The objective of this project is to explore whether a neuro-mechanical synapse can be formed by interfacing a neuron cell with a silicon substrate. It is based on our recent finding that mechanical force applied at a neuromuscular synapse of Drosophila (fruit fly) embryos produces neuronal memory. This mechano-sensing ability of neurons is likely rooted in evolutionarily conserved properties of cells. It has thus the potential to apply to neuro-silico interfaces. The approach to this goal are: (1) investigate whether mechanically applied tension on a neuron using a silicon probe can be transduced into information that can be captured, e.g., as an image; (2) functionalize the probe to explore whether synapse can be engineered, and (3) explore whether engineered synapse can attain usage dependent force sensitive memory.
Intellectual merit: The study will shed light, for the first time, on whether (and how) nature employs mechanical tension to store and process information in an analogue fashion. It then attempts to translate the knowledge to engineer a neuro-mechanical synapse. If this possibility is realized, it will be a breakthrough in the engineering of learning and memory into synthetic systems.
Broader impact: This study will lay the foundation for creating networks of synthetic cells that could learn and remember patterns and share that knowledge with devices and sensors. The knowledge will be integrated with education through (1) development of a new course on neuro-mechanics (2) student seminars, (3) hands on teaching modules at the local Children?s Science Museum, and (4) web page development.
The goal of this project was to explore whether anlogue memory and logic elements can be built from a hybrid of live neurons and silicon devices. The concept is based on our finding that neurons are under tension in vivo. We hypothesized that any perturbation in the tension will be remembered by the neuron. Such memory could be read out by optical imaging. The elementary memory array might be formed from a group of neurons and mechanical cantilevers that transduce external stimulation to a change of tension in the neurons. Basics: Memory and learning in animals is mediated by neurotransmission at the synaptic junctions (end point of axons). A large majority of neurons (single neuron cell) have a long (100µm to meters) axon that forms junctions (synapse) with muscle tissue or another neuron. They carry neurotransmitters enclosed within vesicles, about 50 nm in size. The density of the vesicles is high at the neuro-muscular or neuron-neuron synapse. In response to a stimulus (e.g., mechanical, chemical), the neuron generates an action potential (a spatial voltage wave) that propagates through the axon, arrives at the junction and triggers the release of the neurotransmitters through the exocytosis of the vesicles. The more a synapse is used, the higher is the amount of vesicle release in response to the same input stimulus, i.e., higher is the synaptic plasticity. This usage dependent synaptic strength is believed to be the basis for memory and learning. A central dogma in neuroscience is that, clustering is the result of a complex biochemical signaling process. Intellectual merit: We showed, using embryonic Drosophila (fruit fly), that mechanical tension in axons is essential for vesicle clustering at the synapse. Embryonic axons maintain a rest tension of about 1 nN. Without this tension, clustering disappears, but reappears with the application of tension. Increase of tension mechanically results in increased clustering. We read out the increased accumulation by florescently labeling neurotransmitter vesicles and observing their dynamics using an optical microscope. Under sustained stretch of about 10%, the accumulation increases by more than 30% in about 20 mins. The increased accumulation retains for more than 90 mins after the stretch is released. Thus the neuron remembers its experience of increased stretch or tension that we can read out. We then asked, (1) how do neurons accumulate vesicles at the synapse, i.e., what is the nature of vesicle transport and how do we quantify the different modes of transport; and (2) How does the transport change with neuronal forces or stretch? We found: (a) vesicles carrying peptides and neurotransmitters in neurons employ two modes of transport: directed motion along microtubules carried by motor proteins, and Brownian motion when the vesicles get off the microtubular tracks and move in the plasma due to thermal energy. We quantified the frictional barriers during the Brownian motion as well as during directed motion. We find that the two friction barriers are significantly different; (b) With increased force or stretch on the axon, the vesicles spend more time in directed motion, i.e., the probability of directed motion increases, and the frictional barrier against motion of the vesicles decreases. This implies, with increased tension, the overall mobility of the vesicles increases which might be the mechanism by which more vesicles accumulate at the synapse. Once at the synapse, the vesicle are trapped by actin network and the accumulation sustains even the tension is released. With decrease of force, the change in the probability of directed motion is small. Broader impact: The project has laid the very foundation of developing neuromechanical memory elements, although, significant additional research is needed towards such biohybrid computers of the future. In the face of tremendous advances in computer power and information technology, we still find that computers are vastly inferior to human and other animal nervous systems in the learning and recognition of complex patterns, for example recognition of faces. This is in spite of the fact that computer logical elements are much faster than neurons. A key difference between neurons and digital computer logical elements is the logical richness of neurons. This study has laid the foundation for creating networks of synthetic cells that could learn and remember patterns and share that knowledge with devices and sensors, and/or with other networks of cells. In addition, the project has trained several graduate and undergraduate students at the interface of cell biology and engneering. The research has been dessimated to K-12 students and public at large through childrens’ museum, farmers’ market, engineering open house displays, and visits to elementary to high schools with demonstrations, as well as conference and workshop presentations and journal publications.
