2023.fall Soft Materials and Devices 202393003 Emma Julé-Bonnardel Development of implantable wireless electrodes using AI to restore hearing sensation in deaf patients. 1. Summary The project aims to address severe to profound deafness, affecting over 5% of the global population and projected to double by 2050. The current standard solution, cochlear implants (CI), faces limitations in sound perception crucial for speech, noise, and music. The proposal explores the integration of artificial intelligence (AI) to create a novel hearing aid system capable of encompassing a wider frequency range. This involves developing a chip for sound signal processing, wireless communication between the chip and biocompatible brain-implanted electrodes, and optimizing the overall system performance. The specific goals of the project are the following: 1) Creation of a chip using existing AI algorithms to capture and differentiate relevant sound signals in the environment, converting them into efficient electrical impulses for transmission. 2) Design of electrodes capable of communicating wirelessly with the chip, and biocompatible for safe brain implantation and effective stimulation of auditory nerves. 3) Improved overall system performance by optimizing the efficiency of communication between the chip and the electrodes to ensure device effectiveness. 2 Project detail 2.1 Rationale The phenomenon of diminished auditory acuity, referred to as deafness, denotes a decline in the capacity to perceive sound, quantified in decibels of hearing loss across four categories: mild (20 to 39 dB), moderate (40 to 69 dB), severe (70 to 89 dB), and profound (90 dB and above).[1] Alarming statistics indicate that over 5% of the global population necessitates rehabilitation for debilitating hearing impairment, a figure projected to double by 2050, reaching nearly 10% of the world's population.[1] This underscores the imperative to develop efficient devices facilitating either partial or complete restoration of hearing for individuals afflicted with deafness. The principal objective of this research proposal is to pioneer a system tailored for the treatment of the most severe forms of deafness. Currently, the most prevalent and effective solution is the CI, designed to circumvent compromised auditory structures and facilitate sound perception and interpretation.[2] Comprising an external component affixed to the head and an internally implanted segment housing electrodes within the cochlea, responsible for auditory perception, this device extracts and processes environmental sound signals using a microphone, transmitting the processed information to the internal implant via an antenna and external transceiver.[11] The internal implant then stimulates auditory nerve fibers with electrodes by rendering these stimulations as interpreted sounds in the brain. Despite the widespread adoption of CI, they exhibit substantial limitations, particularly in the impaired perception of sound frequencies crucial for speech comprehension, noise discernment, and music appreciation.[2] Recent research has predominantly concentrated on mitigating these shortcomings through advancements in implant technology and surgical techniques, such as optimizing electrode placement within the cochlea.[2] Concurrently, the evolving landscape of AI introduces novel perspectives for deafness treatment.[3,4] Existing algorithms demonstrate promising results in sound processing within the environment, reportedly surpassing the efficacy of CIs.[3,4] Given the paramount significance of hearing restoration for the deaf, the integration of AI in the development of a novel hearing aid emerges as a promising avenue. Collaborative efforts with researchers implementing functional algorithms can facilitate the creation of a chip that captures and discerns sound signals from the surroundings, converting pertinent signals into electrical impulses transmitted to cochlear electrodes. Current investigations into wireless neuro-engineering technologies indicate the feasibility of wireless communication with electrodes.[5] This prompts consideration for the development of electrodes communicating wirelessly with the chip, facilitating the transmission of electrical signals to stimulate auditory nerves in the chosen cochlea. Moreover, emerging evidence underscores the pivotal role of electrode placement in reconstructing the patient's auditory perception.[12] By leveraging algorithms to discern signals based on their frequency and strategically placing electrodes, optimal stimulation of nerves attuned to the original nature of the sound signal can be achieved. 2. Project Goals Following the rationale provided above, we propose to use an existing algorithm based on AI and to combine our knowledge acquired during our studies in electronics and materials to address the following project goals: • To develop a chip capable of picking up sound signals in the surrounding environment and sorting the important sound information to be transmitted and to convert it into electrical signals. • To develop biocompatible electrodes that can be implanted in the brain and communicate with the chip using wireless technology. • To optimize system performance and make communication between the chip and the electrodes as efficient as possible. 