Symposium F.FL02—Advanced Neural Interfacing Materials, Devices and Microsystems

Implantable electrode arrays have been widely used to treat neurological disorders such as Parkinson’s disease, epilepsy, hearing loss, and chronic pain by applying electrical stimulation and/or recording neural activity. One of the main requirements for these devices is the ability to evoke minimal inflammatory tissue response to guarantee the long-term use of integrated functions. However, this can be limited by a mechanical mismatch between the implant, very stiff, and the tissue, very soft, inducing neural damages and leading to glial scar formation. So far, the most common materials used for fabrication of such neural interfaces are PI and PDMS. On one hand, thin-film microfabrication techniques allow the fabrication of flexible miniaturized and highly integrated stimulating or recording arrays in PI. However, the Young’s modulus of PI is about a few GPa. On the other hand, PDMS because of its low Young’s modulus (approximately 1 MPa) and its elastic behaviour leads to less damage to the neural tissue, but it presents challenging steps in the microfabrication process, especially for the encapsulation of the active sites. Off-Stoichiometric-Thiol-ene Epoxy polymers (OSTE+) were developed firstly for rapid prototyping of microfluidic chips because they are less permeable to gas. OSTE+ materials are a good compromise between PI and PDMS. In fact, by adjusting the stoichiometry of the constituting monomers, the Young’s modulus of OSTE+ can be decreased down to a few MPa at physiological temperature. The preparation of OSTE+ consists of two curing steps in which (1) thiol-ene photopolymerization allows to pattern the polymer surface and remove the non-polymerized areas before (2) thermal curing. Thus, it is possible to encapsulate the electrode array simply using photolithographic techniques. Taking advantage of these properties, we fabricated OSTE+ implantable multielectrode arrays with Pt serpentines in monolayer and bi-layer. The functional characteristics of these arrays have been tested by electrochemical characterization and recordings.

Retinal prostheses require high-density electrode arrays to attain high spatial resolution in vision. However, the transmission lines panning out in the same layer with electrodes may hamper achieving high-density electrodes. Double-faced electrode arrays may circumvent this problem, in which the electrodes and bonding pads are arranged face-to-face and the transmission lines are absent. The electrodes can be connected with stimulation drivers such as photovoltaic cells directly through bonding pads. Here we present the fabrication of a flexible and transparent parylene-based double-faced electrode array and the method of integrating it with the stimulation driver IC.

In recent years, neural interface technologies have been focused on flexible and transparent neural electrodes for high spatiotemporal resolution mapping of neuronal activities using electro-optical measurements. There had been several new material suggestions for the transparent, flexible neural electrodes such as ITO, graphene, conductive polymer, or patterned metal (nanowires, mesh patterns).[1, 2] However, these materials have shown several limitations such as high processing temperature, limited mechanical/electrical reliability, fabrication complexity and incompatibility with conventional microfabrication processes etc. To resolve these problems, in this work, we developed a flexible microelectrode array (MEA) using polymer-metal hybrid structures for high optical transparency and good electrical conductivity.
We used layer-by-layer assembly of polyelectrolytes as a nucleation inducing coating layer on transparent substrates (glass, polymer) to realize highly conductive vacuum deposited gold film at sub 10 nm thickness. We optimized the electrical and optical properties of the ultrathin gold electrodes, showing sheet resistance as low as 7.24 Ω/sq at 8 nm thickness and electrical conductivity of 1.327 × 10⁶ S/m. Despite the opaque characteristic of typical thick gold film, the ultrathin film of gold in this work still showed optical transparency of 77% at 550 nm, which is comparable to or better than most of transparent neural electrode materials. In addition, excellent mechanical stability is also demonstrated with bending radius down to 1 mm for at least 500 times of bending. We further measured the baseline noise properties of our ultrathin 40-µm diameter gold MEA for in vitro neural recording setup. We have confirmed that it shows low RMS noise level of 9.62 ± 2.24 µV_rms (n=60, bandpass filtered in 200 Hz – 3.5 kHz). These results confirm that the polymer-metal hybrid structure based MEA we developed in this work can be used for simultaneous extracellular electrical neural recording and optical neural recording such as calcium imaging.

References
1. Jimbo, Y. et al. Ultraflexible Transparent Oxide/Metal/Oxide Stack Electrode with Low Sheet Resistance for Electrophysiological Measurements. ACS Appl. Mater. Interfaces 9, 34744–34750 (2017).
2. Kshirsagar, P. et al. Transparent Graphene/PEDOT:PSS Microelectrodes for Electro- and Optophysiology. Adv. Mater. Technol. 4, 1–7 (2019)

Neural signal recording and direct stimulation of the brain have been widely used to mitigate neurological disorders, such as epilepsy and Parkinson’s disease. Current available techniques in this field commonly require invasive surgery potentially entailing both transient and permanent complications. A promising strategy designed to overcome these risks involves the exploitation of the cerebrovascular system as access route to the neural tissue. Cardiovascular stent technology, exceptionally designed for a minimally invasive procedure and maximum adherence to the vessel wall, represents an optimal scaffold for the recording/stimulating system. Moreover, given the versatility of this type of implant, peripheral nervous system could also be targeted for neural stimulation aimed at relieving intractable pain or supporting motor rehabilitation. Therefore, our goal is the development of a stent-inspired device, with electrodes embedded on its surface facing the vessel walls, for minimally invasive active and passive interaction with the surrounding neural tissue. In order to reduce the possibility of chronic local inflammation and delayed endothelialization, commonly occurring with metallic implants, our device has been fabricated using only biodegradable and biocompatible soft polymers such as poly-ε-caprolaptone (PCL) and (3-glycidyloxypropyl)trimethoxysilane-doped poly(3,4-ethylenedioxythiophene) : poly(styrenesulfonate) (GOPS-doped PEDOT:PSS), for encapsulation and electrodes respectively. Electrochemical properties of the prototypes have been tested in both static and dynamic conditions, with the latter simulating the vascular environment and its pulsatile flow. Degradation analysis in vitro has been performed as well to define the device durability. The promising results open the doors to further investigation of the potentialities of this innovative device, able to replace invasive systems currently used for neural recording and stimulation. Its development will have a huge impact in the treatment of neuropathological conditions related to both central and peripheral nervous system.

There are roughly 175,000 different medical devices in the US market. The global market for medical devices grows fast yearly, it reached nearly $521.2 billion in 2017 and should reach $674.5 billion by 2022, at a compound annual growth rate (CAGR) of 5.3% for the period of 2017-2022. Some relevant examples include joint replacements, cardiac pacemakers, hernia meshes, subcutaneous real-time glucose sensors, and neural recording devices. An important and growing subset of these devices termed here ‘active’ implants, require electrical contact to tissues. Their function demands that they have long-term electrical contact to host tissue; perhaps the most widely known example is cardiac pacemakers now found in over 235,000 people in the US alone. Important emerging disease treatments, however, also rely on active medical devices including subcutaneous glucose sensors and neural recording devices. In the case of glucose sensors, the real-time information can drive both dietary choices and mitigate Type 2 diabetes or be part of an automatic insulin delivery system that frees patients with Type 1 diabetes from self-administered insulin injections. In these cases, the implant must stay functional and non-changing in their tissue interactions for years.
Every medical device that is implanted in a person encounters the body’s foreign body response (FBR) which in short results in the body physically walling off an object from its physiology. Initially discovered in the 1970s, this multi-stage process is initiated by the innate immune system upon the introduction of foreign material. After injury and implantation, FBR activated by blood/material interaction, followed by protein adsorption, immune cells adhesion, activation, and differentiation. The whole process undergoes acute inflammation, chronic inflammation, the generation of granulation tissue and end up with the fibrous encapsulation generated by FBGCs. Foreign body response (FBR) triggered by the implantation of medical devices limits the lifespan of devices.
One of the widely used solutions is to suppress the immune system response to medical devices. Reactive oxygen species (ROS) generated during the FBR process play a key role in immune responses to implantation. ROS are generally linked to aging, cancer, and nervous system diseases. Anti-inflammatories and antioxidants, such as nonsteroidal anti-inflammatory drugs (NSAIDs), dexamethasone, masitinib, and resveratrol have been combined with implantable devices in local delivery to fight against FBR. However, molecular antioxidants are facing several problems, such as short lifetime, instability in micro-environment, and ineffectiveness induced high doses, limiting the chronic efficiency of implanted devices. Nanoparticles, well known with their high surface to volume ratios, are widely used in biomedical engineering these days. Cerium oxide nanoparticles exhibit recyclable antioxidant properties due to facile electron transfer between Cerium (III) and Cerium (IV). Recent studies reveal that ceria nanoparticles with diameters of 5 nm are non-toxic and efficient for quenching ROS in vivo. We developed a PEGylated polymeric ceria nanoparticle which is stable and biocompatible. We also observed that the antioxidant activity of cerium oxide nanoparticle is enhanced by this PEGylated surface modification which can be widely used in vivo to control the ROS levels at the implant site and relief FBR.

