Home
Search
Collections
Journals
About
Contact us
My IOPscience
High sensitivity recording of afferent nerve activity using ultra-compliant microchannel electrodes: an acute in vivo validation
This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2012 J. Neural Eng. 9 026005 (http://iopscience.iop.org/1741-2552/9/2/026005) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 131.111.95.11 The article was downloaded on 26/07/2012 at 19:00
Please note that terms and conditions apply.
IOP PUBLISHING
JOURNAL OF NEURAL ENGINEERING
doi:10.1088/1741-2560/9/2/026005
J. Neural Eng. 9 (2012) 026005 (7pp)
High sensitivity recording of afferent nerve activity using ultra-compliant microchannel electrodes: an acute in vivo validation Ivan R Minev1,4 , Daniel J Chew2 , Evangelos Delivopoulos1 , James W Fawcett2 and St´ephanie P Lacour3 1 2 3
Nanoscience Centre, University of Cambridge, Cambridge CB3 0FF, UK Cambridge Centre for Brain Repair, University of Cambridge, Cambridge CB2 0PY, UK ´ Laboratory for Soft Bioelectronic Interfaces, Ecole Polytechnique F´ed´erale de Lausanne, Switzerland
E-mail:
[email protected]
Received 8 August 2011 Accepted for publication 1 December 2011 Published 13 February 2012 Online at stacks.iop.org/JNE/9/026005 Abstract Neuroprostheses interfaced with transected peripheral nerves are technological routes to control robotic limbs as well as convey sensory feedback to patients suffering from traumatic neural injuries or degenerative diseases. To maximize the wealth of data obtained in recordings, interfacing devices are required to have intrafascicular resolution and provide high signal-to-noise ratio (SNR) recordings. In this paper, we focus on a possible building block of a three-dimensional regenerative implant: a polydimethylsiloxane (PDMS) microchannel electrode capable of highly sensitive recordings in vivo. The PDMS ‘micro-cuff’ consists of a 3.5 mm long (100 μm × 70 μm cross section) microfluidic channel equipped with five evaporated Ti/Au/Ti electrodes of sub-100 nm thickness. Individual electrodes have average impedance of 640 ± 30 k with a phase angle of −58 ± 1 degrees at 1 kHz and survive demanding mechanical handling such as twisting and bending. In proof-of-principle acute implantation experiments in rats, surgically teased afferent nerve strands from the L5 dorsal root were threaded through the microchannel. Tactile stimulation of the skin was reliably monitored with the three inner electrodes in the device, simultaneously recording signal amplitudes of up to 50 μV under saline immersion. The overall SNR was approximately 4. A small but consistent time lag between the signals arriving at the three electrodes was observed and yields a fibre conduction velocity of 30 m s−1 . The fidelity of the recordings was verified by placing the same nerve strand in oil and recording activity with hook electrodes. Our results show that PDMS microchannel electrodes open a promising technological path to 3D regenerative interfaces. S Online supplementary data available from stacks.iop.org/JNE/9/026005/mmedia (Some figures in this article are in colour only in the electronic version)
1. Introduction
4
Advanced neuroprosthetic systems offer a technological path to treat and assist patients with impaired motor and/or sensory neural functions. They involve electrodes chronically
Author to whom any correspondence should be addressed.
1741-2560/12/026005+07$33.00
1
© 2012 IOP Publishing Ltd Printed in the UK
J. Neural Eng. 9 (2012) 026005
I R Minev et al
confining bundles of axons. They artificially increase the resistivity of the surrounding medium. A tubular array interface may host tens of parallel microchannels equipped with electrodes. In vitro micro-cuffs are usually produced as hybrid platforms where a polymeric overlay with micro-moulded trenches is aligned and bonded onto a planar microelectrode array (MEA) on glass or silicon [7–9] Dissociated neurons seeded on dedicated areas of the lab-onchip are allowed to grow through the microchannel network. Extracellular potentials of associated action potentials are recorded within the confined microstructures with amplitude in the tens of μV range and SNR up to 4. These microfluidic platforms find applications in models studying neuronal network communication, injury and neurodegeneration, as well as in drug delivery [10]. In vivo microchannel electrode arrays are challenging to produce given their 3D design and high aspect ratio. In this paper, we propose to use a long and narrow polydimethylsiloxane (PDMS) microchannel equipped with axially distributed thin gold film electrodes as a building block for an in vivo regenerative microchannel interface. The proposed device benefits from PDMS compatibility with standard microfabrication techniques [10–13], microchannel dimensions optimized to reliably detect extracellular potentials [14] and enhanced mechanical compliance of PDMS-based electrodes [15]. To validate our approach, the PDMS microcuff electrodes were evaluated in vivo. In a typical experiment, strands of dorsal root nerves were surgically teased and inserted into the micro-cuff while still remaining attached to their afferent inputs. Nerve activity was evoked by tactile stimulation of the skin of the rat under anaesthesia and recordings with high signal-to-noise ratio (SNR) were obtained from the embedded microchannel electrodes under saline immersion. An array of microchannel electrodes may then combine in one device the regenerative capabilities and sub-fascicular resolution of sieve electrodes with noise rejection and elevated SNRs of cuff interfaces. The regenerative implant may not only find applications for controlling robotic limbs but also for detecting weak and slowly propagating afferent signals such as those from the bladder. Monitoring bladder fullness would then be integrated with the effector branch of a closed loop prosthetic control system.