2.3 Experimental Approach The experimental approach will follow the logical order defined by the project goals. We organized the project into the following work-packages (WP). WP 1: Development of the chip capable of picking up sound signals in the environment and sorting the important sound information to be transmitted, converting it into electrical signals. WP1.1: Integration of the functional algorithm in a microprocessor To integrate the algorithm onto the chip, the initial step involves implementing it in a microprocessor slated for chip integration. Prudent selection of a microprocessor is essential, considering computational capacity, memory capabilities, and peripheral features, especially for algorithms reliant on deep learning methodologies, where a high-performance multi-core processor is ideal .[6] Given that microprocessor inputs and outputs are digital, considerations must be made during chip prototyping to employ devices converting picked-up sounds into digital signals. Similarly, for the output, converting the digital signal into an electrical one enables wireless transmission to the electrodes. WP1.2: Chip Design and Prototyping In our case the chip's design hinges on pivotal components: the microphone, microprocessor, and transmitter. A directionally chosen microphone ,[7] captures ambient sounds comprehensively, generating an analog electrical signal necessitating conversion to digital for microprocessor compatibility. The microprocessor undertakes vital roles—processing captured sounds, discerning relevant information, determining transmission specifics to the patient, and identifying the optimal electrode for sound retransmission. Processed signals are then transmitted to the electrodes via the radiofrequency (RF) transmitter, selected for its efficacy in existing CI technology,[8] This transmitter integrates a variable frequency oscillator (VFO) for frequency modulation, vital for tailored communication with each electrode. Beyond these components, the design incorporates a power supply block for comprehensive power provision. Component arrangement in the circuit is optimized for chip size reduction. Subsequent simulations validate the theoretical circuit functionality, leading to the creation of a prototype subjected to real-world tests. WP 2: Development of wireless biocompatible silicon-based with platinum-iridium film electrodes. WP2.1: Manufacture of highly conductive and biocompatible silicon-based with platinum-iridium (Pt-Ir) film electrodes The electrode architecture, rooted in silicone for its biocompatibility and electrical conductivity, is designed with an external segment housing the antenna and a lower segment in direct contact with cochlear auditory nerves. The lower segment substructure is coated with a strategically deposited Pt-Ir film to optimize electrode efficacy. The material choice, driven by its proficient electrical conductivity and biocompatible nature, aims to minimize the potential for an immune response.[9] The functional mechanism involves gold particle deposition within the silicon substrate via chemical vapor deposition (CVD),[13] establishing conductive pathways for transmitting electrical signals from the antenna to the Pt-Ir film. WP2.2: Development of antennae to pick up RF signals emitted by the chip Our primary objective is to allocate each electrode for processing a distinct acoustic signal. To achieve this, it is crucial to develop multiple antennas, with dimensions aligned to the wavelength of their designated frequency. Opting for the fabrication of these antennas, we apply a thin layer of gold, a conductive material, onto our silicone substrate using the Physical Vapor Deposition (PVD) technique.[14] Following this, a photosensitive resin is employed on the metal layer, exposed to ultraviolet light through a mask to precisely define the structure of each antenna.[15,16] After resin development, revealing the antennas, non-resin-protected metal is removed, leaving the antenna structure exclusively on the substrate. Varied mask configurations empower us to design antennas capable of capturing signals across different frequencies. WP 3: In-vivo test for real time monitoring of the efficiency of the communication between the cheap and the electrodes WP3.1: Analysis and optimization of the in vitro performance of communication between the chip and the electrodes To validate the comprehensive functionality of the entire system, our focus lies in assessing the efficacy of the electrodes post-reception of the RF signal emitted by the chip. This involves measuring the individual impedance of each electrode, a process executed by sequentially connecting each electrode to an oscilloscope. A test sound signal, corresponding to the reception frequency associated with each electrode, is utilized in this evaluation. This methodology ensures the verification of electric current generation, allowing for the measurement of voltage (V) and current (I), crucial parameters in determining impedance (Z = V/I) .[10] For an electrode designed to stimulate auditory nerves in the cochlea, a typical impedance may range between 5 and 20 kΩ. WP3.2: In-vivo test of the functioning of the communication between the optimised chip and electrodes Given the critical nature of this research and the neuroanatomical parallels shared between human and chimpanzee brains,[17] the chimpanzee stands out as an optimal candidate for in-vivo system testing. A qualified veterinarian will surgically implant electrodes into the auditory region of chimpanzees with severe to profound deafness. Audiometric assessments will evaluate auditory responsiveness before and after electrode implantation. Continuous monitoring will appraise electrode stability, discern potential side effects, and scrutinize auditory assessment data. The primary objective is to observe alterations in behavioral responses, identifying shifts in auditory habits and gauging the capacity to respond to auditory stimuli post-implantation. Following a meticulous analysis, comprehensive conclusions on the efficacy of the devised auditory implant device can be drawn. 