Table of contents

Research and academic professionals in neural engineering know well the promise the field offers for advancing our understanding of basic neuroscience and devising new therapies for treating neurological diseases and disorders. But there is also considerable commercial opportunity for new start-up companies in several areas of neural engineering. The neurotechnology industry, which includes firms that manufacture neuromodulation devices, neural prostheses, neurorehabilitation systems, and neurosensing devices, is forecast to grow to grow from $3.6 billion this year to $8.8 billion in 2012, according to a recently published market research study from Neurotech Reports.

In recent years, there have been several successful spinoffs of neurotechnology startup firms that originated with research at universities and clinical institutions. In many cases, the academic researchers who invented the new technology or product innovation have stayed on with their startup firms after receiving funding from venture capital firms, or after going public.

Among the most successful neurotechnology industry spinoffs in recent years were:

Cyberkinetics Inc., Foxborough, MA, a manufacturer of brain–computer interface devices based on research at Brown University. John Donoghue, a professor and chairman of the department of neuroscience at Brown University and executive director of the brain science program at Brown, founded the company in 2001 and remains on board as the chief scientific officer.

Synapse Biomedical, Inc., Oberlin, OH, a manufacturer of diaphragm pacing systems, based on research at Case Western Reserve University. Raymond Onders, director of minimal invasive surgery and associate professor at University Hospital Case Medical Center in Cleveland, was the primary researcher. He was helped by J. Thomas Mortimer, professor emeritus of biomedical engineering at Case, and it was Mortimer who came up with the name NeuRx for the device. Onders performed his first surgical implant of the device in 2000. Anthony Ignagni, who worked with Mortimer and Onders as project director and chief biomedical engineer, became a co-founder of Synapse and is currently president and CEO.

Afferent Corp., Providence, RI, a manufacturer of sensory stimulation systems based on research at Boston University, Afferent's technology is based on work by James Collins, professor of biomedical engineering at Boston University, who showed that low-level stochastic, or random, vibrations improved sense of touch. Collins went on to demonstrate that generating a subthreshold noise in the sensory pathways with a random electrical stimulation improves detectability of weak mechanical stimuli. In 1999 entrepreneur Jason Harry licensed the stochastic resonance technology from Boston University and received help from the Community Technology Fund-the university's technology transfer incubator.

NeuroNexus Technologies, Inc., Ann Arbor, MI, a manufacturer of implanted neural probes, based on a program at the University of Michigan. NeuroNexus was spun out of UM's Center for Neural Communications Technology. Daryl Kipke, head of the Neural Engineering Laboratory at the university, started up NeuroNexus with several colleagues and currently serves as president and CEO.

Commercialization issues were discussed recently at a preconference workshop at the 2007 meeting of the International Neuromodulation Society in Acapulco, Mexico. In the session, which was chaired by Chris Coburn of the Cleveland Clinic, neurotech entrepreneurs John Bowers from Northstar Neuroscience and Ben Pless, formerly of NeuroPace, shared their experiences bringing neuromodulation therapies to market. Coburn related his observations from the Cleveland Clinic, which has spun off 22 companies over the last five years. He cited several factors that would influence a neurotech startup's market potential, such as identifying the regulatory pathway, any predicate devices that exist, and the revenue potential for potential investors.

Bowers cited several factors that would be critical for the growth of the industry, including reaching out to the 'influential implanters' and pushing reimbursement based on quality of life improvements, rather than mortality. Pless provided an overview of patent-related issues affecting neuromodulation devices. He noted that there have been 2000 issued patents for neuromodulation devices, most of them involving the peripheral nervous system.

The example set by these companies and other startups in the neurotechnology industry should remind neural engineering researchers that there are plenty of opportunities for commercial success based on promising research at university and clinical research laboratories. Given the wide range of help and information that is available to potential entrepreneurs in this field, we expect to see many more examples of neurotechnology spinoffs in the years ahead.