Application of more advanced microelectronics technologies over the past decade has increased routine extracellular recording capacity from dozens to thousands of neurons. The NIH BRAIN Initiative has catalyzed an explosion of innovative small area low noise analog circuit designs. A next generation of devices under development promises capacity of 10’s to 100’s of thousand neurons. I will review our work to develop high channel count devices, Neuropixels, and the pathway to an additional 10-50X capacity increase. Important materials innovation in recording electrodes, stress compensation layers, and circuit design are required to fuel this transformation of electrophysiology capability. This data volume has created bottlenecks for analysis. Discuss the current state of current practice and the paths needed to convert the avalanche of data to neuroscience understanding. The scale of data is so much larger than historically typical that a complete alteration of analysis and visualization is required. Finally, I will open a discussion of limits, how do we understand when we have adequately sampled the relevant neural activity for each behavior. How can we estimate that needed sampling in advance?

Chronic reliability, and thus a reduced foreign body response, is key to realizing the promise of neural probe technology in applications such as brain-machine interfaces, connectome mapping and brain disorder investigations. A variety of flexible, low density probes with small footprints have been designed to minimize the foreign body response in mouse models. However, few are applicable to larger systems such as primate brains. Such implantations are challenging; probes must be inserted ~30cm into the brain, which could require ~100mN of force. Further, accurate targeting requires less than 2 degrees deflection and may involve repeated insertions during the up to 6 hour surgical procedure. Recent micro-scale characterization of insertion mechanics into brain tissue has demonstrated that, following an initial loading peak during tissue penetration, continued load increase is attributable to friction between the implant and the surrounding tissue. These forces often exceed the initial loading peak past 2mm of insertion, and are particularly relevant for large brain implantation. We therefore present a new strategy for reaching large brain targets with smaller probes: minimizing the coefficient of friction.

By minimizing the friction coefficient of the probe we can substantially decrease the insertion load and thus minimize the critical buckling load (P_crit) required for successful implantation. Since for an Euler beam with one free end P_crit= π²EI/4L² , this enables probes with smaller cross-sectional areas (lower moment of inertia I) and lower moduli E to be inserted to the target depth L without deflection due to buckling. To demonstrate this principle, we fabricated a carbon fiber/platinum insertion device with a muscle inspired polydopamine/ poly-2-methacryloyloxyethyl phosphorylcholine ultra low friction coating (previously reported coefficient of friction μ=0.02). We characterized the insertion properties of 84mm samples with a range of diameters (100-430μm) with and without the coating utilizing an agarose brain phantom model and a nm-precision stage. For probes with a 230μm diameter, we observed 1.4 +/- 0.5 degrees deflection in the coated samples compared to 6 .2 +/- 0.3 in the control, indicating a surgically acceptable insertion accuracy at decreased cross-sectional area. We further characterized these insertion properties through MRI following a realistic surgery in porcine brain tissue.

These results suggest that incorporating a thin (<50nm) friction-reducing coating into a neural probe delivery device can enable an overall reduction in probe cross-sectional area, and therefore the expected dead zone during insertion. Such insertion devices could enable the translation of neural probe technologies from mouse brain models to complex primate systems by effectively delivering electrodes to target areas with minimal damage. In particular, we envision shape memory polymer based electrode deployment from the tip of this device, enabling long-term recording outside of any minimized dead or encapsulated zone.

Microphysiological systems, also called organs-on-chips, are rapidly being adopted as advanced in vitro tools for modeling disease, development, and injury, and as high-throughput preclinical drug screening devices. Modeling the nervous system using such devices presents unique challenges owing to its complex hierarchical structure and organization, and its unique physiological role in information transmission and processing. Past designs of neural microphysiological systems based on microfabricated chip-based platforms and 3D neural organoids have limitations in their abilities to recapitulate appropriate cell-cell interactions, morphology, and functionality seen in vivo. Microfabricated platforms can provide specified tissue organization, but typically restrict cell biology to 2D configurations, which can alter cellular physiology. Conversely, self-assembling 3D neural organoids provide a more physiologically-relevant microenvironment but fail to represent macroscale tissue organization. Our laboratory has sought to advance neural organ-on-a-chip platforms that balance simplicity and complexity to recapitulate neural physiology for scientific discovery and drug screening. We developed and commercialized a 3D peripheral nerve-on-a-chip for drug screening that enables both physiological and histological evaluation in manners analogous to clinical evaluation. We have further developed relatively simple methods for assembling multiple neural tissue types and shown the potential of this approach for modeling the transmission of physiological signals associated with pain. Recent efforts have been focused on the integration of these tissue systems with integrated microelectrode technology to enhance our ability to investigate the electrophysiological behavior of these model systems. In all of these endeavors, materials continue to serve as both enabling technologies and opportunities for innovation. Interestingly, these same model systems may serve as convenient platforms for testing novel neural interfacing materials and devices. In this presentation, I will highlight our advances in the development of neural microphysiological systems and discuss intersections with materials innovation.

Due to the trade-off between their electrical/electrochemical performance and underwater stability, realizing polymerbased, high-performance direct cellular interfaces for electrical stimulation and recording has been very challenging. We developed conductive direct cellular interfaces based on a water-stable, high-performance poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) film via solvent-assisted crystallization. The crystallized PEDOT:PSS exhibited excellent electrical/electrochemical characteristics, long-term underwater stability without dissolution/delamination, and good viability for primarily cultured neurons and cardiomyocytes over several weeks. Furthermore, the highly crystallized, nanofibrillar PEDOT:PSS networks enabled dramatically enlarged surface areas and electrochemical activities, which were successfully employed to modulate cardiomyocyte beating via direct electrical stimulation. Finally, the high-performance PEDOT:PSS was seamlessly incorporated into transparent microelectrode arrays and free-standing 3D fiber network-based interfaces for efficient, realtime recording of neuronal and cardiac tissue action potentials with a high signal fidelity. All these results demonstrate the strong potential of crystallized PEDOT:PSS as a crucial component for a variety of versatile bioelectronic interfaces.

