University of Southern California
“The history of electrophysiology has been decided by the history of electrical recording instruments.”
The desire to decipher targeted neural activities in the mammalian nervous system has inspired the development of many innovative technologies that incorporate a variety of signaling modalities (electrical, chemical, and mechanical). While decades of neural engineering have been dedicated to electronic interfaces with neural tissue, more recent advances acknowledge the multiple modalities of neural activity and make use of other interface methods. Scaling of such technologies to acquire data from large numbers of neurons remains a challenge, as is the long-term stability of the recording interface, which is susceptible to foreign body response. The goal of advancing interface technologies is both to better understand proper functioning of the brain and to address a variety of neurological, neuro-degenerative, psychiatric, and neuromuscular conditions and deficits.
BRAIN COMPOSITION AND ANATOMY
The anatomical and functional complexity of the brain poses a great challenge for engineering the interfaces needed to record and modulate its activity with precision and over short and long time scales. The human brain (~1.3 cm3) has 100 billion (1011) electrically active neurons (cell body ~20 mm diameter) that are interconnected to one another via chemically active synapses (20–40 nm wide gaps). Each neuron possesses about 103 synapses and therefore 100 trillion (1014) such connections exist in the brain. Furthermore, chemical synapses communicate
via 102 different neurotransmitters, with events occurring at a rate of 0.1–200 Hz (“firing” rate). For comparison, the number of estimated stars in the observable universe is on the order of 1022–1024.
Electrically active neurons account for only 50 percent of the cells in the brain. The other 50 percent are electrically inactive support cells (e.g., oligodendrocytes, astrocytes, and microglia; Chen et al. 2017). To support all the cell types, the organ is bathed in and cushioned by cerebrospinal fluid and nourished via blood vessels.
Neurons are anatomically organized into different regions, each with specific functions. Interfaces to neurons target specific brain regions associated with functions of interest. While it is possible to engineer devices that are smaller and computationally faster than neurons, it is not possible to fully recreate functional neural tissues. Hence, the ability to reliably interface with different brain regions is of immense scientific and clinical interest.
THE NATURE OF BRAIN ACTIVITY
Brain activity can be recorded from single neurons (single unit spikes or action potentials, APs) or from groups of neurons (multiunit recordings or local field potentials, LFPs) in a particular region. An inactive neuron has a resting potential of −70 mV (cell membrane voltage). An active neuron exhibits an AP of 80–100 mV, lasting a few milliseconds, that propagates down a neuron. This measurable electrochemical gradient results from the cell’s selective permeability to specific ionic species mediated by voltage-sensitive ion channels in the membrane and produces a measurable electrochemical gradient.
At the synapse, the AP induces the release of neurotransmitters, which cross the gap and bind to receptors on the cell membrane of the recipient neuron, triggering the opening of its ion channels. Signal transmission between two neurons connected by such a chemical synapse goes from electrical to chemical to electrical, where electrical signaling involves the movement of ions and not electrons. It is possible to modulate this natural activity using artificial stimuli such as electrical current or chemical agents.
The AP or “spike” from a single neuron (single unit) represents communicative activity and is measurable using invasive electrodes placed in (intracellular) or next to (extracellular) neurons. The “recording” electrode can be used in conjunction with a distant reference electrode to enable measurement of current or potential difference. Given the difficulty of targeting the interior of a single neuron, the destructive nature of such an interface, and the limited information it provides, the focus here is on extracellular interfaces.
Extracellular interfaces can also be used to record the LFP, the collective activity of a nearby group of neurons. Less invasive FPs can be recorded on the surface of the brain or through the protective dura membrane covering via electrocorticogram (ECoG) or exterior to the skull via electroencephalogram
(EEG). The less invasive the interface, the lower the resolution and the higher the tissue volume from which recordings are obtained (e.g., ~0.1 mm for an invasive penetrating electrode, compared to 2 mm for ECoG and 10 mm for EEG; Borton et al. 2013). With advances in recording interfaces, multiple penetrating electrodes can be placed to obtain recordings from multiple single units, thus achieving both high resolution and access to information from different regions of the brain.
