MULTIELECTRODE & IMAGING ANALYSIS OF CULTURED NETWORKS

目前，在动物体内行为学研究方面存在较大差距

At present, there is a large gap between in vivo behavioral studies of learning and memory, and in vitro studies of the cellular mechanisms of synaptic plasticity. A neuroscience research tool will be created that bridges this gap by providing a network of cultured cortical neurons with a computer-simulated body, and a virtual reality in which to behave. This new paradigm for studying learning in vitro is enabled by recent advances in computing power and multi-electrode array substrates. A long-term, 2-way interface between a computer and a cultured neural network will be created. Software tools that recognize emergent patterns of network activity will be developed, and used to trigger distributed patterns of electrical stimulation in real time. The effects of this sensory-motor feedback loop will be studied at the millisecond time scale by optical recording using our custom high-speed CCD camera. Changes in neuronal connectivity and morphology will be followed on the scale of minutes, hours and days using our 2-photon laser scanning microscope. Two-photon microscopy allows extended high- resolution imaging of living cells without harming them or bleaching the fluorescent label. It is clear that neural systems process and store information in a distributed fashion. Single-unit neurophysiology research is likely to miss many of the emergent properties of distributed information processing. By combining many-unit electrophysiology and optical recording with non-destructive 2-photon imaging in an in vitro system capable of behaving and learning, it will be possible to observe changes in subcellular, cellular, and network properties that underlie learning and memory. By studying the mechanisms of information processing and storage in small networks of neurons, models of mental impairment from disease or aging can be examined with unprecedented detail. Information about how living neural networks function will promote the creation of more human-like computing systems.