimental to the experiment. We have found here that Göttingen minipigs, approximately 5-6 months old (and 8 kg) at the time of implantation, should give the best results.
For evaluating the performance of the implanted ECoG for electrophysiological recordings, we have set up a rapid protocol for auditory evoked potential (AEP) recording that can be used in freely moving animals and under sedation. It consists of presenting a series of acoustic tone bursts at specific frequencies over the course of a few minutes. The advantage of such a protocol is the fact that it can be tuned to the available length of recording by reducing the number of frequencies probed. One challenge when recording cortical signals under anesthesia is that the level of consciousness of the animal should be taken into account when analyzing and comparing the data.
The protocol for perfusion was adjusted over time by observation of the extracted brain's quality. Indeed, we found it easier to catheterize the carotid artery only, and not the jugular vein. Initially, the literature presents methods where the jugular vein is catheterized to drain waste20. Practically, this limits the flow out of the brain and leads to poorer extraction of blood and the overall quality of perfusion. By cutting the jugular vein and leaving the liquid to escape in a large container where the animal lies, the efficiency of the perfusion increases.
We have developed a robust tissue preparation method that works with antibodies routinely used for inflammation tracing. We have separated the two hemispheres for practical reasons, as half the pig's brain fits on standard microscope slides and is thus compatible with most imaging equipment available in histology laboratories. By cutting the brain in blocks, direct access to the zone of interest is made possible without requiring further cutting of the whole brain or trimming extensive parts of the tissue. The brain slices at 40 µm can be pooled in standard well plates and stained in a free-floating fashion without major protocol changes from other species' immunostainings. Full brain immunostaining could also be envisioned by using, for example, CLARITY methods21.
Overall, this protocol, which covers personalized implant design to implantation, functionality follow-up, and biosafety assessment, is robust and consistent. We demonstrated here its feasibility to study the auditory system, but it can be transposed to test other physiological functions. Moreover, the strength of our method resides in the fact that it is not restricted to minipigs, but fully transposable to other species such as sheep, goats, or non-human primates. To a certain extent, it can also be easily adapted to rats.
The authors have nothing to disclose.
The authors would like to acknowledge financial support from the Bertarelli Foundation and the SNSF Sinergia grant CRSII5_183519. The authors would also like to thank Katia Galan of EPFL for her help on developing the staining protocol for the histology, the staff at the Neural Microsystems Platform of the Wyss Center for Bio and Neuroengineering in Geneva for their help with the fabrication processes, the staff of the of animal platform in the University Medical Center (CMU) at the University of Geneva (UNIGE) for animal care, surgical assistance, and postoperative management of the minipig (John Diaper, Xavier Belin, Fabienne Fontao, and Walid Habre), the team members of the Center for Biomedical Imaging (CIBM) at the University of Geneva (Julien Songeon, François Lazeyras, and Rares Salomir), the staff members of the Pathology Department at the University Hospital Geneva (HUG) (Sami Schranz, Francesca Versili, Ruben Soto, and Coraline Egger), and Blaise Yvert from Université Grenobles-Alpes for his input and exchanges on chronic minipig experiments. The authors would like to acknowledge the help of employees of Neurosoft Bioelectronics SA, for their help with the fabrication process and for their help in the minipig experiments (Benoit Huguet and Margaux Roulet).
Bone drill | BBraun | Elan 4 with GA861 handpiece | |
Bone drill bit | BBraun | Neurocutter GP204R | |
Bonewax | Ethicon | W31G | |
Catheter | Venisystems | Abbocath 14G | |
Confocal Microscope | Zeiss | LSM 880 | |
Cryostat | Leica | CM1950 | |
Gelfoam | Pfizer | Gelfoam | |
Insert speakers | Etymotic | Etymotic ER2 insert Earphones | |
Multimeter | Fluke | Fluke 1700 | |
Oscilloscope | Tektronix | MDO3014 Mixed Domain Oscilloscope | |
Perfusion pump | Shenzen | LabS3/UD15 | |
Potentiostat | Gamry Instruments | Reference 600 | |
Primary Antibody Anti-GFAP | Thermofischer | Anti-GFAP, Rat, # 13-0300 | |
Primary Antibody Anti-Iba1 | Fujifilm | Anti Iba1, Rabbit, 019-19741 | |
Primary Antibody Anti-NeuN | SigmaAldrich | Anti-NeuN, GuineaPig, ABN90 | |
Pulse Generator | AM Systems | Model 2100 Isolated Pulse Stimulator | |
Recording headstage | Multichannel systems | W2100-HS32 | |
Recording system | Multichannel systems | W2100 | |
Screwdriver | Medtronic | Handle: 001201, Shaft: 8001205 | |
Secondary Antibody 488 | Thermofischer | Goat anti-Rat IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor 488, # A-11006 | |
Secondary Antibody 555 | Thermofischer | Goat anti-Guinea Pig IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 555, # A-21435 | |
Secondary Antibody 647 | Thermofischer | Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 647, # A-21245 | |
Slide Scanner | Olympus | VS120 | |
Snapfrost | Excilone | Excilone Snapfrost | |
Stab knife | Fine Science Tools | 10316-14 | |
Suture wire dermal | Ethicon | Vicryl 2-0 | |
Suture wire dura mater | Ethicon | Mersilk 5-0 | |
Suture wire for catheter | Ethicon | Vycril 3-0 without needle | |
Suture wire for lifting dura | Ethicon | Prolene 6-0 with BV-1 needle | |
Suture wire subcutaneous | Ethicon | Vicryl 4-0 | |
Titanium bridge | Medtronic | TiMesh 015-2001-4 | Cut out the required size |
Titanium screws | Medtronic | 9001635, 9001640 | |
X-ray system | GE | GE OEC 9800 Plus C-Arm |