Fiering et al.
S. Methods Supplement
S.1 Microfluidics
The individual components making up the fluid network are as follows (Fig 1): pump, reservoir, T-junction, cannula, fluid resistor, fluid capacitor, and connecting tubing. The pump was fitted with custom machined stainless steel reducing adapters to step down the existing 1/16th in. inlet and outlet tubes to connecting 28 gauge stainless steel hypodermic tubing. The adapters were epoxied with LCA-4 [Bacon Industries Inc., Watertown, MA]. The glass capillary reservoir had an ID of 1 mm.
A T-junction was fabricated from 4 layers of laminated adhesive-backed polyimide films [R/Flex 1000, Rogers Corp., Chandler, AZ]. Microchannels and tubing sockets were cut in the polyimide films with a Quickcircuit 7000 [T-Tech, Norcross GA] X-Y router table. The cannula and two sacrificial inlet and outlet tubes were inserted into their respective sockets in the polyimide sheets and the assembly was laminated under vacuum in a custom press according to standard procedures. The cannula, a 10-mm length of 34-gauge stainless steel hypodermic tubing was cut by wire electrical discharge machining. Alternately, in some experiments, a cannula was made from polyimide-sheathed fused silica having 75-mm ID and 150-mm OD (Polymicro, Phoenix AZ). The sacrificial tubes were 10-mm lengths of Teflon perfluoroalkoxyalkane (PFA) capillary tubing with an ID of 50 mm and an OD of 360 mm. Following lamination, the sacrificial Teflon tubing was removed and replaced with an 8.5-cm inlet tube and a 3-cm outlet tube, each of 28-gauge stainless steel. Epoxy (907, Miller Stephenson, Danbury CT) was applied to all three junctions where the tubing entered the laminated polyimide.
The fluid capacitor was laminated from 4 layers of milled polyimide sheets using methods similar those used for the T-junction. After lamination, the membrane outer surface was sputter coated with 30 nm Ti followed by 200 nm Au to reduce water vapor permeability. Also to reduce permeability, the opposite face of the capacitor was covered with a 250 mm thick glass coverslip, fixed in place with consumer-grade cyanoacrylate glue.
The pump was primed with deionized (DI) water according to the manufacturer’s recommendations. The T-junction and capacitor were also primed, and the three components connected in series using Tygon tubing (R3603), which was simply pressed over the stainless steel and PEEK tubing with an overlap of 3-5 mm. The tubing connecting the pump outlet to the T-junction was 19 cm long, to provide a small amount of compliance at the pump outlet. The other press fit connections, for example, the tubing connecting the T-junction to the capacitor was of minimal length (~8 mm). The final connection, from capacitor outlet to pump inlet, was made with epoxy after all components were filled. It is notable that the epoxy cured successfully at room temperature despite the water contents at the joint. The reason for priming and assembling the system in this sequence is that priming the pump with the resistor already attached was unsuccessful.
All stainless steel hypodermic tubing was purchased from Vita-Needle (Needham MA) and all PEEK and Teflon capillary tubing was purchased from Upchurch Scientific (Oak Harbor, WA).
S.2 Electronics
The lithium polymer rechargeable battery was purchased from Ultralife, Newark, NY. The microcontroller was a BS2 Basic Stamp® (Parallax Inc., Rocklin CA). All electronic components were soldered to a double-sided circuit board, which was fastened to the housing lid with nylon screws. The pump and battery were each connected to the circuit board with subminiature latching connectors (Omnetics, Minneapolis MN).
S.3 Housing and Mounting
The pedestal was machined from G-10 fiberglass composite. The heads of several (3-5) titanium #2 screws were trimmed to a thickness of 0.020 in. and a width of 0.086 in to form a bolt with a thin T-shaped head. Slots were bored in the skull and the bolts were inserted head-down between the skull and dura. The bolts were rotated 90 degrees about the shaft axis and secured with stainless steel nuts and washers. Scalloped features in the pedestal base mated to the bolt shafts as the pedestal was positioned on the skull. The entire assembly was additionally secured with dental cement. The guinea pig was allowed to heal for at least 6 days before proceeding.
The housing base socket attached to the tapered pedestal and was fastened with a nylon screw. The base and lid were designed in Solidworks (Concord MA) 3D computer aided design (CAD) software and exported to a Stratasys Titan (Eden Prairie MN) rapid prototyping system for fabrication in polycarbonate. To minimize mass of the housing, numerous cutouts were implemented in the 3D CAD model.
S.4 In-Vitro Testing
The fine bore tubing used to measure pulse volume was a 30-cm length of Tygon peristaltic tubing (ID 190 mm), which had been partially filled with water. The opposite end of the tube was open to atmosphere. The mass measurements were made with a Mettler Toledo (Columbus, OH) analytical balance. Individual components were filled with DI water and sealed. The water loss rates of the sealing fittings had been previously measured. The components were weighed at intervals for periods 3 to 10 days. For multi-day tests of the total system, the mass measurements were also made every few days. Between measurements the T-junction and cannula were immersed in an open reservoir of DI water while the pump continued to run at its programmed rate.
S.5 In-Vivo Testing – Acute
The composition of artificial perilymph was: 120 mM NaCl ; 3.5 mM KCl; 1.5 mM CaCl2; 5.5 mM glucose; 20 mM HEPES; titrated with NaOH to pH 7.5; total Na+=130 mM). The device was filled with AP by slowly displacing its original water content, taking care to not introduce air or exceed the pressure limits of the fluid capacitor membrane. In this procedure, the press-fit joint between the T-junction and the resistor was opened and a syringe pump infusing AP was attached to the resistor inlet. Simultaneously the device pump was turned on at a rate of 0.2 Hz. As fluid drained from both the cannula outlet and the opened connector, the infusion rate of the syringe pump was adjusted to match the solenoid pump rate. Flushing continued for at least 1hr and then the syringe pump was removed and the press fit joint was closed. The animal was anesthetized with a combination of pentobarbital sodium (Nembutal; 25 mg/kg i.m.), fentanyl (0.2 mg/kg i.m.), and droperidol (10 mg/kg i.m.) with half delivered as an initial bolus and the other half delivered over the next 1 to 1.5 hrs. Anesthetic boosters (1/4 the original dose) were administered as needed to maintain an adequate depth of anesthesia.