The theoretical groundwork of the 1930s and 1940s and the technical advance of computers in the following decades provided the basis for dramatic increases in human efficiency. While computers continue to evolve, and we can still expect increasing benefits from their use, the interface between humans and computers has begun to present a serious impediment to full realization of the potential payoff. This paper is about the theoretical and practical possibility that direct communication between the brain and the computer can be used to overcome this impediment by improving or augmenting conventional forms of human communication. It is about the opportunity that the limitations of our body's input and output capacities can be overcome using direct interaction with the brain, and it discusses the assumptions, possible limitations and implications of a technology that I anticipate will be a major source of pervasive changes in the coming decades.

A half-day forum to address a wide range of issues related to translational neural engineering was conducted at the annual meeting of the Biomedical Engineering Society. Successful practitioners of translational neural engineering from academics, clinical medicine and industry were invited to share a diversity of perspectives and experiences on the translational process. The forum was targeted towards traditional academic researchers who may be interested in the expanded funding opportunities available for translational research that emphasizes product commercialization and clinical implementation. The seminar was funded by the NIH with support from the Rehabilitation Institute of Chicago. We report here a summary of the speaker viewpoints with particular focus on extracting successful strategies for engaging in or conducting translational neural engineering research. Daryl Kipke, PhD, (Department of Biomedical Engineering at the University of Michigan) and Molly Shoichet, PhD, (Department of Chemical Engineering at the University of Toronto) gave details of their extensive experience with product commercialization while holding primary appointments in academic departments. They both encouraged strong clinical input at very early stages of research. Neurosurgeon Fady Charbel, MD, (Department of Neurosurgery at the University of Illinois at Chicago) discussed his role in product commercialization as a clinician. Todd Kuiken, MD, PhD, (Director of the Neural Engineering for Artificial Limbs at the Rehabilitation Institute of Chicago, affiliated with Northwestern University) also a clinician, described a model of translational engineering that emphasized the development of clinically relevant technology, without a strong commercialization imperative. The clinicians emphasized the importance of communicating effectively with engineers. Representing commercial neural engineering was Doug Sheffield, PhD, (Director of New Technology at Vertis Neuroscience, Inc.) who strongly encouraged open industrial–academic partnerships as an efficient path forward in the translational process. Joe Pancrazio, PhD, a Program Director at NIH's National Institute of Neurological Disorders and Stroke, emphasized that NIH funding for translational research was aimed at breaking down scientific barriers to clinic entrance. Vivian Weil, PhD, (Director of Center for the Study of Ethics in the Professions at the Illinois Institute of Technology) a specialist on ethics in science and engineering, spoke of the usefulness of developing a code of ethics for addressing ethical aspects of translation from the bench to clinical implementation and of translation across disciplines in multi-disciplinary projects. Finally, the patient perspective was represented by Mr Jesse Sullivan. A double-arm amputee and patient of Dr Kuiken's, Mr Sullivan demonstrated the critically important role of the patient in successful translational neural engineering research.

Recent advances in Kalman filtering to estimate system state and parameters in nonlinear systems have offered the potential to apply such approaches to spatiotemporal nonlinear systems. We here adapt the nonlinear method of unscented Kalman filtering to observe the state and estimate parameters in a computational spatiotemporal excitable system that serves as a model for cerebral cortex. We demonstrate the ability to track spiral wave dynamics, and to use an observer system to calculate control signals delivered through applied electrical fields. We demonstrate how this strategy can control the frequency of such a system, or quench the wave patterns, while minimizing the energy required for such results. These findings are readily testable in experimental applications, and have the potential to be applied to the treatment of human disease.

