Project 3: Mechanics of Long-Shank 5 mm Neural Probe Insertion into the Rat Brain: Effects of Geometry and Vibration-Assisted Insertion
Motivation
Reaching deep subcortical targets in rodents — such as the dorsal striatum, hippocampus, and thalamus, which sit 3–5 mm below the cortical surface — requires probes significantly longer than those used in standard intracortical applications. But as shank length increases, critical buckling force drops sharply, making flexible, low-cross-section probes highly susceptible to mechanical failure during insertion. Reducing probe cross-sectional area is desirable for minimizing foreign body response and glial encapsulation, but this directly trades off with the structural stiffness needed for reliable implantation. For 5 mm probes in particular, no systematic characterization of how shank geometry and insertion technique interact to affect buckling risk, penetration force, and cortical dimpling had been established.
Our Solution
We designed and evaluated two 5 mm amorphous silicon carbide (a-SiC) probe geometries with identical length, thickness (~12 µm), and tip angle (11°): one with a uniform 175 µm shank width, and one with a tapered profile narrowing from 175 µm proximally to 75 µm at the distal tip. The tapered design reduces cross-sectional area at the point of entry — where penetration force and tissue interaction are greatest — while the wider proximal shank preserves bending stiffness during deeper insertion. Finite element modeling (FEM) in COMSOL was used to predict critical buckling force for both designs, including the tapered geometry where Euler's equation does not apply. We then validated FEM predictions experimentally and characterized insertion mechanics in 1.2% agarose phantom across four insertion speeds (20–1000 µm/s) before conducting acute in vivo insertions in rat cortex, also evaluating the effect of vibration-assisted insertion using a NeuralGlider inserter at 0.5 W actuation.
My Contributions
I conducted the experimental buckling force measurements for both probe designs, driving each probe against a rigid silicon wafer using the NeuralGlider inserter while recording force via a 20 g S-beam load cell. These measurements validated the FEM-predicted critical buckling forces (1.1 ± 0.07 mN for the uniform design, 0.7 ± 0.04 mN for the tapered), with no statistically significant difference between predicted and measured values for either design. I also prepared all 1.2% agarose gel phantoms and ran the full insertion speed characterization — testing both probe designs at 20, 100, 500, and 1000 µm/s — capturing force-displacement curves and extracting Fmax, penetration force (FP), and dimpling depth (Dd) at each condition. From this data, 100 µm/s was identified as the optimal insertion rate, balancing low Fmax, high success rate, and surgical feasibility. I then performed all in vivo probe insertions in rat brain, implanting both uniform and tapered probes at 100 µm/s to a depth of 4 mm under stereotaxic guidance, and repeated these insertions with vibration actuation enabled on the tapered design. Across all experimental conditions, I processed the full force-displacement curves to extract FP, Dd, and Fmax, which formed the basis of the primary in vivo findings: the tapered probe produced significantly lower FP (0.37 ± 0.17 vs. 0.55 ± 0.23 mN, p = 0.017) and a 100% insertion success rate compared to 55% for the uniform design, and vibration-assisted insertion further reduced FP and Dd, with 40% of actuated insertions showing no detectable penetration event.
Project Outcomes
Looking to discuss further? Contact me at research@mkmaharana.com
Project 2: Improved Intracortical Insertion Mechanics via Staggered Microelectrode Shank Design
Motivation
Multi-shank microelectrode arrays offer higher channel density and broader spatial coverage than single-shank designs — but they tend to fail faster. A leading reason is the mechanical trauma they cause during insertion: when four shanks enter the brain simultaneously, the cortex deforms significantly before penetration, straining tissue, disrupting vasculature, and triggering inflammatory responses that ultimately degrade the electrode-tissue interface. Reducing this insertion trauma without sacrificing the structural rigidity needed for unaided implantation is a core unsolved problem in neural interface engineering. Existing approaches — pneumatic inserters, vibration-assisted tools, surface coatings — all require additional surgical equipment or steps. A passive, fabrication-integrated solution would be far more practical.
Our Solution
Working with NeuroNexus, we developed and tested a custom four-shank MEA with staggered shank lengths (2.2, 2.4, 2.75, and 3 mm) and a non-uniform longitudinal thickness profile — 6 µm at the recording tip and 16 µm proximally. The staggered geometry means each shank enters the brain sequentially rather than simultaneously, distributing mechanical load across time. The stepped thickness preserves the bending stiffness needed to prevent buckling while keeping tip cross-sectional area small. Finite element modeling confirmed that the critical buckling force exceeded the measured cortical penetration force by a factor of six or more across all shanks. The design was validated in vivo in rat cortex against a conventional four-shank NeuroNexus array with equal 3 mm shanks, using insertion energy — the integral of the force-displacement curve — as a cumulative metric of total mechanical work imparted to the tissue.
My Contributions
I performed all in vivo MEA insertions and force-displacement data collection. Using a 20 g S-beam load cell mounted to a NeuralGlider inserter, I inserted both the staggered (STGRD_NNx) and conventional (CONV_NNx) MEAs into rat somatosensory cortex at 100 µm/s to a fixed depth of 2 mm across five trials per group, capturing complete force-displacement traces throughout each insertion. The data showed the staggered design reduced total insertion energy by 57% (0.78 ± 0.12 vs. 1.8 ± 0.7 µJ, p = 0.013) and peak cortical dimpling depth by 71% (190 ± 114 vs. 650 ± 183 µm, p = 0.0014), with approximately 50% of staggered shank insertions producing no detectable penetration force peak or dimpling.
Project Outcomes
Looking to discuss further? Contact me at research@mkmaharana.com
Project 1: Chronic Stability of Ruthenium Oxide Electrodes in Rat Motor Cortex
Motivation
For a neural implant to be clinically useful, it has to work not just on day one — it has to maintain reliable electrical contact with neurons for months or years. A central challenge is the electrode coating: the material that sits at the interface between the device and the brain, transducing ionic neural signals into electronic measurements. Iridium oxide coatings like SIROF are effective and well-characterized, but iridium is expensive and increasingly limited in supply. Ruthenium oxide is a lower-cost alternative with similar electrochemical properties in acute settings — but its long-term stability in the living brain hadn't been rigorously tested. Without that data, it couldn't be seriously considered for chronic implants. We needed to know whether RuOx could maintain stable impedance, charge storage, and recording quality over a clinically relevant timescale.
Our Solution
We implanted RuOx-coated amorphous silicon carbide (a-SiC) microelectrode arrays in rat motor cortex and tracked their performance weekly for 16 weeks. Each week we measured electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) in vivo to assess impedance at clinically relevant frequencies and charge storage capacity at two sweep rates, then immediately followed with neural recordings to quantify active electrode yield, signal amplitude, and signal-to-noise ratio. By pairing electrochemical and electrophysiological measures at every timepoint, we could distinguish true electrode degradation from recording variability — and track both independently across the full implantation period.
My Contributions
I led the data collection of this study. Performing all baseline in-vitro electrochemical characterization before implant, then executed weekly in-vivo EIS and CV sessions using a Gamry potentiostat across the full 16-week implantation period — tracking charge storage capacity at 50 mV/s and 50,000 mV/s sweep rates and 1 kHz impedance at every timepoint. I collected weekly neural recordings and quantified active electrode yield (AEY%), peak-to-peak voltage (Vpp), and SNR across all animals. Results confirmed stable electrochemical performance through week 16, with AEY sustained between 40–80% and stable Vpp (~70–170 µV, p=0.17 for trend).