Clinical mapping of electrophysiological activity in the human brain is routinely used for delineating margins between functional and pathological tissue, for therapeutic applications and for brain-machine interfaces. Trade-offs between spatial resolution and cortical coverage, specificity in recording and contact impedance, conformity/compliance to brain curvature/movement and complexity of grid designs often limit the scalability and the potential use of brain mapping electrodes. In this work, we developed a scalable manufacturing process to realize human-compatible electrocorticography surface grids with thousands of channels. The 7” long cortical electrodes together with our sterilizable compact connector overcome challenges associated with the required large separation between the sterile surgical field and the non-sterile acquisition electronics enabling reliable intraoperative recordings from patients undergoing neurosurgical resections. The electrode diameter and pitch were adjustable providing a spatial resolution from micrometer scale to millimeter scale. Pt nanorod electrodes were used as contact materials and demonstrated low impedances at frequencies of interest for brain activity mapping.
The spatiotemporal recording capability of the electrode is investigated first with anesthetized rat model undergoing whisker air-puff stimulations using high-density 32×32 array electrodes with 150μm pitch and 4.8mm×4.8mm coverage. The spatial mapping of stimulation evoked high gamma activity (HGA) measured from the electrode clearly distinguished functional boundaries of individual whisker barrels at 150μm resolution.
We recorded from awake patients undergoing tumor resection while performing motor tasks with our grid placed on the motor cortex. The geometry of the array included a 46×44 array with multiple 30μm diameter contacts and 1.8mm spacing extending over an area of 8cm×8cm. Motion sensors and an accelerometer were attached to a glove worn by patients who performed finger flexion/contraction tasks between thumb and the other four fingers. The motion signals were time-locked to the recordings using our custom-made intraoperative monitoring system. Both local field potentials and broadband HGA showed distinctive neural correlates of fine-motor movements for different fingers when compared to baseline. Furthermore, we recorded phase reversal boundaries during motor mapping to precisely localize the central sulcus in sub-mm scale resolution. These results with our high-density grids offer an unprecedented view of the functional organization and coordination of motor function over large brain regions in the human cortex.

We introduce sharp, needle-like Si-based neuronal probes on ultra-thin, biocompatible flexible substrates for penetrative, high-resolution, 3D mapping of brain activity. The flexible polyimide/parylene-C substrate carrier allows conformal adhesion to the cortical surface and embeds electrical leads for data streaming. We utilize optically transparent polyimide/parylene-C layer and Si substrates with conventional Si micromachining processes and double-side photolithography alignment in order to fabricate 3D Si micropillars on a flexible substrate. We report in this work two devices layouts. Device A consists of 32 individually-addressable, high aspect ratio pillars (wire-to-wire spacing of 400 μm, 6×6 array, L=100 μm and D=10 μm) on a sub-10μm thin flexible substrate. Electrode sensing sites at the tip of the pillars were engineered by coating Pt nanomesh structure which significantly increased the electrochemically active surface area of electrodes thus achieving sub-1MΩ electrochemical impedance in phosphate buffered solution (PBS). We demonstrated up to 196 days of reliable chronic implantation on mice and multimodal capabilities of the device with air puff stimulation-evoked activities from barrel cortex, optogenetically stimulated responses, and 2-photon intravascular imaging of the implant sites on the brain cortex. The scalability of this technology is shown by the preliminary recordings from Device B with 1024 microwire arrays (wire-to-wire spacing of 400 μm, 32×32 array, L=200-300 μm and sharp tips with sub-1 μm diameter) which holds the potential for large-scale, higher fidelity recordings for brain-machine interfaces.

The whole-cell patch-clamp technique is probably the most commonly used technique to record neuronal electrical properties. Although this approach provides maximum achievable quality for intracellular recordings, it does not provide a sustainable method for simultaneous recordings and it inevitably leads to cell dialysis and loss of the intracellular machinery that may be essential for the expression of certain biological functions. Neuronal in vitro extracellular electrophysiology, instead, commonly relies on recording techniques which monitor neuronal activity by means of electrodes placed in close contact with the cells at their extracellular domain [1]. Since the interface scenario between cells and electrodes clearly affects the quality of recordings, this interaction is classically investigated applying equivalent electrical circuit models.
Although conventional modelling approaches provide insights on the role of multiple interface parameters, they yet limit the studied systems in specific conditions, thus producing numerical results that are valid only for a fixed set of variables [2], ultimately impeding the understanding of the interface mechanical dynamics. However, systems theory [3], as already employed to analyse biological systems [4] or to study the cell-electrode interface in the frequency domain [5], promises to overcome these limits providing a modelling platform for multiparameter dynamic analysis.
Moving from the base of classical electrical modelling frameworks to a novel interpretation of systems theory, a time-varying state-space model is presented, achieving a clear separation between non-linear spike-generation dynamics [6] and linear cell-electrode coupling, while also recovering interface dynamics. Here the systematic variation of multiple parameters allows to perform a simulated spike sorting and to follow a target recorded spike while changing the same coupling parameters over time. Hence, by screening different planar electrodes sizes, the recording detection limits for each specific situation are identified. Extending the approach to 3D electrodes, time-varying simulations are then exploited to highlight the effect of micro protrusions and shape-curvatures on the recorded signals. In this way, by continuously defining the electrical detection limits, a dynamic geometrical optimization is enabled during the electrodes 3D design.

[1] Fromherz P., Offenhauser A., Vetter T., Weis J., A neuron-silicon junction: a Retzius cell of the leech on an insulated-gate field-effect transistor, Science, 252, 1991.
[2] Martinoia S., Massobrio P., Bove M., Massobrio G., Cultured Neurons Coupled to Microelectrode Arrays: Circuit Models, Simulations and Experimental Data, IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, 51, 2004.
[3] Magni L., Scattolini R., Advanced and Multivariable control, Pitagora, 2014.
[4] Izhikevich E.M., Dynamical Systems in Neuroscience, The Geometry of Excitability and Bursting, The MIT Press, 2007.
[5] Joye N., Schmid A., Leblebici Y., An Electrical Model of the Cell-Electrode Interface for High-density Microelectrode Arrays, 30th Annual International IEEE EMBS Conference, Vancouver, 2008.
[6] Hodgkin A. L., Huxley A. F., Propagation of electrical signals along giant nerve fibres, Proc. R. Soc. Lond. B, 140, 1952.

There is a pressing need for the development of novel materials-based tools that enable seamless biointerfaces across various spatiotemporal scales and can interrogate the nervous system with unprecedented resolution. Nanostructured materials have the potential of meeting the demands for next-generation devices given their small feature size, synthesis scalability, mechanical properties, and diverse signaling capabilities. Many forms of nanostructures, such as nanoparticles, nanowires/nanotubes, and nanomembranes, have been used for neural interfaces. They can enable electrical, optical, or magnetic interrogation of tissues, single cells, or even organelles.
In this talk, I will focus on nanoporous materials-based neural interfaces. The materials include nanoporous carbon membranes synthesized from micelle-enabled self-assembly, nanoporous silicon particles developed from nano-casting method, and nanoporous silicon membranes fabricated by wet chemical etching. They can deliver stimuli to cells and tissues through electrochemical, photothermal, and photoelectrochemical mechanisms. I will discuss the fundamentals and rationales of using the nanoporous structures for these studies, as well as diverse neural interface applications. Finally, I will discuss new materials and biological targets that could catalyze future advances.

Optogenetics, which uses light to control neurons genetically modified with light-gated ion channels, is revolutionizing neuroscience research by the precise deconstruction of neural circuits with neuron-type specificity. However, the poor tissue penetration of visible light has limited the application of optogenetics due to invasive craniotomy and intracranial implantation of obtrusive optical fibers or light-emitting diodes (LEDs). Here, we present a new form of deep-tissue light source based on mechanoluminescence nanoparticles, which can be delivered intravenously, recharged during circulation, and turned on and off by brain-penetrant focused ultrasound with millisecond precision. With this circulation-delivered nanoscopic light source, we have achieved non-invasive ‘sono-optogenetic’ stimulation of ChR2-expressing neurons in the intact mouse brain, successfully evoking hindlimb motion that was synchronized with the ultrasound pulses. Our approach provides an ‘inside-out’ light source anywhere and anytime inside the body at will, and we envisage that it can be used for any application that needs a light source deep in the body, such as in organs (e.g., spinal cord, heart and lungs) that are usually refractory to fiber implantation.