The performance of current EEG-based self-paced brain–computer interface (SBCI) systems is not suitable for most practical applications. In this paper, an improved SBCI that uses features extracted from three neurological phenomena (movement-related potentials, changes in the power of Mu rhythms and changes in the power of Beta rhythms) to detect an intentional control command in noisy EEG signals is proposed. The proposed system achieves a high true positive (TP) to false positive (FP) ratio. To extract features for each neurological phenomenon in every EEG signal, a method that consists of a stationary wavelet transform followed by matched filtering is developed. For each neurological phenomenon in every EEG channel, features are classified using a support vector machine classifier (SVM). For each neurological phenomenon, a multiple classifier system (MCS) then combines the outputs of the SVMs. Another MCS combines the outputs of MCSs designed for the three neurological phenomena. Various configurations for combining the outputs of these MCSs are considered. A hybrid genetic algorithm (HGA) is proposed to simultaneously select the features, the values of the classifiers' parameters and the configuration for combining MCSs that yield the near optimal performance. Analysis of the data recorded from four able-bodied subjects shows a significant performance improvement over previous SBCIs.

To explore the reliability of a high performance brain–computer interface (BCI) using non-invasive EEG signals associated with human natural motor behavior does not require extensive training. We propose a new BCI method, where users perform either sustaining or stopping a motor task with time locking to a predefined time window. Nine healthy volunteers, one stroke survivor with right-sided hemiparesis and one patient with amyotrophic lateral sclerosis (ALS) participated in this study. Subjects did not receive BCI training before participating in this study. We investigated tasks of both physical movement and motor imagery. The surface Laplacian derivation was used for enhancing EEG spatial resolution. A model-free threshold setting method was used for the classification of motor intentions. The performance of the proposed BCI was validated by an online sequential binary-cursor-control game for two-dimensional cursor movement. Event-related desynchronization and synchronization were observed when subjects sustained or stopped either motor execution or motor imagery. Feature analysis showed that EEG beta band activity over sensorimotor area provided the largest discrimination. With simple model-free classification of beta band EEG activity from a single electrode (with surface Laplacian derivation), the online classifications of the EEG activity with motor execution/motor imagery were: >90%/∼80% for six healthy volunteers, >80%/∼80% for the stroke patient and ∼90%/∼80% for the ALS patient. The EEG activities of the other three healthy volunteers were not classifiable. The sensorimotor beta rhythm of EEG associated with human natural motor behavior can be used for a reliable and high performance BCI for both healthy subjects and patients with neurological disorders. Significance: The proposed new non-invasive BCI method highlights a practical BCI for clinical applications, where the user does not require extensive training.

Due to the relative noise and artifact insensitivity, steady-state visual evoked potential (SSVEP) has been used increasingly in the study of a brain–computer interface (BCI). However, SSVEP is still influenced by the same frequency component in the spontaneous EEG, and it is meaningful to find a parameter that can avoid or decrease this influence to improve the transfer rate and the accuracy of the SSVEP-based BCI. In this work, with wavelet analysis, a new parameter named stability coefficient (SC) was defined to measure the stability of a frequency, and then the electrode with the highest stability was selected as the signal electrode for further analysis. After that, the SC method and the traditional power spectrum (PS) method were used comparatively to recognize the stimulus frequency from an analogous BCI data constructed from a real SSVEP data, and the results showed that the SC method is better for a short time window data.

In models of electrical stimulation of the nervous system, the electric potential is typically calculated using the quasi-static approximation. The quasi-static approximation allows Maxwell's equations to be simplified by ignoring capacitive, inductive and wave propagation contributions to the potential. While this simplification has been validated for bioelectric sources, its application to rapid stimulation pulses, which contain more high-frequency power, may not be appropriate. We compared the potentials calculated using the quasi-static approximation with those calculated from the exact solution to the inhomogeneous Helmholtz equation. The mean absolute errors between the two potential calculations were limited to 5–13% for pulse widths commonly used for neural stimulation (25 µs-1 ms). We also quantified the excitation properties of extracellular point source stimulation of a myelinated nerve fiber model using potentials calculated from each method. Deviations between the strength–duration curves for potentials calculated using the quasi-static (σ = 0.105 S m⁻¹) and Helmholtz approaches ranged from 3 to 16%, with the minimal error occurring for 100 µs pulses. Differences in the threshold–distance curves for the two calculations ranged from 0 to 9%, for the same value of quasi-static conductivity. A sensitivity analysis of the material parameters revealed that the potential was much more strongly dependent on the conductivity than on the permittivity. These results indicate that for commonly used stimulus pulse parameters, the exact solution for the potential can be approximated by quasi-static simplifications only for appropriate values of conductivity.