interfaced with the central or peripheral nervous systems. Chronic stimulation implants such as deep brain stimulators, cochlear and retinal implants are currently used in the clinical context. They operate by injecting current pulses to stimulate large populations of neurons in the vicinity of the electrode, thus transmitting information to the brain stem and cortex [1]. The long-term capability of these devices to stimulate the nervous system is not greatly affected by the inflammation and scarring that surrounds them as this can be overcome by gradually increasing stimulation currents. Recording from the nervous system, however, poses more challenges and no devices are in routine clinical use. Reliable recording of single or compound action potentials is the basis for deciphering movement intention from the nervous system. Recordings can be used for control of external devices (a prosthetic limb or a muscle stimulator) and are key for understanding how sensory functions are encoded. In vivo recording interfaces rely on extracellular electrodes. Electrodes sit in low resistivity extracellular space and have to detect travelling action potentials typically 10 to 200 μV in amplitude and with frequency content in the 100 Hz to 10 kHz range. In myelinated fibres, extracellular currents are concentrated at nodes of Ranvier (which can be up to a couple of millimetres apart depending on the axon diameter and fibre type), thus the electrode needs to be close to these nodes to record an optimal signal. Furthermore, recordings are subject to interference from physiological noise and stimulus artefacts. Several strategies to interface the nerves with recording electrodes are under development. Slanted Utah electrode arrays or micro-wire arrays penetrate the peripheral nerve and thus establish a close electrical coupling with axons inside [2, 3]. They are however extremely stiff, can easily be displaced by tethering forces and do not accommodate micro-movement from nerves and surrounding tissues. Cuff electrodes are a less invasive alternative. A silicone tube bearing electrodes on its inside surface is wrapped around a whole nerve. Cuffs are a convenient platform for utilizing tripolar electrode configurations for noise suppression. Cuff interfaces with enhanced selectivity such as FINEs (flat interface nerve electrodes) are currently in clinical trials [4]. Flexible thin-film LIFEs (longitudinal intrafascicular electrodes) or TIMEs (transverse intrafascicular multichannel electrodes) have recently been developed for invasive, multi-contact, sub-fascicular peripheral nerve recordings. A test on a human volunteer illustrates promising control and feedback data but the level of selectivity remains relatively low [5]. Polymeric regenerative electrodes are designed to interface a higher number of nerve fibres and can bridge the gap across severed nerve stumps. They are a promising route to high sensitivity, high resolution chronic interfaces. Thin sieve electrodes (usually inserted in a hollow silicone tube) support relatively successful in vivo axon ingrowth and subsequent neural interfacing. They rely on a close proximity of the ring electrode to the axons and the corresponding nodes of Ranvier, therefore not all electrodes record equally well and are subject to cross-talk [6]. Regenerative microchannel electrodes combine features of cuff and sieve electrodes. They are long tubular devices
2. Materials and methods 2.1. Device fabrication The micro-cuff device consisted of a top and a bottom PDMS layer. Five thin gold film electrodes spaced at a 0.6 mm pitch were patterned onto the bottom PDMS layer using electronbeam evaporation and a flexible shadow mask (figure 1(a)). Each electrode was 140 μm wide, 13 mm long and formed as a titanium/gold/titanium (Ti/Au/Ti) multilayer of 5/40/3 nm thickness, respectively. The titanium films ensured the adhesion of the gold electrode to the top and bottom PDMS layers. The top PDMS layer as illustrated in figure 1(b) served as an electrically insulating cap for the electrodes underneath. 2
J. Neural Eng. 9 (2012) 026005
I R Minev et al
(d)
(a)
(b)
(e)
(c)
Figure 1. Device fabrication. (a) The bottom PDMS layer carries five evaporated Ti/Au/Ti electrodes. (b) The top PDMS layer is aligned and plasma bonded to the electrode array to form a microchannel. (c) Wire connections are established and the contact pads flooded with silicone. (d) Optical micrograph of the channel entrance. (e) Assembled microchannel device subjected to a torsion test.