The recording of the action potentials in electrogenic cells plays a central role in many fields of research spanning from neuroscience to cardiology, cell biology and pharmacology. To this aim, many techniques have been developed both in research and in clinical applications. Broadly speaking they can be divided into two categories: direct electrical measurements performed with micro-electrodes (such as patch-clamp and multi-electrode arrays (MEAs)) and optical measurements collected from fluorescent reporters administrated to the cell cultures. However, regardless of the field of application some fundamental needs are still not met. For instance, methods based on electrodes can provide very accurate measurements but they require inserting the electrode inside the cell by porating the cell membrane. Hence, they are very invasive, thus reporting on a perturbed system. Furthermore, only a limited number of cells can be investigated (few cells for patch-clamp, thousands for MEAs). Even in the latest generation CMOS –MEAs, that aims to monitor hundreds of thousands of cells, the distances between the electrodes remain large with respect to cell size thus providing fragmented information. The latter issue can be overcome with optical methods which can record a million points on multiple scales with no fragmentation. However, dyes are well known to be cyto-toxic. Even if nontoxic, fluorescent dyes may bind to membrane receptors altering the response of the cells to the administered treatment in unpredictable ways. These are serious issues in many applications, for example to investigate compound toxicity in pharmacology. In this regard, cardio-toxicity and brain-toxicity of potential drug-compounds account for 60% of rejection cases during trials. This massive rejection rate significantly contributes to making the drug development process extremely expensive (now exceeding 2 billion per drug on average) and extremely long (13 years on average). This example, as with many others in neuroscience and cell biology, necessitates the need for developing new technologies providing non-invasiveness and large spatial scalability combined with accurate monitoring. Namely, the possibility of measuring both shape and duration of the action potential with precision approaching the one achieved in patch-clamp (intracellular).
In this talk, we will show a radically new concept for monitoring action potentials in human cells that overcomes all of the mentioned limitations. It bases on the concept of “mirror charge” in classical electrodynamics: electric charges placed in proximity of a conductor affect its spatial charge distribution thus generating mirror charges into the conductor itself. Hence, by monitoring the dynamics of the mirror charges one can monitor the dynamics of the “source charges” and the related electric potential, i.e. the action potential. The latter can be done by using charged dyes whose dynamics can be monitored by conventional optical methods. However, being the dyes placed in microfluidic chip, separated from the cell culture, the cells are subjected neither to dye contact, nor to direct light illumination, but are in a perfectly unperturbed physiological state. Remarkably, the optical signal perfectly resembles an action potential (AP) even without the need of cell membrane poration as for conventional electrical recording. We reported a set of experiments on human-derived cardiomyocytes that uniquely support our statements.
The working principle can be extended to other nanostructures such as Quantum Dot or plasmonic particles. More in general, this new approach based on “mirror or virtual cells” paves the way toward a new family of in-vitro bio-devices that can lay a new milestone in the field of electrophysiology. Such an innovative configuration could inspire a new disruptive class of electro-optical sensors or hybrid devices for optical computing.

Optoacoustic imaging is increasingly attracting the attention of the biomedical research community due to its excellent spatial and temporal resolution, centimeter scale penetration into living tissues, versatile endogenous and exogenous optical absorption contrast. State-of-the-art implementations of multi-spectral optoacoustic tomography are based on multi-wavelength excitation of tissues to visualize specific molecules within opaque tissues. As a result, the technique offers unmatched capacities for simultaneous visualization of structure (e.g., vascular anatomy, organ structure), function (total hemoglobin, oxygen saturation, blood flow, pH, metabolic rate of oxygen consumption), and molecular information (intrinsic bio-chromes, targeted agents, genetic reporters, sensors) from living tissues. Our recent work has established ultrafast functional optoacoustic neuro-tomography (FONT) technology to enable acquisition and visualization of spectroscopic information from the rodent brain at video rates. In this way, bulk changes in multiple hemodynamic parameters can be monitored non-invasively in whole mouse brains. We were also able to optoacoustically track neural dynamics using genetically encoded calcium indicators, thus overcoming the longstanding penetration barrier of optical microscopy in scattering brains. FONT captures hemodynamic responses and calcium fluxes as true three-dimensional information across the entire brain with 10 ms temporal and 100 µm spatial resolution, thus enabling large-scale neural recording at penetration depths and spatio-temporal resolution scales not covered with the existing neuroimaging techniques. Several hybrid imaging solutions have further been developed to support concurrent acquisition of various combinations of functional optoacoustic, ultrasound and fluorescence data as well as concurrent ultrasound stimulation of the brain.

Ion channels offer direct means to manipulating neural activity with specificity and temporal precision. The nanometer dimensions of these molecular machines inspire the design of neuromodulation tools operating at comparable spatial scales. In this talk I will discuss how magnetic nanotransducers based on synthetic magnetic materials can be applied to convert external weak magnetic fields into stimuli, such as heat, force, torque, pH change, and chemicals, commonly perceived by ion channels. I will illustrate the importance of magnetic properties in controlling magnetically-mediated neuronal modulation in the context of ion influxes and rodent behavior. This talk will also highlight opportunities offered by wireless magnetic modulation of cell function in neuroscience and beyond.

Ceramics find many applications in electronics, optics, energy storage, and medical implants. Synthesis of ceramic materials conventionally requires hours long processing time in a furnace. While this classic technique allows for large scale fabrication, it is not suitable for high throughput screenings necessary for efficient material discovery. New methods, such as flash sintering, rapid thermal annealing, and ultrafast high-temperature annealing, have been developed to address this issue. These methods rely on fast energy transfer to achieve uniform heating within the substrate, provide stable synthetic conditions, and speed up the processing time. However, all these approaches provide relatively standard synthetic conditions and do not enable new phases and structures.

Laser processing can enable new materials, increase research throughput, and allow for the instant prototyping of new devices. Here, we developed a modified laser processing to produce an intense photothermal effect over an elastic matrix, which allows the fast ceramics formation with a laser power as low as 10 mW. Out-of-equilibrium nature of the process and complex strain relaxation at the synthetic interface between the soft substrate and hard ceramic product create hierarchical anisotropic micro-structuring of the material on the scale from nano- to micrometers. Beyond understanding the fundamental principles of this laser-assisted synthesis, we have also explored their applications as diverse biointerfacing materials.

Next-generation brain-machine interfaces may use thin-film metals or non-metallic materials (e.g. carbon) for electrical connection to the brain. However, thin-films are not robust to electrical stimulation and carbon does not form a low impedance electrode interface with neural tissue. Thus, improving the electrical and mechanical stability of carbon fiber microelectrodes is a major challenge facing this promising technology. Surface roughening can improve interface properties, while maintaining small geometric area for targeted recording and stimulation of local neural populations. A platinum–iridium electrodeposition process has been developed for carbon fiber microelectrodes to improve the interfacial properties. We study the coating effect on impedance as well as morphology and composition of the deposited film. Carbon fiber electrodes (6.8 micron diameter) with parylene insulation and an exposed tip of approximately 10 microns were coated with a film of platinum-iridium. Accelerated soak test in a heated (T = 50 degrees C) saline solution showed impedance variation of less than 8 % after 36 days (89 equivalent) of soaking. Compared to uncoated fibers, impedance modulus of carbon fibers coated with platinum-iridium was decreased (on average) by 96% at 10 Hz, 89% at 1 kHz, and 13% at 1 MHz, due to the high surface roughness of the platinum-iridium film. Loss of platinum-iridium film was noted in 3 of 43 electrodes after soaking. The small decrease in impedance at 1 MHz indicates that geometric area was maintained. SEM analysis confirmed that deposition was confined to the area with exposed carbon. Energy dispersive X-ray spectroscopy (EDS) quantified the percentages of platinum and iridium as ~ 70% Pt, ~30 % Ir on the surface and the film appeared uniformly deposited. Focused ion beam was used to cut the coated microelectrode for cross sectional analysis. The film was about 400 nm thick and mostly continuous with some cracks. EDS was performed on the cross section using scanning transmission electron microscopy. As with the surface analysis, the cross-sectional analysis showed a homogeneous distribution of platinum and iridium with relative percentages similar to the surface measurements. Electrodeposition of platinum iridium is a feasible approach to decreasing the impedance of carbon fiber electrodes.

