Model-based design of stimulus trains for selective microstimulation of targeted neuronal populations

Selective activation of targeted neuronal populations is required for central nervous system (CNS) neuroprosthetic device efficacy. However in many regions of the CNS, cells and fibers of passages are intermingled. The goal of this project was to design stimulus trains that would enhance selectivity between microstimulation of cells and fibers of passage. Detailed computer-based models were developed that accurately reproduced the dynamic firing properties of mammalian neurons. The neuron models were coupled to a three-dimensional finite element model of the spinal cord that solved for the potentials generated in the tissue medium by an extracellular electrode. The results demonstrate that alterations in the stimulus frequency, based on differences in the post-action-potential recovery cycles of cells and axons, enabled differential activation of cells or fibers of passage. The results also show that asymmetrical charge-balanced biphasic stimulus waveforms, designed to exploit the non-linear conductance properties of the neural elements, can be used in combination with the appropriate stimulus frequency to further enhance selectivity. These outcomes provide useful tools for selective stimulation of the CNS, as well as basis for understanding frequency-dependent outputs during CNS stimulation.