Current activities in brain neuroscience research focus on cognitive processes, mapping of neuronal activity in ever-larger segments, in-silico simulations and experiments. Given the number of neurons and the complex neural interconnection web that exist in the human brain, it is unlikely that an artificial brain of significant size can be achieved using approaches based on semiconductor devices.
Can we actually build a section of an artificial brain that has the “look-and-feel” of the biological organ and act like it? Can an artificial brain be implanted in the body and connected to the rest of the “system”? This is clearly a grand challenge to duplicate the enormous web of functions in the most complex organ developed by nature.
We will first define and discuss main aspects that need to be demonstrated materials properties, electrical properties, energy consumption, interconnections, functionality, testability and integration. Then we introduce a possible concept for a “real artificial” brain that has some tentative answers for each set of requirements. This “semi-soft” approach it combines some elements that are organic, but not necessarily biological, with cellular components.
Our approach starts with the observation of similarities between biological neuronal arrays and polymer fibers in membranes formed by electrospinning. Typical fiber diameter range (from nm to µm) and length (from sub-mm to meters) is consistent to that of axons. The number of fiber cross-connects in a typical membrane are ~ 1011-1012/cm3, similar to brain synapse density.
The presentation will review the work to date of several groups. At Cincinnati, we have fabricated polymer-based carbon nanofiber (CNF) electrospun membranes with electrical resistivity range of ~0.1 to 1000 W-cm. This range covers typical axon resistance of ~100 W-cm. In addition, we have fabricated core-sheath fibers by coaxial electrospinning that contain an inner conducting medium shielded by an insulating cover, similar to myelin sheath covered axons. Among other reports, Nielsen/Aarhus University has demonstrated that coaxial fiber membranes enhanced the growth of neuronal cells, indicating not just biocompatibility but ability to sustain long term viability. Jakobson and Ottosun (Lund) have demonstrated a unique electrospun fiber membrane with an uncompressed 3D structure that mimics the extracellular matrix of real brain tissue and can serve as scaffold for neuronal cell growth. Lee/Postech has demonstrated that artificial organic synapses fabricated on core-sheath electrospun nanofibers can operate with femtojoule energy consumption.
There is a whole universe of issues associated with the possible integration of an artificial brain section into a living brain in order to repair malfunctioning aspects. Nonetheless, the potential benefit of this concept is of such magnitude that this long and arduous journey is worth embarking on.