2.4 Timetable and milestones MS1 Task Year 1 MS2 Year 2 WP1.1: Integration of the functional algorithm in a microprocessor WP1.2: Chip Design and Prototyping WP2.1: Manufacture of highly conductive and biocompatible silicon-based with platinum-iridium (Pt-Ir) film electrodes WP2.2: Development of antennae to pick up RF signals emitted by the chip WP3.1: Analysis and optimization of the in vitro performance of communication between the chip and the electrodes WP3.2: In-vivo test of the functioning of the communication between the optimised chip and electrodes MS1: The chip and the electrodes with the antennae have been developed and the two devices are functional independently. MS2: The entire system is tested in-vivo and in-vitro. 2.5 Expected output The development of a device utilizing AI to address severe to profound deafness is a groundbreaking and ambitious project that requires a synthesis of diverse disciplines. At its core, the innovation involves integrating an AI-based algorithm onto a microprocessor, establishing communication with implanted electrodes to stimulate auditory nerves. This strategic integration has the potential to eliminate the need for patients to use bulky external processors, addressing practical and aesthetic concerns. The successful realization of this project, supported by compelling empirical outcomes, could usher in a paradigm shift in auditory prosthetics research. 2.4 Impact and significance of the project The rise of hearing restoration devices for individuals with profound to severe hearing impairment holds great importance in healthcare. As the healthcare sector grapples with the growing number of individuals facing hearing loss, the need for innovative and effective solutions becomes more pronounced. This pressing issue emphasizes the critical importance of research and development initiatives to engineer advanced hearing restoration technologies. This effort not only tackles the immediate challenges faced by those currently affected but also anticipates the growing demand for sophisticated interventions in the years to come. 3. References [1] “Deafness and Hearing Loss.” N.p., n.d. Web. 14 Nov. 2023 . <https://www.who.int/news-room/factsheets/detail/deafness-and-hearing-loss>. [2] Fletcher, Mark D, Nour Thini, and Samuel W Perry. “Enhanced Pitch Discrimination for Cochlear Implant Users with a New Haptic Neuroprosthetic.” Scientific Reports 10.1 (2020): 10354. Web. [3] Lesica, Nicholas A. et al. “Harnessing the Power of Artificial Intelligence to Transform Hearing Healthcare and Research.” Nature Machine Intelligence 3.10 (2021): 840–849. Web. [4] Diehl, Peter Udo et al. “Restoring Speech Intelligibility for Hearing Aid Users with Deep Learning.” Scientific Reports 13.1 (2023): 2719. Web. [5] Won, Sang Min et al. “Wireless and Battery-Free Technologies for Neuroengineering.” Nature Biomedical Engineering (2021): n. pag. Web. [6] Reuther, Albert et al. “Survey and Benchmarking of Machine Learning Accelerators.” 2019 IEEE High Performance Extreme Computing Conference (HPEC). IEEE, 2019. 1–9. Web. [7] Sivonen, Ville et al. “The Efficacy of Microphone Directionality in Improving Speech Recognition in Noise for Three Commercial Cochlear-Implant Systems.” Cochlear implants international 21.3 (2020): 153–159. Web. [8] Wolfe, Jace et al. “Evaluation of Speech Recognition of Cochlear Implant Recipients Using a Personal Digital Adaptive Radio Frequency System.” Journal of the American Academy of Audiology 24.8 (2013): 714– 724. Web. [9] Petrossians, A, J J Whalen, and J D Weiland. “Improved Electrode Material for Deep Brain Stimulation.” Conference Proceedings: Annual International Conference of the IEEE Engineering in Medicine and Biology Society 2016 (2016): 1798–1801. Web. [10] Tykocinski, Michael, Lawrence T Cohen, and Robert S Cowan. “Measurement and Analysis of Access Resistance and Polarization Impedance in Cochlear Implant Recipients.” Otology & Neurotology 26.5 (2005): 948–956. Web. [11] “How Do Cochlear Hearing Implants Work? | Cochlear.” N.p., n.d. Web. 17 Nov. 2023 . <https://www.cochlear.com/us/en/home/diagnosis-and-treatment/how-cochlear-solutions-work/cochlearimplants/how-cochlear-implants-work>. [12] Durán-Alonso, María Beatriz. “Stem Cell-Based Approaches: Possible Route to Hearing Restoration?” World journal of stem cells 12.6 (2020): 422–437. Web. [13] Carlsson, Jan-Otto, and Peter M. Martin. “Chemical Vapor Deposition.” Handbook of Deposition Technologies for Films and Coatings. Elsevier, 2010. 314–363. Web. [14] Mattox, Donald M. “Physical Vapor Deposition (PVD) Processes.” Metal Finishing 100 (2002): 394–408. Web. [15] Jun, S et al. “Circular Polarised Antenna Fabricated with Low‐cost 3D and Inkjet Printing Equipment.” Electronics letters 53.6 (2017): 370–371. Web. [16] Rashidian, A et al. “Photoresist-Based Polymer Resonator Antennas: Lithography Fabrication, Strip-Fed Excitation, and Multimode Operation.” IEEE Antennas and Propagation Magazine 53.4 (2011): 16–27. Web. [17] Baizer, Joan S et al. “Species Differences in the Organization of the Ventral Cochlear Nucleus.” Anatomical Record 301.5 (2018): 862–886. Web.