Our long-term goals are to better understand the response of neurons to artificial stimulation, and, to use this knowledge to develop new and more effective strategies for stimulating non- or improperly-functioning neurons of the CNS. The development of models that comprehensively and accurately predict the response of neural populations to electric stimulation has proven challenging, in part because of the significant morphological differences that can exist even between nearby cells, and, a lack of understanding as to how such differences shape each cell?s response to stimulation. A comprehensive understanding of the activation process would not only allow the development of models that would more accurately predict population responses but would also support the development of more effective stimulation strategies. In the retina for example, cells that respond to increases in luminance (ON cells) typically lie adjacent to cells that respond to luminance decreases (OFF cells); the two do not typically fire action potentials in response to the same stimulus and therefore, a prosthesis that activates both simultaneously creates a pattern of neural activity that is non-physiological. Mis- match between natural and artificial signals limits the quality of vision that can be obtained by a retinal prosthesis and similarly limits the effectiveness of other CNS-based prostheses as well. Here, we propose to comprehensively study how individual cellular properties each influence the response to artificial stimulation. Our approach will be to map sensitivity across a cell, and then compare physiological maps to cellular morphology, including the expression of voltage-gated ion channels; this will allow us to identify the specific cellular regions that have the strongest influence on responsivity. Computational models based on our precise anatomical measurements can be calibrated from the physiological maps to optimize the accuracy of the models; they will also help to unequivocally identify the relative sensitivity of individual features. Comparison of multiple cells within the same cell type will help to further identify the features that have the strongest influence on threshold and repeating the process across multiple cell types, different CNS regions and multiple species will lead to a comprehensive understanding of the activation process, along with the concurrent development of models that accurately predict the response of large populations of neurons to many different forms of stimulation. The inclusion of non-human primate tissue in the study will enhance the translation value of our findings. Validated models will be used to study responses to more advanced stimulating strategies, e.g. the high-rate stimulus trains that produce selective activation in ON vs. OFF cell types of the retina, and, the use of magnetic stimulation from implantable micro-coils to selectively target pyramidal neurons in the cortex while avoiding nearby passing axons from distal neurons. Models will be further enhanced from each new set of experiments and the comprehensive set (of models) will be made widely available to the research community.
Existing neural prostheses that target the CNS offer the potential to treat a wide array of neurological disorders. Progress with these devices has been limited, in part because of an inability to stimulate targeted neurons in an effective manner, e.g. in a way that re-creates the neural activity that arises naturally. Our goal here is to study the fundamental principles that govern how retinal and cortical neurons respond to artificial stimulation, and to incorporate this knowledge into a series of computational models that can be used to better evaluate the effectiveness of new, and potentially more effective, stimulation strategies.
