Minimally-invasive insertion strategy and in vivo evaluation of multi-shank flexible intracortical probes

Chronically implanted neural probes are powerful tools to decode brain activity however, recording population and spiking activity over long periods remains a major challenge. Here, we designed and fabricated flexible intracortical Michigan-style arrays with a shank cross-section per electrode of 250 μm\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^2$$\end{document}2 utilizing the polymer paryleneC with the goal to improve the immune acceptance. As flexible neural probes are unable to penetrate the brain due to the low buckling force threshold, a tissue-friendly insertion system was developed by reducing the effective shank length. The insertion strategy enabled the implantation of the four, bare, flexible shanks up to 2 mm into the mouse brain without increasing the implantation footprint and therefore, minimizing the acute trauma. In acute recordings from the mouse somatosensory cortex and the olfactory bulb, we demonstrated that the flexible probes were able to simultaneously detect local field potentials as well as single and multi-unit activity. Additionally, the flexible arrays outperformed stiff probes with respect to yield of single unit activity. Following the successful in vivo validation, we further improved the microfabrication towards a double-metal-layer process, and were able to double the number of electrodes per shank by keeping the shank width resulting in a cross-section per electrode of 118 μm\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^2$$\end{document}2.

. Exemplary impedance spectra of PEDOT:PSS in comparison to bare Pt and Au microelectrodes with a GSA of 113 µm 2 .

Method S1: Microfabrication
The double-metal-layer fabrication of the second generation of flexible devices was started with the chemical vapour deposition of a 5 µm thick PaC film as substrate layer on a cleaned 4" Si wafer. In the first metallization step, the metal traces with a minimum line width and pitch of 5 µm were evaporated via electron beam evaporation and structured using a lift off process.
Due to the small dimensions of the interconnects, the photolithography process was changed from the double resist system (used for the first generation) to using only AZ nLOF2020. Furthermore, the exposure dose was adjusted to 17 mJ/cm 2 . A 500 nm thick PaC was deposited as insulation between the two metal layers. Openings with a diameter of 5 µm were introduced in the interlayer over the feedline ends via reactive ion etching. Recording sites and bond pads were defined in the second metallization step as described above. To ensure electrical connectivity between the two conductive layers, the thickness of the second metal stack was matched to the thickness of the PaC interlayer. Thus, the metal stack consisting of 10 nm titanium and 500 nm gold were evaporated. After the deposition of a 5 µm thick PaC film as passivation layer, the probe shapes were structured and the recording sites were exposed within the final etching step. Finally, the probes were dry released from the carrier wafer and soldered via flip-chip bonding to custom-made printed circuit boards.

Method S2: Animal surgeries
Barrel cortex: In preparation for the recordings, mice underwent two surgeries. One week prior to the experiments, a headholder was implanted onto the skull of the mice. For this, mice were anesthetized using inhaled isoflurane (5 % for induction, 1.5 -2.5 % during surgery). Mice were placed on an electrical heating blanket (TC200 Temperature controller, Thorlabs GmbH, Germany) for the duration of the surgery. For analgesia, mice were injected with buprenorphin (0.1 mg/kg, subcutaneous, Buprenovet, Bayer AG, Germany). Additionally, bupivacain (PUREN Pharma GmbH & Co. KG, Germany) was used as local anesthetic and was injected at the incision site. The hair above the skull was removed and the skin sterilized using iodine. A small incision was made along the midline. The skull underneath was cleaned, and the muscles gently pushed aside and fixed in place using vetbond. A small craniotomy was made above the cerebellum and a ground pin was implanted. A custom designed headbar was placed above the skull and both the headbar and ground pin were fixed in place using dental cement. Any remaining exposed skull area was covered in vetbond for protection. 9 mg/l buprenorphin and 1 ml/l baytril (Bayer AG, Germany) were added to the drinking water after surgery until the experiment. On the day prior to the experiment, the mice underwent a second surgery where a craniotomy was made above the barrel cortex (AP:-1, ML:-3, diameter:3 mm). Again, isoflurane was used for anesthesia and buprenorphin was injected for analgesia. Additionally, prednisolonacetat was injected (1.2 mg, intramuscular, CP-Pharma, Germany) to prevent brain swelling. For the recording, mice were lightly anesthetized with inhaled isoflurane (0.5 -1 %) and headfixed. Mice were placed on an electrical heating pad for the duration of the experiment. Prednisolonacetat was injected to prevent brain swelling. Slits were introduced to the dura that served as entry points for the probe shanks. Olfactory bulb: Extracellular recordings of olfactory bulb output activity were performed as previously described 9 . Briefly, mice were anesthetized by intraperitoneal injection of pentobarbital (50 mg/kg, Narcoren (16g/100 ml), Boehringer Ingelheim Vetmedica GmbH, Germany). Further doses of pentobarbital were administered via a shunt placed intraperitoneally into the mouse's abdominal region. Mice were placed on a heating blanket with circulating water (T/pump Professional, Stryker, USA) to keep core body temperature at about 37°C for the duration of the experiment. Bupivacain (PUREN Pharma GmbH & Co. KG, Germany) was used as local anesthetic and was injected at all incision sites. Mice underwent a double tracheotomy and an artificial sniff paradigm was used to reliably control air and odorant inhalation independent of respiration 10 . Following the double tracheotomy, mice were fixed in a custom-built stereotaxic frame. A dental drill was used to first thin and later remove the bone covering both olfactory bulbs. Finally, with the help of fine tweezers, the dura was also removed to permit probe penetration.

Method S3: Histology
The DiI solution was prepared by diluting the powder (D282, Invitrogen, USA) in ethanol (50 mg/ml) 11 . The flexible probes were labelled with DiI by dipping several times into the dye solution and allowing them to dry in air for a few seconds between dips. After successful insertion, the probes coated with DiI were left in the brain for roughly one hour for postmortem reconstruction of the penetration tracks. At the end of the acute recordings, the mice were overdosed with pentobarbital (Narcoren (16g/100 ml), Boehringer Ingelheim Vetmedica GmbH, Germany) and perfused with PBS and 4 % (v/v) paraformaldehyde (Carl Roth GmbH & Co. KG, Germany) in PBS. Perfused mouse heads were stored in paraformaldehyde in PBS at least for 24 hours before extracting the brain. After extraction, the brains were either stored in paraformaldehyde solution or in 30 % (w/v) sucrose solution. Sectioning was performed by embedding the brain in 3 % low-melting point agarose, and 100-200 µm thick coronal slices were prepared using a vibratome (752/M vibroslice; Campden Instruments Limited, UK). Dil fluorescence was evaluated using a confocal-laser scanning microscope (TCS SP2; Leica GmbH, Germany). Confocal evaluations of fluorescence were