The top layer was equipped with a trench opening at the surface of the film. When the top and bottom layers were brought together, the trench formed a channel through which a nerve strand was later threaded. The top layer was 100 μm thick and the channel had a rectangular cross section of 70 μm by 100 μm. It was fabricated by spin casting PDMS on a wafer containing a raised ridge of the desired channel dimensions. The top and bottom layers were permanently bonded by exposing both surfaces to brief air plasma then manually aligning the two surfaces into contact so that the channel was perpendicular to the electrode tracks. Any film overhang was trimmed. The tip of the implant was 13 mm long, 3.5 mm wide and 0.6 mm thick. Connections to off-board electronics were made with Teflon-coated wire, which was attached to the connecting pads of the electrodes through a small globule of conductive paste. The connection pads were then flooded with silicone rubber and the complete recording device left to set overnight (figure 1(c)). The optical micrographs in figures 1(d) and (e) show the channel entrance and a device subjected to a torsion test, respectively. Several devices were made and used for characterization and in vivo testing.
to the working electrode terminal of the impedance analyser. A platinum mesh and a Ag/AgCl probe were dipped in the electrolyte around the device and were connected as counter and reference electrodes, respectively. Frequencies between 100 Hz and 100 kHz were scanned with stimulation voltage amplitude of 5 mV. 2.3. Surgery and electrophysiology All animal experiments were performed in accordance with the Animals (Scientific Procedures) Act 1986. Adult SpragueDawley rats (300 g) were anaesthetized with an intraperitoneal injection of 1.5 g kg−1 urethane (Sigma Aldrich, UK). A laminectomy to three of the lumbar vertebrae was performed to expose the dorsal side of the spinal cord. The skin was sutured to an overhanging bar, to create a contained pool. The dura was transected to expose the dorsal roots and the pool flooded with PBS. The left dorsal rootlets of the lumbar 5 (L5) root were cut approximately 3 mm caudal from the dorsal root entry zone. The cut end of the rootlet was free to move but the whole strand remained attached to its afferent inputs and dorsal root ganglion. The tip of the free end of the nerve root was held with fine forceps (Fine Science Tools, UK) and carefully teased caudally into two smaller strands along its length. The splitting was repeated until a strand (less than 100 μm in diameter) small enough to fit inside the channel was obtained. Next a nylon surgical suture (Ethilon No 9.0, Ethicon) was threaded through the PDMS microchannel. The end of the suture was tied to the tip of the free end of one of the rootlet strands with a single suture knot. The suture was then retracted rostrally, guiding the rootlet strand into the channel as illustrated in figures 2(a) and (b). Stimulation of the afferents
2.2. Impedance spectroscopy The electrical properties of electrodes in the channel were evaluated using impedance spectroscopy. In order to fill the microchannels with electrolyte, devices were immersed in phosphate buffered saline (PBS) and subjected to a gentle vacuum in a desiccator. Impedance spectra were obtained with an Autolab μIII impedance analyser in potentiostatic mode at room temperature. Each microelectrode in an array was tested separately at the open circuit potential by connecting it 3
J. Neural Eng. 9 (2012) 026005 (a)
I R Minev et al (b)
(c)
Figure 2. Implantation in the rat spinal cord. (a) Schematic of the recording device interfaced with root L5. (b) Optical micrograph of the device in place during a mock-recording. Please note that for the purpose of taking this image, the saline has been aspirated. This prevents reflections from the liquid surface. Neural signal recording is however conducted under complete saline immersion of the micro-cuff. (c) Schematic of the connectivity of electrodes within the device.
of the L5 dorsal root was achieved by rhythmically stroking the corresponding dermatomes with a plastic rod. For the L5 root this was the ipsilateral (left) hind paw. As a control, noninnervated areas such as the contralateral (right) leg or the tail were stroked. Simultaneous recordings from the inner three electrodes were obtained by connecting each of them to one input of a dedicated amplifier. The other input of each amplifier received signal from the two outer electrodes which were shorted. This formed three recording tripoles positioned inside the channel as illustrated in figure 2(c). The experiment was conducted inside a Faraday cage to reduce external electrical interference. The animal and apparatus were connected to ground. Signals were amplified using a Neurolog electrophysiology system (Digitimer, UK). For each tripole, the signal was pre-amplified 1000 times with an ac-coupled amplifier (Neurolog NL104A) fed through a Humbug 50 Hz noise eliminator (Quest Scientific, Vancouver, Canada) and ten times further amplified (Neurolog NL106) before filtering out frequencies below 200 Hz and above 5 kHz. The amplified signal was sampled at 50 kHz and digitized with an analogueto-digital converter (Micro 1401 Mk II, Cambridge Electronic Design, UK). Control recordings with hook electrodes were performed in differential mode with the same amplification, filtering and digitization settings as above. The rootlet strand was placed on the hooks and the saline in the spinal canal was replaced with paraffin oil to provide an electrical seal.
Figure 3. Electrical characterization of electrodes inside microchannels. The traces are averaged Bode plots for the impedance modulus and phase angle of 15 electrodes across three devices. The error bars represent standard error. The inset is a scanning electron micrograph of the surface of an electrode following stretch to 20% tensile strain and relaxation.
reproducible impedances within a device and across devices. This observation indicates that the top PDMS passivation layer offers an adequate electrical seal. The impedance spectra suggest largely capacitive behaviour for Ti/Au/Ti electrodes in the 100 Hz–10 kHz frequency range as evidenced from phase angles