An accurate description of the electrode–electrolyte interfacial impedance is critical to the development of computational models of neural recording and stimulation that aim to improve understanding of neuro–electric interfaces and to expedite electrode design. This work examines the effect that the electrode–electrolyte interfacial impedance has upon the solutions generated from time-harmonic finite-element models of cone- and disk-shaped platinum microelectrodes submerged in physiological saline. A thin-layer approximation is utilized to incorporate a platinum–saline interfacial impedance into the finite-element models. This approximation is easy to implement and is not computationally costly. Using an iterative nonlinear solver, solutions were obtained for systems in which the electrode was driven at ac potentials with amplitudes from 10 mV to 500 mV and frequencies from 100 Hz to 100 kHz. The results of these simulations indicate that, under certain conditions, incorporation of the interface may strongly affect the solutions obtained. This effect, however, is dependent upon the amplitude of the driving potential and, to a lesser extent, its frequency. The solutions are most strongly affected at low amplitudes where the impedance of the interface is large. Here, the current density distribution that is calculated from models incorporating the interface is much more uniform than the current density distribution generated by models that neglect the interface. At higher potential amplitudes, however, the impedance of the interface decreases, and its effect on the solutions obtained is attenuated.

Organic field effect transistors can be integrated into micromachined polyimide-based neural stimulation electrode arrays in order to build active switching matrices. With this approach, a matrix of N × M electrode contacts requires only N + M interconnects to a stimulator when active switching elements are used instead of N × M interconnects. In this paper, we demonstrated that pentacene-based organic field effect transistors (OFETs) can be used to drive stimulation currents through neural electrodes in a physiological-like environment. In order to prove the general applicability as an implant material, the cytotoxicity of pentacene was evaluated with respect to potential effects on cell viability. The results of these tests indicate that extracts from pentacene inhibit neither proliferation nor metabolism of the tested mouse fibroblasts. However, some effect on cell spreading was observed when cells were in direct contact to pentacene for 48 h. In pilot experiments it was demonstrated for the very first time that pentacene transistors can be used as switching elements, acting as voltage-controlled current sources, capable of driving currents suitable for electrical stimulation of a peripheral nerve via a tripolar cuff electrode.

We show here that a brain–computer interface (BCI) using electrocorticographic activity (ECoG) and imagined or overt motor tasks enables humans to control a computer cursor in two dimensions. Over a brief training period of 12–36 min, each of five human subjects acquired substantial control of particular ECoG features recorded from several locations over the same hemisphere, and achieved average success rates of 53–73% in a two-dimensional four-target center-out task in which chance accuracy was 25%. Our results support the expectation that ECoG-based BCIs can combine high performance with technical and clinical practicality, and also indicate promising directions for further research.

The performance of five non-parametric, univariate seizure detection schemes (embedding delay, Hurst scale, wavelet scale, nonlinear autocorrelation and variance energy) were evaluated as a function of the sampling rate of EEG recordings, the electrode types used for EEG acquisition, and the spatial location of the EEG electrodes in order to determine the applicability of the measures in real-time closed-loop seizure intervention. The criteria chosen for evaluating the performance were high statistical robustness (as determined through the sensitivity and the specificity of a given measure in detecting a seizure) and the lag in seizure detection with respect to the seizure onset time (as determined by visual inspection of the EEG signal by a trained epileptologist). An optimality index was designed to evaluate the overall performance of each measure. For the EEG data recorded with microwire electrode array at a sampling rate of 12 kHz, the wavelet scale measure exhibited better overall performance in terms of its ability to detect a seizure with high optimality index value and high statistics in terms of sensitivity and specificity.

Equation (3) in this paper contains an error. The correct equation is given below: