Skip to Main Content
 

Global Search Box

 
 
 
 

ETD Abstract Container

Abstract Header

Chronic Peripheral Nerve Recordings and Motor Recovery with the FINE

Eggers, Thomas Elliott

Abstract Details

2018, Doctor of Philosophy, Case Western Reserve University, Biomedical Engineering.
Objective Upper limb amputations severely inhibit an individual’s ability to care for themselves and hold down employment, as well as cause social isolation. Commercially available techniques to replace the lost functionality fall short of restoring natural movement to artificial limbs. These available methods rely on recording muscle activity from the residual limb, and thus the functional restoration is directly related to the level of amputation; high level (above elbow) amputations provide the fewest number of recording sites to restore the greatest number of movements. Even among transradial amputees where functional restoration is high, acceptance rates of these myoelectric devices are low, sometimes below 50%. Peripheral nerve interfaces provide an avenue to tap into the rich neural information the brain sends to the amputated limb, theoretically containing all the necessary information to fully replicate the intended motion. The challenge lies in chronically interfacing with these fragile structures and interpreting the neural activity to allow such signals to drive a prosthetic limb. In this work chronic peripheral nerve recordings are obtained in a freely walking animal model, and a novel signal processing algorithm demonstrates the ability to predict the dynamic motor activations to the intended muscles using the flat interface nerve electrode (FINE). Methods A novel signal processing algorithm, Hybrid Bayesian Signal Extraction (HBSE), which combines several previously developed approaches is first derived and presented. A benchtop setup which allows a user to precisely place artificial sources within the cuff is shown. This setup allows us to generate realistic neural signals to test the ability of different algorithms to recover the location and original source activity. For the chronic preparation, canines were implanted with 16-channel multicontact FINE on the sciatic nerve. Data presented comes from two animals (one unilateral and one bilateral). Muscle electrodes were also implanted in the target muscles, the gastrocnemius (GN) and tibialis anterior (TA), as well as possible interfering muscles in the hamstrings. Recording trials consisted of having the animals walk on a treadmill at a moderate pace for 1-2 minutes. Results Benchtop testing demonstrates improved recovery abilities of the proposed algorithm, HBSE, as measured by the signal to noise ratio (SNR), correlation coefficient (CC) and signal to interference ratio (SIR). HBSE is further shown to reject external noise, such as contamination from nearby muscles. Chronic recordings were achieved over the implant durations. These recordings are demonstrated to be free of interfering muscle activity from the nearby hamstring muscles by comparing the simultaneously recorded neural and interfering EMG. Independent component analysis (ICA) validates the presence of two independent signals in the neural recordings which correlate with the target muscle activities. The HBSE algorithm reliably recovers these two activities in all three legs, with correlation coefficients of 0.84±0.07 and 0.61±0.12. This analysis is extended to calculate the information recovered from the neural signals to the muscle activations, demonstrating 4 and 1 bit per second (bps) of transmitted information. Significance These results demonstrate that signals which correlate to the intended muscle function can be recovered from peripheral nerves using extraneural electrodes. The main limitation of this work lies in the investigated anatomy; canine sciatic nerves contain two large fascicles, while human upper limb nerves contain 20+. Nevertheless, this work demonstrates a chronic peripheral nerve interface which can predict the activity of downstream muscles could in turn be used to control a prosthetic limb.
Dominique Durand (Advisor)
Gustafson Kenneth (Committee Chair)
Tyler Dustin (Committee Member)
Triolo Ron (Committee Member)
Mohseni Pedram (Committee Member)
123 p.

Recommended Citations

Citations

  • Eggers, T. E. (2018). Chronic Peripheral Nerve Recordings and Motor Recovery with the FINE [Doctoral dissertation, Case Western Reserve University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=case1516624330839376

    APA Style (7th edition)

  • Eggers, Thomas. Chronic Peripheral Nerve Recordings and Motor Recovery with the FINE. 2018. Case Western Reserve University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=case1516624330839376.

    MLA Style (8th edition)

  • Eggers, Thomas. "Chronic Peripheral Nerve Recordings and Motor Recovery with the FINE." Doctoral dissertation, Case Western Reserve University, 2018. http://rave.ohiolink.edu/etdc/view?acc_num=case1516624330839376

    Chicago Manual of Style (17th edition)