Carbon nano-structured neural probes show promise for magnetic resonance imaging applications

Abstract

Objective Previous animal studies have demonstrated that carbon nanotube (CNT) electrodes provide several advantages of preferential cell growth and better signal-to-noise ratio when interfacing with brain neural tissue. This work explores another advantage of CNT electrodes, namely their MRI compatibility. MRI-compatible neural electrodes that do not produce image artifacts will allow simultaneous co-located functional MRI and neural signal recordings, which will help improve our understanding of the brain. Approach Prototype CNT electrodes on polyimide substrates are fabricated and tested in vitro and in vivo in rat brain at 9.4T. To understand the results of the in vitro and in vivo studies, a simulation model based on numerical computation of the magnetic field around a two-dimensional object in a tissue substrate is developed. Main Results The prototype electrodes are found to introduce negligible image artifacts in structural and functional imaging sequences in vitro and in vivo. Simulation results confirm that CNT prototype electrodes produce less magnetic field distortion than traditional metallic electrodes due to a combination of both superior material properties and geometry. By using CNT films, image artifacts can be nearly eliminated at magnetic fields of strength up to 9.4T. At the same time, the high surface area of a CNT film provides high charge transfer and enables neural local field potential (LFP) recordings with an equal or better signal-to-noise ratio (SNR) than traditional electrodes. Significance CNT film electrodes can be used for simultaneous MRI and electrophysiology in animal models to investigate fundamental neuroscience questions and clinically relevant topics such as epilepsy.

Computation of Magnetic Field Distortions and Impact on T 2 *-weighted MRI, with Applications to Magnetic Susceptibility Parameter Estimation

“…For the silver wire, we would expect the minimum A% value to be −15 ppm, and for the tungsten wire we expect to find the minimum Δχ value as 86 ppm. For CNT, the expected value of Δχ is between −1 and −21 ppm when bulk-scale density of CNT is used to convert mass magnetic susceptibility reported in literature to volume magnetic susceptibility [22]. These estimates are quite close considering the simplicity of implementing this approach.…”

“…Experimental in vitro gradient echo imaging was performed on a silver wire with diameter 254 μm, a tungsten wire with diameter 127 μm, and a prototype CNT electrode [22]. We then utilized the scalability of Φ with respect to χ for estimating the magnetic scalar potential value of these three samples.…”

“…From T 2 * , we can generate a map that estimates the signal strength in each voxel using equation (22), where S is the signal intensity. While this simple approach only accounts for the influence of T 2 * and the echo time (T E ) on the magnitude map, reasonable correspondence to the experimental data can be obtained without having to estimate longitudinal relaxation time (T 1 ), proton density, etc throughout the sample, assuming that the MRI signal is acquired under fully relaxed condition.…”

“…Three samples were prepared using: a silver wire with outer diameter of 254 μm, a tungsten wire with outer diameter of 127 μm, and a prototype CNT neural electrode 282 μm wide with approximately 5 μm conductive functionalized CNT on 76 μm polyimide and insulated by polydimethylsiloxane estimated to be of similar thickness to the polyimide layer. The CNT electrode is described in detail elsewhere [22]. The samples were held in plastic tubes such that the cross sections of interest would be oriented perpendicular to the static magnetic field in the MRI experiment.…”

“…T 2 was estimated for water as 40 ms at 9.4 T [23] and set to 500 μs for polyimide and 10 ms for polydimethylsiloxane. We then computed the signal strength map from the map of T 2 * using equation (22).…”