It should be noted that while most synaptic transmission is chemical in nature (involving neurotransmitters), there are also electrical synapses that form a mechanical and electrically conductive link via a structure known as a gap junction (~3–4 nm). And some neurons possess specialized ion channels that respond to physical stimuli such as pH, temperature, pressure, and tension (Chen et al. 2017).
HISTORY OF ELECTRICAL STIMULATION AND RECORDING
Early Explorations and Applications
Electrical interfaces that interact with the nervous system have been used since ancient times, when Egyptians and Romans used electric shocks delivered by electric eels to treat pain. The foundations of bioelectricity and electrophysiology were laid much later in experiments conducted by Luigi Galvani in the 1780s, in which dead frog’s leg muscles moved in response to current applied to nerves via metal wires, a phenomenon dubbed “animal electricity.”
With the discovery of the effects of electricity on the human body (often the investigator’s own body), “medical electricity” research commenced shortly thereafter (Bresadola 1998). Giovanni Aldini, nephew of Luigi Galvani, discovered that electrical stimulation of the cerebral cortex could elicit physical responses (facial grimaces on decapitated prisoners; Aldini 1804). This discovery inspired work on brain stimulation both to understand function and as a means of therapy. In 1938, Ugo Cerletti applied electric current to the skull to induce “therapeutic” epileptic seizures to treat severe psychosis (Cerletti 1940).
Another paradigm shift occurred in 1947 when electrodes were used for intraoperative electrical stimulation to determine the location of lesioned targets with the assistance of stereotactic techniques; until that point, electrodes were used clinically to lesion the brain in neurosurgery (Spiegel et al. 1947). Brain stimulation was also investigated for pain control in the 1950s.
Together these efforts provided the foundation for new clinical therapies such as transcranial magnetic stimulation, cortical brain stimulation, and deep brain stimulation (DBS). DBS has borrowed heavily from cardiac pacemaker and defibrillator electrode concepts that were developed earlier.
Evolution of Technologies
Recordings of animal electricity were first reported by Leopoldo Nobili in 1828 using an electromagnetic galvanometer, but the first true recordings of the resting and action potentials were made in 1868 by Julius Bernstein using a differential rheotome that allowed measurement of fast electrical processes (Verkhratsky et al. 2006). The detection of currents from the brain was achieved using an early form of EEG by Richard Caton in 1875 (Grimnes 2014). Wire electrodes were used to record from behaving animals in the 1950s.
Advances in microelectronics led to the development of miniaturized multi-electrode arrays on planar surfaces in the 1970s. They interfaced with cell and tissue cultures and demonstrated that electrodes could be made at the scale of a single neuron. At about the same time, microelectrodes on penetrating probes were introduced. The technology was independently developed by multiple groups, leading to commercially available products for research in animals and investigational studies in humans, including use in clinical trials in 2004 (Chen et al. 2017). But while silicon microelectrode arrays have been developed over several decades, the inability to achieve reliable and stable long-term device-tissue interfaces has spurred interest in the development of more compliant polymer probes.
Electrodes used for stimulation and recording have opposing requirements, which prevent their simultaneous use. Smaller recording electrodes are preferred to isolate activity from single cells, whereas stimulation electrodes should have larger surface area to increase the charge injection capacity available to excite neurons. Because electrical stimulation indiscriminately activates nearby neurons and produces a large artifact that interferes with recording, its use in understanding brain activity and therapy is limited; but lowering the electrode area to minimize activation proportionately increases the input charge densities and the risk of tissue damage. Electrical stimulation is unable to inhibit activity. These drawbacks of electrical interfaces have given rise to alternative interface modalities.
Advances in genetic engineering of cells have opened new avenues to interface with neurons. In optogenetics, a neural population is genetically manipulated so that it can be selectively perturbed optically and probed electrically at the same time. This is accomplished by injecting a cell with light-sensitive microbial ion channels (opsins), which can change their conformation in response to light and affect ion transport. Unlike electrical stimulation, optogenetic approaches can both excite and inhibit neural activity.
Chemical stimulation, whether excitatory or inhibitory, can be achieved by infusing (through conventional cannulae or microfluidics) chemical agents or biological (genetic) agents to modulate activity. Electrochemical sensors provide a means of detecting neurotransmitters and can be specific to particular
electroactive neurotransmitters. They may be located nearby conventional microelectrodes, even on the same supporting substrate, and provide information about the concentration of molecules.