We continue to investigate the design, synthesis, and characterization of functionalized poly(3,4-ethylenedioxythiophene) (PEDOT) for use in a variety of applications including neural interfaces, as well as chemical and biological sensors. Here, we will present our recent results focused on maleimide-functionalized PEDOT variants that provide us a wide range of novel opportunities for post-polymerization modification with particular biomolecules. These copolymers can be precisely deposited onto electrodes using oxidative electrochemical deposition methods. Of particular current interest to us are the development of new materials that show specific responses to such biological and chemical agents as proteins, peptides, neurotransmitters, and opioids. We examine the morphology of our materials with optical and electron microscopy, as well as X-ray and electron diffraction. We examine the physical chemistry with UV-Vis, FTIR and Raman spectroscopy. Information about electrical transport is obtained by CV, electrochemical impedance spectroscopy, four-point probe, and transconductance measurements.

Stimulation of the peripheral nervous system through implanted devices, often referred to as bioelectronic medicines, electroceuticals, or neuromodulation therapies, is increasingly being used to treat a variety of diverse conditions including FDA approvals for pain, overactive bladder, obesity, sleep apnea, migraine, essential tremor, and hypertension. Unfortunately, the availability of neuromodulation therapies is usually limited to patients that have failed conventional treatments. This is largely due to the need for invasive implantation procedures and resulting surgical risk. we have developed a minimally invasive method for stimulating the peripheral nervous system through a syringe-injectable electrode which is delivered as a pre-polymer composite material and cures in vivo, with the goals of reducing surgical risk, increasing the availability of neuromodulation therapies, and enabling the development of novel therapeutic strategies,. This electrode, referred to as the Injectrode, can conform to complex target structures and is orders of magnitude less stiff compared to conventional neural stimulation electrodes.
This talk will cover our work developing the Injectrode, validating the material properties of the Injectrode, and our ongoing progress towards non-invasively interfacing with the Injectrode through a transcutaneous interface. This will include microscopic and electrochemical testing to assess the material properties most relevant to electrical stimulation of the nervous system mimicking FDA pre-clinical test requirements, as well as a comparison to clinical vagus nerve stimulation electrode. Dr. Ludwig will present a finite element model of the transcutaneous Injectrode concept, validated in a swine cadaver model, to investigate a method for non-invasively coupling to the Injectrode without use of an implantable pulse generator. Lastly, Dr. Ludwig will present feasibility studies in rats a swine showing stimulation of complex peripheral nerve anatomy, comparisons to a clinical nerve cuff electrode, and in vivo non-invasive coupling to the Injectrode. The ability to deliver the Injectrode through a minimally invasive procedure - as well as non-invasively couple to it without the need for an implantable pulse generator - will lead to reduced surgical risk and the potential for fewer failure points. As such, this technology fits an unmet need for less invasive stimulation systems and could lead to novel and more widely applicable neuromodulation therapies.

Biodegradable devices represent an emerging type of electronics that can degrades in physiological environments and can therefore eliminate a second surgery for device retraction and minimize associated infection risks. Such devices could play a critical role in many therapeutic and diagnostic processes at neural interfaces, including promoting tissue regeneration, probing neurotransmitters, monitoring neural signals, etc. Herein, we introduce novel materials strategies and fabrication schemes that enable a biodegradable and self-powered device for peripheral nerve regeneration. Successful nerve regrowth and motor functional recovery are realized in rodent models. Approaches detecting neurotransmitters and neural signals will also be discussed. These works will provide new insights for modulating and probing neural activities.

Techniques to study brain activities have been evolving tremendously, yet challenges remain in acquiring high throughput of simultaneous electrophysiological recordings. Here, we developed a soft neuroelectrode array that is on-site multiplexed. This novel array achieved up to 256 channels on a thin flexible substrate in only 2.3×0.3mm² area, enabled by on-site multiplexing from Si thin film transistors, and with the electrode size of only 238μm², which is capable of measuring single-unit neuronal activity. The array showed high electrical performance of average SNR of ~71 dB with over 91% yield. Mechanical bending test of 10,000 cycles proved great robustness of the device. We used 1 transistor current sensing method to reduce aliased white noise and multiplexer charge injection of the active electrodes. Acute auditory experiments in mice successfully validated in vivo use of this array. Together, these results demonstrate the promise of this on-site multiplexed, soft, penetrating neuroelectrode array for neuroscience and neural engineering.

Intraoperative neuromonitoring is a standard clinical practice in spinal cord surgeries as it is a detection method for neurological injury and helps limit postoperative impairments. Neuromonitoring is conducted by recording motor evoked potentials (MEPs) or somatosensory evoked potentials (SSEPs) via pulsed peripheral nerve or transcranial electrical stimulation (TES). These recordings are accomplished with electroencephalography (EEG) or electromyography (EMG), which require the use of large stimulation amplitudes (up to 1kV in TES and 60mA with peripheral nerve stimulation). Direct recordings from the spinal cord can reduce the stimulation thresholds required to detect neurological activity and provide another avenue for neural interface devices in the recovery from spinal cord injury. When performed with high channel count electrodes, direct recordings from the spinal cord can help to better localize pathologic spinal tissue. Standard clinical electrodes cannot provide intimate contact with spine and their spatial resolution is not suitable for these types of recordings.

Here we report and discuss the applicability of thin-film conformal and low impedance microelectrodes with hundreds of channels for neuromonitoring during spinal cord surgical resection and the potential translation of this technology to neuromodulation after spinal cord injury (SCI). We recorded from the pial surface of the spinal cord from patients undergoing an intramodular tumor resection during both transcranial and peripheral nerve stimulation. We recorded evoked potentials at subclinical threshold currents (as low as 4mA) for peripheral nerve stimulation compared to thresholds amplitudes (16mA) for clinical detection in these patients. The subdural placement permitted us to record activity that was otherwise missed with standard clinical and even high-density epidural electrodes. With the high channel count, we observed well-defined spatial variability depending on stimulus type: sensory versus motor and upper versus lower limb. Using these patterns, we investigated the applicability of our electrodes for neuromodulation after SCI via a brain-spine interface (BSI) by classifying sensory and motor events in the spinal cord. We took the first steps in performing such classification by using a supervised learning model to classify the applied stimulus type based on the spatiotemporal features of a single-trial recorded response. Taken together, these advances highlight the potential applicability of thin-film microelectrode arrays to transform neuromonitoring and to develop a soft subdural spinal cord implant for a BSI.

