Unmet Need
Electrical stimulation with neural implants plays a crucial role in restoring sensory and motor function, as well as treating neurological disorders, including Parkinson’s disease[i], seizures, and even psychiatric disorders. Despite decades of research and development, neural implants only result in limited restoration of function in patients and optimal outcomes remain a challenge and patient recovery typically remains significantly below normal levels of function. System-specific explanations have been offered, ranging from unnatural recruitment of neurons (network-level adaptation[ii]), to local interference based on physiology.[iii] [iv]
To illustrate this issue, consider cochlear implants. Over 1 million people in the U.S. are considered deaf, with severe sensorineural hearing loss that do not benefit from hearing aids. Instead, cochlear implants are used to restore speech perception. However, cochlear implants cannot successfully restore pitch perception or music appreciation due to limitations in how sound is processed and converted into an electrical stimulus. Studies suggest that if auditory neurons were activated with fine timing closer to that of natural responses, pitch would be restored.
Technology Overview
Johns Hopkins (JHU) inventors have developed a fast algorithm for determining the optimal stimulation pattern to produce naturalistic responses in single neurons. This algorithm has two components. (1) It uses machine learning to efficiently determine desired local single neuron responses to inputs that would be naturalistic at real-time speeds. On a test set of auditory inputs, which is hidden during the training of the algorithm, it achieves a 4.4% error and runs 300 times faster than the state-of-the-art phenomenological model of sound processing in generating target cochlear responses to auditory input. This suggests that it is possible to integrate artificial intelligence into cochlear implants that can accurately predict normal physiological responses to sound and restore pitch perception, allowing patients with hearing loss to appreciate and listen to music again. (2) It includes a back-end that applies novel mathematical rules to pulsatile stimulation to account for disruptive effects of sequences of pulses on neural firing. The algorithm uses the range of viable pulse rate and pulse amplitude parameters and a measure spontaneous firing rate of neurons to predict the optimal pulse parameters to generate the target firing rate over time. In this case, the output of (1). The method can be applied to various types of neural implants, including cochlear, vestibular, retinal, deep brain, and spinal cord stimulators.
Stage of Development
Currently, the inventors have completed bench-top testing of the algorithm and expect to release the early version of the software on GitHub.
Patent
Provisional patent application – filed 2/18/21
PCT/US2022/17087 - filed 2/18/22
WO2022178316A1 - published 8/25/22
Publications
Steinhardt, C. R., & Fridman, G. Y. (2021, November). A Machine Learning-based Neural Implant Front End for Inducing Naturalistic Firing. In 2021 43rd Annual International Conference of the IEEE
Steinhardt, C. R., & Fridman, G. Y. (2020, July). Predicting response of spontaneously firing afferents to prosthetic pulsatile stimulation. In 2020 42nd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) (pp. 2929-2933). IEEE.Engineering in Medicine & Biology Society (EMBC) (pp. 5713-5718). IEEE.
Steinhardt, C. R., Mitchell, D. E., Cullen, K. E., & Fridman, G. Y. (2021). Pulsatile electrical stimulation creates predictable, correctable disruptions in neural firing. bioRxiv, 2021-08. (in revision Nature Communication)
________________________________________________
[i] Marquez-Chin, C. & Popovic, M. R. Functional electrical stimulation therapy for restoration of motor function after spinal cord injury and stroke: a review. Biomed. Eng. OnLine 19, 34 (2020).
[ii] Mitchell, D. E., Santina, C. C. D. & Cullen, K. E. Plasticity within non-cerebellar pathways rapidly shapes motor performance in vivo. Nat. Commun. 7, 11238 (2016).
[iii] Frijns, J. H. M., Schoonhoven, R. & Grote, J. J. The influence of stimulus intensity on spike timing and the compound action potential in the electrically stimulated cochlea: a model study. Proc. 18th Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. 1, 327–328 vol.1 (1996).
[iv] Kalkman, R. K., Briaire, J. J. & Frijns, J. H. M. Stimulation strategies and electrode design in computational models of the electrically stimulated cochlea: An overview of existing literature. Netw.: Comput. Neural Syst. 27, 107–134 (2016).