Nanoscale transducers introduced into brain tissue can modulate brain activity through the conversion of optical, acoustic, and magnetic stimulation into voltage or electric fields. These nanotransducers include quantum dots, gold nanoparticles, up-conversion nanoparticles, and magnetic nanoparticles. The latter can activate mechanosensitive ion channels by producing the required piconewton-level forces in the presence of a magnetic field gradient. The delivery of these nanomaterials and control of their targeting remain a challenge.
Interfaces need not be invasive. Acoustic waves and magnetic fields can be harnessed to modulate activity in the brain. Whereas electromagnetic waves in the visible and infrared spectrum have limited penetration depth (1 mm), transcranial focused ultrasound can access deeper regions (>50 mm), although at inferior spatial resolution (1 mm3). Transcranial magnetic stimulation can access the upper 10 mm but with reduced spatial resolution (Chen et al. 2017).
Although electrical and nonelectrical interfaces are discussed separately here, several have been combined to leverage the advantages of the particular technique for research purposes.
CHALLENGES AND OPPORTUNITIES
The availability of appropriate interface technologies for the brain strongly affects the ability of researchers and clinicians to understand it and develop new therapies. Yet even the limited and imperfect information currently available has resulted in clinically implemented technologies with only a few stimulating electrodes that have dramatically improved lives. DBS, for example, has been approved by the US Food and Drug Administration for the treatment of tremor (1997), Parkinson’s disease (2002), dystonia (2003), and obsessive compulsive disorder (OCD; 2009) (Sironi 2011).
Future advances seek to seamlessly integrate neural interfaces with the brain to enable both long-term recording and modulation of neurons with a higher number of input and output channels (Wellman et al. 2017). To achieve this, the health of the tissue-device interface needs to be improved by addressing tissue damage related to surgical delivery, the biological immune response, and the stability of the materials used in the construction of the interfaces. More information is needed about the effects of the complex interplay between material selection, device design, and fabrication methodology on the long-term performance and function of the device in the body. These advances are critical to obtain chronically stable high-density and large-scale recordings.
In addition, modulation technologies need improvements in reliability and precision. When used together, recording and modulation can achieve exciting new concepts in closed-loop therapeutic systems of the future. With rigorous
engineering focused on reliability, the next generation of life-changing medical technology breakthroughs can be realized.
Adrian E. 1932. The Mechanism of Nervous Action: Electrical Studies of the Neurone. Philadelphia: University of Pennsylvania Press.
Aldini G. 1804. Essai théorique et expérimental sur le galvanisme, avec une série d’expériences faites en présence des commissaires de l’Institut National de France, et en divers amphithéaâtres anatomiques de Londres par Jean Aldini. Paris: De l’imprimerie de Fournier fils.
Borton D, Micera S, Millan J del R, Courtine G. 2013. Personalized neuroprosthetics. Science Translational Medicine 5(210):210rv2.
Bresadola M. 1998. Medicine and science in the life of Luigi Galvani (1737–1798). Brain Research Bulletin 46(5):367–380.
Cerletti U. 1940. L’elettroshock. Reggio-Emilia: Istituo psichiatrico di s. Lazzaro.
Chen R, Canales A, Anikeeva P. 2017. Neural recording and modulation technologies. Nature Reviews Materials 2:16093.
Grimnes S. 2014. Bioimpedance and Bioelectricity Basics. Boston: Elsevier.
Sironi VA. 2011. Origin and evolution of deep brain stimulation. Frontiers in Integrative Neuroscience 5:42.
Spiegel EA, Wycis HT, Marks M, Lee AJ. 1947. Stereotaxic apparatus for operations on the human brain. Science 106(2754):349–350.
Verkhratsky A, Krishtal OA, Petersen OH. 2006. From Galvani to patch clamp: The development of electrophysiology. Pflügers Archiv 453(3):233–247.
Wellman SM, Eles JR, Ludwig KA, Seymour JP, Michelson NJ, McFadden WE, Vazquez AL, Kozai TDY. 2017. A materials roadmap to functional neural interface design. Advanced Functional Materials 1701269.