Recent advances in techniques and methodologies for dissecting complex neural circuits are poised to unravel the function of neural networks and the underlying causes of neurological diseases¹. Electrical stimulation of neural tissue and recording of neural activity play critical roles for our understanding of information transfer and processing within the nervous system. The traditional strategy for this purpose mainly exploits metal-based microwires that are mechanically and biologically incompatible with tissue. Recently, thermally-drawn multimaterial fibers that integrate electrodes and other modalities have emerged as promising tools for deciphering the neural circuitry in a unique way^2-4. The development of in-fiber electrodes with low impedances, small feature sizes, arbitrary geometries and excellent electrochemical properties that allow for efficient both electrical stimulation and recording, however, remains challenging. Here we, for the first time, achieve such capability via thermal drawing of a peculiar material system — metallic glasses (MGs) — that is metallic yet exhibits disordered atomic-scale structure within a polymer matrix⁵. Thanks to the control over the interplay between fluid instabilities and crystallization dynamics, we are able to fabricate ultra-long MG electrodes spanning a wide size range from a few micrometers to around 40 nm. Designing the structures in a macroscopic scaled-up preform enables the fabrication of structured MGs with arbitrary transverse geometries previously unachievable. The electrochemical characterization reveal that these novel MG electrodes surpass in all accounts typical in-fiber electrodes including metallic or polymeric electrodes found in state-of-the-art fiber probes. The charge storage capacity is superior compared to typical metallic electrodes such as commercialized Pt and PtIr. The electrochemical stability is demonstrated by the long-term cyclic voltammetry and voltage transient cycles, is even on par with that of the state-of-art carbon nanotube and graphene fibers⁵. These merits can be attributed to the intrinsic microstructural homogeneity and isotropy of structured MGs. We demonstrate the unique applications of these fiber probes integrating multiple MG conductors embedded in a biocompatible polymer, and microfluidic channels for electrical stimulation and recording plus localized pharmacological manipulation in the deep brain of animals. Specifically, chronic experiments in rats show the ability to elicit robust behavioral responses when delivering electrical neurostimulation in deep structures, to record neuronal activity during unconstrained locomotion, and to deliver pharmacological agents to manipulate local circuits. Our MG probes are implantable, miniaturized and fully functional to form stable brain-machine interfaces in freely moving animals paving the way towards innovative long-term, multi-functional neuro-probes.

W.Y, N.J contributed equally.
1. Wei Yan, et al. Advanced Multimaterial Electronic and Optoelectronic Fibers and Textiles. Advanced Materials 31 (2019) 1802348.
2. Canales, A. et al. Multifunctional fibers for simultaneous optical, electrical and chemical interrogation of neural circuits in vivo. Nat. Biotechnol. 33, 277–284 (2015).
3. Guo, Y. et al. Polymer Composite with Carbon Nanofibers Aligned during Thermal Drawing as a Microelectrode for Chronic Neural Interfaces. ACS Nano 11, 6574–6585 (2017).
4. Frank, J. et al. Next-generation interfaces for studying neural function. Nature Biotechnology 37, 1013-1023 (2019).
5. Wei Yan, Inès Richard, Güven Kurtuldu, Niclolas James, Giuseppe Schiavone, Tung Nguyen-Dang, Tapajyoti Das Gupta, Yunpeng Qu, Jake Cao, Reinis Ignatans, Stéphanie P. Lacour, Vasiliki Tileli, Grégoire Courtine, Jörg F. Löffler, Fabien Sorin*. Structured nanoscale metallic glasses fibers with extreme aspect ratios. Nature Nanotechnology. (In press, 2020).

Neuromodulation of the central nerve system (CNS) and periphery nerve system (PNS) is promising next-generation therapeutics for many neurological disorders. Direct physical stimulation such as electrical, magnetic, optical, and acoustic is an alternative approach to a pharmaceutical approach. Among various modalities, ultrasound is an attractive modality that offers distinct advantages, such as high spatial resolution, in-depth targeting, mature technology on beam-forming, and substantial data on biosafety of ultrasound on human. However, as neuromodulation based on ultrasound is in the early stage of development, further investigations on neuromodulation mechanism, therapeutic effects, and optimum neuromodulation protocols are still required. Thus, many new technological advancements such as miniaturized devices based on microelectromechanical systems (MEMS) technology have been recently introduced for the specific purpose of neuromodulation. In this talk, these new neurotools which are essential to uncovering the fundamental mechanisms of ultrasound brain stimulation and essential to developing an effective therapeutic means for brain disorders will be discussed. Specifically, this talk will discuss the development of miniaturized ultrasound transducers for small animal experiments and also our research efforts to achieve higher spatial resolution. The functionalities and potential applications of our systems will also be discussed. This research was supported by the Brain Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (2016M3C7A1904343).

Clinically, the safety thresholds for electrical stimulation have long been empirically established by the Shannon’s equation. The Shannon’s equation models the charge injection needed to produce tissue damage but does not take into account the electrochemical safety limits for the stimulating electrodes. In this work, we investigated the electrochemical safety limits for bipolar current clamped electrical stimulation using Platinum based depth and strip macroelectrodes in pig cortical tissue. We contrasted these limits against those established by benchtop testing and investigated the observed differences.
To characterize these limits, we first measured the tissue impedance using a 4-electrode setup. This allows us to isolate the electrode-tissue impedance, which can be subsequently modeled in terms of resistive and capacitive circuit elements and used to analyze the voltage drop across each. We subsequently performed cyclic voltammetry measurements to establish the water window both in saline and in-vivo. In bipolar stimulation, the electrochemical safety windows are larger than the traditionally established -0.6V and 0.8V and is dependent on the geometry of the electrode.
We measured the voltage transients for both types of electrodes, using different current amplitude levels and pulse widths. This allows us to study the variation of the Charge Injection Capacity (CIC) for each individual electrode as a function of the pulse width. Our results show that the CIC is much lower in the brain, as compared to that calculated from benchtop testing. This is a consequence of the higher tissue-electrode impedance in-vivo. We developed basic circuit models that capture the complete electrochemical interface and can reproduce the experimental results with good accuracy. We compared these results with the clinical levels used for stimulation, especially those determined from the Shannon’s equation, and found that the electrochemical safety limits established are lower than those used in clinical stimulation. These electrochemical limits are found to be lower than the reported tissue damage thresholds. We argue that the electrochemical limits and currents established experimentally in this study should therefore be considered the upper bounds for clinical stimulation.

For the last several decades the neural engineering community has developed biocompatible interfaces for modulating the central and peripheral nervous systems. Related to such development there has been an enhanced design for reliability and robustness in implantable devices. The impacted areas range from retinal implants to electrode coatings. In this talk I will focus on those two devices and designs. From finite element modeling that leverages experimental results and merges electrochemical data (impedance and cyclic voltammetry) to device testing in chronically implanted animals, I will present results from our lab that show how materials can be combined in order to achieve impedance matching for applications in stimulation or recording of neural activity.
The sensorial prostheses designed for human use have gone through several design cycles. In this talk I will present a background of both epiretinal and subretinal designs for implants, and how materials were adapted for such objectives. Not only deep brain stimulators took advantage of platinum and iridium metal coatings, also retinal implants provided valuable experimental and modeling data to guide further designs. One critical example, which I will discuss in detail, is how the impedance of different layers of the retina makes it impossible to specify a single cell when stimulating using epiretinal or subretinal implants, and how the recording is impacted by this same feature of layered brain structure. I will discuss options to overcome such limitations based on engineering design. Novel combinations of materials designed to match Young’s modulus of the target tissue or to bio-dissolve over time after implantation.
The second topic to be discussed in detail is the effectiveness of coatings to enhance robustness and reliability of chemical, optical, and electrical implantable interfaces such as electrodes and sensors. For example, combining high surface area with several coating materials allow us to keep electrode micrometric size while increasing the charge delivery and guaranteeing biocompatibility of the device.
Finally, I will present possible future developments in engineered devices for immediate applications in biomedical engineering, specifically in combining multimodal sensors and hybrid systems with electronics embedded.

Microelectrodes can treat injuries and disorders such Parkinson’s disease through electrical stimulation, and can be used as powerful tools to help us learn more about neurological functions through neural recording. However, the foreign body response (FBR) can contribute to microelectrode failure over long implant durations. After insertion, glial cells eventually surround the implant, decreasing neural adhesion. The inflammatory environment may also inhibit neural function. As published previously, Janus base nanotubes (JBNTs) are synthetic molecules with a double ring mimicking DNA G^C bases and a lysine side chain. They present a potentially greater degree of biocompatibility than many microelectrode materials, eliciting fewer negative responses from the body. This project involves applying these nanotubes to the surfaces of microelectrodes, while avoiding negatively impacting the electrochemical properties of the microelectrode itself. Other chemical and medicinal coatings have also been shown to elicit positive physical responses but tend to disrupt the electrochemical properties of microelectrodes.

