(J-398) Fabrication Process Development Toward Coupling Mechanically-Adaptive Intracortical Electrodes with Microfluidic Channels

Intracortical microelectrodes record neural signals and are part of the brain-computer interfaces that aids in restoring motor, sensory and cognitive abilities in individuals who have suffered neurological impairments1. Due to the neuroinflammatory response and electrode degradation, intracortical microelectrodes often fail shortly after implantation in the brain2. Previous research indicates that electrodes fabricated from a mechanically-adaptive polymer-based nanocomposite – polyvinyl acetate with tunicate cellulose nanocrystals (PVAc/t-CNC) – induce less neuroinflammation compared to traditional silicon-based electrodes3. The mechanically-adaptive material undergoes transformation from a stiff to a compliant state in response to heat and hydration4. This allows for insertion into the motor cortex without buckling and less mechanical mismatch between the soft cortical tissue and the implant. In addition, the delivery of an antioxidant, resveratrol, has been shown to reduce neuronal death and the high density of inflammatory cells around the implant5. To prolong the release of the antioxidant to extend its effects, microfluidic channels were integrated into PVAc/t-CNC neural probes to continuously deliver therapeutic agents locally from the probe and into the cortical tissue. Previously, our lab has successfully created recording devices that release a fixed dose of antioxidant loaded into the PVAc/t-CNC, and integrated microfluidic channels into PVAc/t-CNC probes without recording microelectrodes. Here, we present a design and fabrication process for integrating both recording electrodes and microfluidic channels into a mechanically-adaptive polymer nanocomposite neural probe for the first time.

Microfluidic probes were fabricated by producing micromolded channel layers and laser-micromachined planar cover layers in parallel. PDMS molds were fabricated by curing a Sylgard 184 elastomer. For channel layers, the PDMS was cured over a SU-8 mold in a polytetrafluoroethylene (PTFE) dish, and for the capping layer, the PDMS was cured directly in a PTFE dish. For film fabrication, PVAc and CNCs were dispersed in dimethylformamide (DMF) separately, combined, and the solution was cast on the negative PDMS mold and the flat PDMS slab within PTFE dishes. The solutions were placed in an oven under vacuum at 70°C for 1 week to evaporate the solvent from the films. Before releasing from the PDMS, PVAc/t-CNC films were pressed at 90°C for 1 hour6.

The channel and capping layers were soaked in deionized  water for 24 hours and the layers were thermally bonded on a hotplate at 50°C for 15 minutes. Laser-micromachining was utilized to cut out individual probes. A 3D-printed microfluidic connector with a wax base was aligned and attached to each probe at 60°C for 90 seconds. Connectors facilitate PE-50 tubing connections between probe inlets and pumps7. The PE-50 tubing is secured to the connector using two-part epoxy. 

To test microfluidic channels, devices were submerged in DI water and PE-50 tubing was connected to a syringe pump that pushed air 100 microliters per minute through the tubing. The force on the syringe was measured using a load cell and a microscope camera monitored air flow through the device. 

The devices were tested using the flow test as described in the methods.  If the microfluidic channels had no leaks or blockages and air was observed to successfully pass through the channel outlet, the device was deemed to have passed the flow test. The force plot shown in Figure 1 displays a representative load cell data plot that visualizes the progression of load cell data, which reflects the pressure of air pushed through the tubing and microfluidic channels. As shown in Figure 1, the pressure increases as the force applied by the syringe pump increases. In successful microfluidic channels, the force on the syringe begins leveling off around 120 g. The point in the pressure recording indicates the point at which air begins to pass through the end of the microfluidic channel into the surrounding water. The air flowed out of the microfluidic channels in a rapid succession of small bubbles, as shown in Figure 2.

These results show that the process of pressing films, which was adapted from a process used for shorter microfluidic channels, allows for successful fabrication of long microfluidic channels. The long microfluidic channels are 8.053 mm, as shown in Figure 3, and twice the length of channels in previous microfluidic probes from our labs. By creating space for recording electrode connections, longer microfluidic channels will allow for the coupling of mechanically-adaptive recording electrodes and microfluidic channels. Notably, the force plots demonstrate that long microfluidic channels can withstand significant pressure without channel collapse or leakage. This suggests that long microfluidic channels can be successfully implanted into the cortex for in vivo testing. 

Future research will explore methods of replacing the capping layer of the probes with a mechanically-adaptive recording electrode layer. This substitution would allow for simultaneous recordings of neural signals and continuous release of an antioxidant. Combining two methods that aim to reduce cortical inflammatory responses would allow for longer lasting recording electrodes and improved brain health following implantation. Future research will also investigate methods to improve the fabrication yield  of mechanically-adaptive probes with long microfluidic channels.