P30 - Modality Specific Signals are Required for Asymmetric Recording in Amputated Nerves Interfaced with Y-shaped Regenerative Microelectrode Arrays

Introduction: : Authors: *K. R. HUSSEIN1, S. COGAN2, M. I. ROMERO-ORTEGA1; Biomedical Engineering, 1Univ. of Arizona, Tucson, AZ; Bioengineering, 2Univ. of Texas at Dallas, Richardson, TX

Bionic integration of advanced limb prosthesis relies on peripheral interfaces implanted to the nerve stump such as extra neural cuff electrodes, indwelling tLIFE, TIME electrodes, or regenerative multielectrode arrays. We have previously used a Y-shaped regenerative interface – with specific growth factors in the two distal arms – to defasciculate a mixed transected nerve and guide its regeneration into multielectrode chambers for motor or sensory modality enrichment. We compared neural activity in capped and end-to-end RUMAs to define the nature of the motor efferent signals (capped group).

Materials and

Methods: : RUMA neural interface: A Y-shaped nerve conduit with 30 electrodes; 9-12 in the common, left and right arms of the Y-shaped interface.
Animals. Six adult Lewis rats (250-350g) were used for this study. All procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of the University of Arizona. Under gas anesthesia the sciatic nerve was exposed and transected. The proximal stump was sutured to the common arm of the Y-RUMA, and the distal arms were either capped with silicone or sutured one arm to the tibial (motor) and the other to the sural (sensory) fascicles. A 4.5-cm wire was placed subcutaneously and connected to a 32-channel Omnetics connector in a titanium pedestal anchored to the back muscles.
Treadmill Locomotion and Neural Recording. Two-weeks after implantation and weekly for 8 weeks, all animals were placed on a treadmill for 5 min and the neural activity recorded at 40kHz sampling rate, high-pass filtered (4-pole Bessel) at 300 Hz, and analyzed with Offline Sorter (Plexon) and NeuroExplorer. Neural data recording sessions were video-synchronized (30 frames/s) via Cineplex Behavioral Research System (Plexon, Inc.) with gait cycle data (heel and toe trajectories, swing, stance phases) measured via Kinovea.

Results, Conclusions, and Discussions:: Baseline neural activity in the capped RUMA common arm was qualitatively double that observed in the distal arms in agreement with the expected equal defasciculation. We observed burst neural activity correlated with toe and heel positions during gait, similarly patterned, but not identical, in the three RUMA arms. This indicates equal axon content in each arm and efferent activity as the nerves at the distal end are isolated from the limb muscles. Power spectral densities show a 10-fold decrease in cRUMA distal arms compared to common arm. Histological analysis is being conducted to confirm the asymmetry in amputated nerves.
As expected in the capped configuration, little to no signal differentiation or identifiable activity patterns were initially observed, whereas our end-to-end implanted rat presented with several spiking patterns distinct for Arm 1 compared to Arm 2, indicating potential separation of motor and sensory modalities. This highlights the need of molecular guidance to separate functional modalities, which may improve natural motor control and sensation of advanced bionic limbs.

Acknowledgements and/or References (Optional):: We thank Professor Stuart Cogan (Univ. Texas, Dallas) for array fabrication of UMEAs, Darianne Sanchez and Hugo Libreros for assistance in data collection and funding by NIH NINDS 1R01NS124222-01 (MI).