Initial tests of JBNTs have been conducted on developed microelectrodes. The microelectrodes have 2.5 mm long silicon shanks, gold/platinum/titanium traces, plasma-enhanced chemical vapor deposition (PECVD) silicon nitride and silicon dioxide insulation, and electrode sites with areas 2000-4000 µm² made of iridium. The microelectrodes were coated with RNTs through dipping, and SEM was used to examine the nanotube distribution on the electrode surfaces.

The tubes aggregated on the surfaces of the electrode sites, forming bundles around the edges. Electrochemical testing showed charge injection values remained largely unchanged, as did the impedances at 1 kHz. The cyclic voltammograms showed similarly consistent results, indicating the nanotubes did not contribute to a significant alteration of the microelectrodes’ electrochemical characteristics. Additional electrochemical testing is ongoing, with electrodes of different sizes and materials. We expect that someday, JBNTs may be demonstrated as an important component of highly biocompatible, long-lasting neural microelectrodes.

Monitoring neurotransmitters in vivo necessitates sensors that approach the spatiotemporal resolution of neuronal communication, while differentiating similarly structured neurochemicals with high selectivity. We have developed biosensors based on aptamers, coupled to ultra-thin In₂O₃ (~3-nm thick) field-effect transistors (FETs). These biosensors show high sensitivity and selectivity for serotonin and dopamine with femtomolar detection limits in vitro. To enable in vivo measurements, we miniaturized our sensor architecture to fabricate neuroprobes with arrays of aptamer-field-effect transistors. The initial prototype for an implantable microprobe is fabricated using micro-electro-mechanical systems (MEMS) technologies on silicon (150-μm thick). Each neuroprobe has lithographically patterned interdigitated gold electrodes deposited atop ultrathin, semiconducting indium oxide channels, where the shank is of 150-μm wide and 150-μm thick. Functionalization of the exposed semiconducting surfaces with neurotransmitter-specific aptamers facilitated electronic neurotransmitter sensing. Reducing individual FET footprints leads to an increase in the numbers of independent transistors on a single neuroprobe for multi-site/multi-target detection capabilities. We characterized target sensitivity ranges, detection limits, and sensor stabilities in vitro prior to conducting in vivo measurements. We assessed the biofouling of neuroprobes using ex vivo measurements in brain tissue homogenates. We also fabricated flexible neuroprobes using polyimide as substrates instead of rigid Si to reduce immunological responses. Shuttle devices were used to assist the implantation of flexible polyimide neuroprobes. We are now working on shuttle-free flexible neuroprobes with temperature-responsive polymers, which are hard at room temperature for implantation and soft in the brain after implantation. Implantable aptamer-field-effect transistor-neuroprobes, including additional designs on flexible polymers, will enable in vivo detection of neurotransmitters in the brain with high spatial resolution and selectivity.

Implantable neural interfaces can record neural signals and electrically stimulate neurons. Several neural interfaces integrated with additional functions such as optical stimulation and drug delivery have been introduced recently. Among various neural electrodes, needle-type neural electrodes, for example, the Utah electrode arrays, have been widely used in non-human primate (NHP) and clinical studies. However, drug delivery function for the needle-type neural electrodes has not been implemented yet because of their structural characteristics. We previously developed a neural interface based on a flexible penetrating microelectrode array (FPMA), which is a needle-type neural electrodes array. Here, we propose a microfluidic flexible interconnection cable (µFIC), which is capable of delivering drugs to the FPMA, resulting in drugs delivered to the implant site. The neural signal recording was performed while delivering drugs to the brain in vivo experiments using the fabricated neural interface. The KCl-evoked excitation signals of neurons in the rat brain were recorded during the KCl solution delivery. Approximately 50% of the electrodes detected an elevation of the number of spikes and amplitude.

High-density neural interfaces are continuously progressing towards seamless integration of brain and electronics. As more effort has been dedicated to implantable devices, effective and non-invasive data communication between the implant and outside electronics becomes a critical unresolved issue. Although radio-frequency (RF)-based communication can provide a wireless link with high bandwidth, RF circuits suffer from high power consumption, complex rigid circuitry for signal transmission, and wave reflection that limit their capacity for prolonged used inside biological tissue. There is a tremendous need for a safe, low consumption, high-bandwidth wireless communication method for use in neural interface devices. Here we propose a fully integrated, conformable, high-channel count neural interface device equipped with a high-speed ionic communication (IC) link. We implanted the device in rodents and validated its performance in acquiring chronic in vivo electrophysiological data from freely moving animals without any tissue extruding components. To achieve this, we first designed a conformable ionic communication circuit for transmitting and receiving ion based signals. We then characterized the performance of this ion-based communication in terms of its bandwidth, depth of propagation, and capacity for parallel transmission at different operating frequencies. Next, we designed and tested IC transmitter and receiver pairs with various geometries to better understand the key parameters defining bandwidth and stability of the ionic link. Lastly, we were able to integrate the optimized configuration of each subsystem into a compact, fully implantable device. The proposed ionic communication scheme and the resulting devices demonstrate concrete steps toward a high performance, fully implanted neural interface system with prolonged effective operation time in vivo.

Bioresorbable electronic materials serve as the basis for emerging classes of biomedical implants with operational lifetimes designed to match those of transient biological processes. These devices provide important functionality during a relevant timeframe and then naturally disappear within the body, thereby eliminating the need for surgical extraction. This talk presents our latest work in this area, with emphasis on materials aspects of (1) platforms for wireless neuromuscular electrical stimulation, with demonstrations of alleviated muscle atrophy resulting from denervation in rodent models via multiple episodes of the distal nerve engagement and (2) devices for programmable release of drugs from multiple, independently addressable reservoirs, with illustrations in insulin and lidocaine delivery in animal models for control of blood glucose concentrations and pain responses, respectively.

Neural interfaces promise to deliver more effective treatment for conditions affecting the brain and body by improving the specificity and timing of therapies. While miniaturization in key to improve biocompatibility and safe deployment of these technologies, we must also improve the specificity of devices that sense and actuate biological signals. In this talk, I will describe on-going work to co-design genetically engineered cells and miniature magnetic/optical devices. These biohybrid technologies combine the speed and programmability of electronics with the biochemical specificity of synthetic biology. Specifically, we will describe how magnetoelectric and magnetothermal biohybrids enable rapid and precise actuation of cell activity, while planar imaging devices combined with computational reconstruction algorithms enable high-resolution measurement of biological activity. Together these technologies can enable future minimally and minutely invasive links between the body and external electronic devices.

One of the grand challenges in neuroscience is to understand how neuronal activity is orchestrated to produce function. Single modality technologies offer either high spatial or high temporal resolution and with varied degrees of coverage. Multimodal approaches that can probe the brain’s activity from the surface, at depth, and with high spatiotemporal resolution are believed to yield results that may complete our understanding of behavior and function. We report the seamless integration of the electrical and optical modalities for comprehensive and interactive 3D probing and manipulation of cortical neuronal activity and optogenetic (OG) stimulation/inhibition. Our “Windansea” technology, named after the famous Windansea beach in La Jolla, will replace the current cranial glass windows providing measurements of the extracellular local field potential (LFP) and multiunit activity (MUA), as well as a see-through optical access. The Windansea allows us to obtain depth-resolved LFP/MUA information, which can be used for estimation of spiking and synaptic activity in specific cortical neuronal populations, while simultaneously sampling electrical neuronal activity from the cortical surface and implementing single/multiphoton optical imaging and optogenetic stimulation/inhibition in behaving animals.

We leveraged microelectromechanical systems (MEMS) surface micromachining techniques to developed the hybrid surface-depth electrode. The depth electrode can be maintained inside the cortex chronically and does not block optical access. Therefore, both the depth and surface electrodes are flush with the cortical surface: The integrated depth e;ectrode in our device utilizes an inserter with a small neck-region (narrow-constriction) of SU8 that is sheared (thermally or mechanically) to separate the handle from the implanted portion of the depth electrode so that a chronic transparent cranial window can be placed on the cortical implant. The device was implanted chronically in mice for several weeks and we recorded electrophysiological signals and obtained single/multiphoton optical images in awake animals.

