Title: Project 1 - Annotated Bibliography
In neural engineering, there currently exist many gaps between current technology available to the public, and where public necessity lies. Demand comes primarily from the clinical sector, wherein individuals suffering from neurodegenerative diseases are left at a deficit in their ability to sense and interact with the world around them. Challenges facing the technology mainly center around decoding the biochemical signals within the central nervous system, encoding information technology into biochemical signals, and creating devices that are able to function within or in tandem with biological systems. Neural engineering as a discipline attempts to close these gaps by taking neurobiological, materials science, biotechnological, and computer science research and designing devices and techniques to integrate technology with neurobiology. Much of the work takes place in pre-clinical and clinical research contexts, and the applied aspect of the discipline is still by-and-large in its infancy, as technology begins to catch up to what is demanded by the applications. Ultimately, the goal of the discipline is to create devices that work seamlessly with biological systems, are non-invasive or biologically integrated, and allow complete restoration of function without loss of fidelity.
Science Translational Medicine, 7(313), 313ra179. http://doi.org/10.1126/scitranslmed.aac7328, , (2015).Virtual typing by people with tetraplegia using a self-calibrating intracortical brain-computer interface.
The primary author of this paper, Beata Jarosiewicz, is a professor at Brown University's Department of Neuroscience, who has, in her career, contributed to or produced 18 publications, garnering nearly 800 citations thus far, and is renowned for her work using biotechnology and neuroscience to investigate ways to use and optimize brain-computer interfaces (BCIs). Jarosiewicz, et. al., sought to demonstrate that the nonstationarity that is convolved with neural signaling, that normally forces brain-computer interfaces to require frequent recalibration, can be mitigated using software to automatically account for such drift, in a point-and-click model used by tetrapalegic patients. The investigators used technological innovations, namely: tracking the statistics of neural activity (neurobiological observations), correction for velocity bias (computer engineering), and recalibration of the decoder using data inferred from the user's past actions (computer engineering) to develop software that extends the time that BCIs can function properly before needing to be manually recalibrated. The innovations made through this investigation can be applied to a wider range of devices, not merely point-and-click oriented BCIs. This will further enable the viable use of intracortical BCIs that patients needing restoration of independent communication and other assistive devices depend on, and lay the groundwork for more optimized methods of accounting for nonstationarities.
Scientific Reports, 5, 15890. http://doi.org/10.1038/srep15890, (2015). A Gaze Independent Brain-Computer Interface Based on Visual Stimulation through Closed Eyelids.
At the time of publication, Hwang was a postdoctoral researcher at the Machine Learning Group of the Berlin Institute of Technology (TU Berlin), and the supporting authors were members of the Neurotechnology Group of TU Berlin. These groups are highly prolific, contributing a dozen or more journal publications, conference papers, and book materials each year. Most Brain Computer Interfaces (BCIs) function using the visual system of patients, often requiring individuals to gaze at elements of a display to provide a signal to the computer. This study seeks to introduce a BCI model that does not require voluntary gaze, to help patients who are visually paralyzed or suffer a similar injury or deficit. The study takes the engineering process approach of finding a gap in current technology, in this case a lack of BCIs that are usable by patients with ocular dysfunction, and innovates a novel paradigm in which patients attend to one of three stimuli presented to their closed eyelids. Understanding neurobiological phenomenon associated with attention, and usnig electrical engineering to modify existing technology, the researchers were able to come up with a novel solution. The investigation demonstrated significantly effective use of the novel paradigm with patients for whom use of standard BCI devices are not possible or effective, and was the first to do so using this method. As a result, technology has been adapted to aid individuals who were previously unable to be helped.
Brain Sciences, 5(4), 419–440. http://doi.org/10.3390/brainsci5040419, (2015). A Real-Time Magnetoencephalography Brain-Computer Interface Using Interactive 3D Visualization and the Hadoop Ecosystem.
The primary team contributing to this work is based out of Northeastern University in conjunciton with the Lawrence Livermore National Laboratory. Northeastern is a highly prolific private research university that encourages groundbreaking technological advancement research, and the LLNL is widely recognized as being committed to seeking engineering solutions to nationally-scoped problems. Understanding that advances in technology have allowed for better handling of big data, the investigators have developed a BCI that utilizes Hadoop (a library of big data processing programs) to distinguish distinctive thought actions using magnetoencepholography (MEG) in a demonstration of flight visualization, where the user's thoughts of left or right are used to turn a virtual aircraft in that direction. By combining computer engineering and computer science methods with principles of neurobiology, the investigators have developed a system that can pick out specific thoughts using a non-invasive technique. The basis of this BCI technology can be used to translate user thought into specific actions like keyboard or mouse button presses, and the history of the user's input can be stored and analyzed using Hadoop to both access the information later for observation, and to ultimately perform comprehensive post-hoc analyses or calculations. These advances pave the way for further integration of big data computation and BCI technology.
Science, 347(6218), 159–163. http://doi.org/10.1126/science.1260318, (2015). Electronic dura mater for long-term multimodal neural interfaces.
The Center for Neuroprosthetics of the Ecole Polytechnique Fédérale de Lausanne (EPFL) is a Switzerland-based research organization with notable sponsors including the International Foundation for Research in Paraplegia and Medtronic. It is involved in an array of cutting-edge research endeavors with bases in neuroscience and biomedical engineering. Working from the challenge that stiff neural-integrated devices often don't interface well with soft biological systems over extended periods of time, the researchers developed a neural implant with properties similar to dura mater, the membrane that covers brain and spinal cord tissue. The resulting construct has been demonstrated to be durable with regards to stretch cycles, chemical injections, and electrical stimulation, allowing it to be viable in an array of different tissue areas and treatment/function scenarios. The research team has integrated neurobiological scientific knowledge with materials science and electrical engineering innovations to create implants that better integrate with biological tissue, improving on previous devices that were previously made of stiffer, less bio-friendly materials. The advances made by this group establish technological innovations that will allow others in the future to use similar techniques, materials, and methodologies to continue to develop solutions to the problems touched on by this work. The implants were demonstrated to function both to extract neural activity in behaving animals, as well as restoring locomotion to animals suffering from spinal cord paralysis with electrochemical inputs. Thus, continued study of these materials will allow for devices that can extract neural activity and provide supplemental inputs to paralyzed systems.
Clinical Neurophysiology, 125(8), 1556–1567. http://doi.org/10.1016/j.clinph.2013.11.034, , , (2014). Complexity and familiarity enhance single-trial detectability of imagined movements with electroencephalography.
The Brain and Mind Institute of Western University is a distinguished institution that focuses on an integrative approach to neuroscience, incorporating an array of different disciplines' perspectives to address challenges facing the field. The work in this publication centers on testing whether sensorimotor rhythms that result from motor imagery are more distinguishable via electroencepholography if the actions are more complex and are familiar to the user, such as requesting a pianist to imagine playing a complicated piece of music. Through study with healthy volunteer subjects, the research has found that EEG responses are indeed more reliably classified when motor imagery actions are complex and familiar. This research is based in a scientific method approach of testing a hypothesis based off of other conclusions in the field, and integrates cognitive neuroscience, EEG technology, and computational analytics to find ways to optimize reliable signal classification in EEG scenarios. As a result of this research, Brain-Computer Interfaces can serve more effectively when decoding the intentions of the user when the motor imagery used is more complex and/or are familiar to the user. This study will therefore improve the ability for the intentions of individuals who are unable to communicate normally to be interpreted by BCIs, and assist in the overall goal of the discipline to utilize less invasive technologies such as EEG and its eventual descendents.