A telemetric methodology can be used for conducting efficient and effective preclinical research in small nonhuman primates during free movement. In this study, we propose a fully implantable brain-measurement system using stretchable components and a small measurement instrument that wirelessly records brain potentials, stimulates neural activities, and recharges its own battery. When the integrated system was fully implanted in a marmoset, simultaneous measurements of the cortical and local field potentials were successfully performed during free movements. Even after a month of the implantation, the marmoset remained in good health and displayed activity equivalent to the amount before implantation. The results confirm that the proposed system is minimally invasive.

A sheet-type stretchable electrode fabricated using Au/Au core/shell nanowire (AgNW/Au) was used to measure the somatosensory-evoked potentials after a few months of implantation. Antithrombogenic polymer treatment was introduced in the electrode to stabilize the monitoring, thus preventing the growth of granulation tissues that could interrupt access to the targeted neuron [1]. Furthermore, the sheet-type stretchable electrode showed mechanical durability at 60% strain and optical transparency over 80%. The multifunction simultaneously realized an optogenetic stimulation and in vivo brain ECoG monitoring. The integrated system is equipped with a micro-LED light irradiation system and is compatible with the optogenetic stimulation, which could contribute to brain–machine interfaces with a feedback sensation for diagnoses and therapy of patients suffering from intractable diseases [2,3,4].

[1] J. M. Anderson, et al., Semin. Immunol., 20 (2), 86–100, 2008.
[2] S. Royer et al., Eur. J. Neurosci., 31 (12), 2279–91, 2010.
[3] A. L. Miguel et. al., Nat. Rev. Neurosci., 10 (7), 530–40, 2009.
[4] Dong-Wook Park et al., Nature Comm., 5, 5258, 2014.

Investigation of neural circuits in the immature brain is critical to understand how brain functions emerge and can be disrupted. The developing brain undergoes significant structural and molecular changes postnatally, some of which are experience-dependent. Mice provide one of the most common and accessible methods to model the human brain and its disorders. However, longitudinal monitoring of high spatiotemporal resolution neural activity in small, fragile mouse pups as they develop from newborns to adults is currently not possible. Here, we propose a flexible, stretchable neural interface device that allows chronic in vivo acquisition of neurophysiologic signals during postnatal development. The high biocompatibility, conformability and ultralow thickness of the device allow it to be hosted in the body for a duration of at least 15 days without inducing tissue reactions. The probe is sufficiently flexible to match the low elastic modulus of the tissue and designed in a stretchable configuration to expand with the pup’s growth via serpentine-shaped interconnects. An electrocorticography array (NeuroGrid) is placed on the cortical surface and each electrode is connected to a net of internal ion-gated organic electrochemical transistor (IGT)-based voltage amplifiers. The high speed and transconductance of the IGTs along with their independently gated activation create a scalable integrated circuit that amplifies and multiplexes the recorded signals. The interconnects elongate inside the subcutaneous space of the pup’s back, and electrical contact pads placed on the back of the animal are sutured through the tissue to the backend electronics via conductive PEDOT:PSS fibers. This approach will enhance knowledge of brain development in health and disease.

Complex interactions between neurons in different brain regions have been studied using a variety of technologies including microwires, silicon-based multielectrode arrays, and electrode arrays with flexible substrates. Such technologies have been shown to provide high-resolution electrophysiological recordings with high signal-to-noise ratio. However, with single implantation, the physical properties of these devices limit their access to one, small brain region. To overcome this limitation, we developed a platform that provides three-dimensional coverage of brain tissue through multifunctional polymer fiber-based neural probes capable of interfacing simultaneously with neurons in multiple sites. A scaffold with helix hollow channels is employed to achieve the 3D expansion of the multifunctional fiber probes in the deep brain while a femtosecond laser micromachining technique is utilized to expose electrode recording sites, microfluidic channel openings, and waveguide windows along the single fiber branch. By using our 3D fiber probes, we have obtained chronic endogenous neural activities with single-unit resolution from the wild type mice and captured optically stimulated neural signals with chemical modulation from transgenic mice expressing channelrhodopsin-2demonstrate. Furthermore, varying electrographic brain activities from different brain regions of interest were detected by our customizable multisite 3D fiber-based neural probes in a mouse model of epilepsy, suggesting the potential of using these probes for the investigation of brain disorders such as epilepsy. Ultimately, we envision that this 3D multifunctional neural interface can bring new insights for deciphering complex brain functions and dynamics in the near future.

Bioresorbable electronic materials serve as the basis for emerging classes of biomedical implants with operational lifetimes designed to match those of transient biological processes. These devices provide important functionality during a relevant timeframe and then naturally disappear within the body, thereby eliminating the need for surgical extraction. This talk presents our latest work in this area, with emphasis on materials aspects of (1) platforms for wireless neuromuscular electrical stimulation, with demonstrations of alleviated muscle atrophy resulting from denervation in rodent models via multiple episodes of the distal nerve engagement and (2) devices for programmable release of drugs from multiple, independently addressable reservoirs, with illustrations in insulin and lidocaine delivery in animal models for control of blood glucose concentrations and pain responses, respectively.

Ion channels offer direct means to manipulating neural activity with specificity and temporal precision. The nanometer dimensions of these molecular machines inspire the design of neuromodulation tools operating at comparable spatial scales. In this talk I will discuss how magnetic nanotransducers based on synthetic magnetic materials can be applied to convert external weak magnetic fields into stimuli, such as heat, force, torque, pH change, and chemicals, commonly perceived by ion channels. I will illustrate the importance of magnetic properties in controlling magnetically-mediated neuronal modulation in the context of ion influxes and rodent behavior. This talk will also highlight opportunities offered by wireless magnetic modulation of cell function in neuroscience and beyond.

Application of more advanced microelectronics technologies over the past decade has increased routine extracellular recording capacity from dozens to thousands of neurons. The NIH BRAIN Initiative has catalyzed an explosion of innovative small area low noise analog circuit designs. A next generation of devices under development promises capacity of 10’s to 100’s of thousand neurons. I will review our work to develop high channel count devices, Neuropixels, and the pathway to an additional 10-50X capacity increase. Important materials innovation in recording electrodes, stress compensation layers, and circuit design are required to fuel this transformation of electrophysiology capability. This data volume has created bottlenecks for analysis. Discuss the current state of current practice and the paths needed to convert the avalanche of data to neuroscience understanding. The scale of data is so much larger than historically typical that a complete alteration of analysis and visualization is required. Finally, I will open a discussion of limits, how do we understand when we have adequately sampled the relevant neural activity for each behavior. How can we estimate that needed sampling in advance?

There is a pressing need for the development of novel materials-based tools that enable seamless biointerfaces across various spatiotemporal scales and can interrogate the nervous system with unprecedented resolution. Nanostructured materials have the potential of meeting the demands for next-generation devices given their small feature size, synthesis scalability, mechanical properties, and diverse signaling capabilities. Many forms of nanostructures, such as nanoparticles, nanowires/nanotubes, and nanomembranes, have been used for neural interfaces. They can enable electrical, optical, or magnetic interrogation of tissues, single cells, or even organelles.
In this talk, I will focus on nanoporous materials-based neural interfaces. The materials include nanoporous carbon membranes synthesized from micelle-enabled self-assembly, nanoporous silicon particles developed from nano-casting method, and nanoporous silicon membranes fabricated by wet chemical etching. They can deliver stimuli to cells and tissues through electrochemical, photothermal, and photoelectrochemical mechanisms. I will discuss the fundamentals and rationales of using the nanoporous structures for these studies, as well as diverse neural interface applications. Finally, I will discuss new materials and biological targets that could catalyze future advances.