From selective stimulation to single-spike detection: interfacing neocortical neurons with thousands of microelectrodes at subcellular spatial resolution

Abstract

Complex operations performed by neuronal networks arise from the orchestrated activities of individual neurons. Therefore, an experimental ability to simultaneously observe and elicit activity in individual neurons over extended periods of time is a prerequisite for exploring how neural circuits work. Recent advances in microelectronics and microfabrication technology have produced novel high-density microelectrode arrays (HD-MEAs) that enable measuring and stimulating in vitro neuronal activity across thousands of microelectrodes noninvasively. The arrays’ dense arrangements of microelectrodes provide hundreds of locations to comprehensively access the electrical activity of a single neuron. The microelectrodes can simultaneously observe and elicit activities in virtually any neuron within the network. However, selectively targeting a specific neuron and not its neighbors and extracting the small signals of a cell’s neurites from the background noise both pose difficult challenges that must be overcome. With this in mind, the first part of this thesis presents strategies for the electrical identification and selective stimulation of individual neurons within cortical networks. The second part describes a method for observing the small individual action potentials (APs), importantly without the need to average multiple events, emitted by the hundreds of micrometers long axonal arbors. The combination of these techniques enables precisely controlling and observing electrical activity across an entire neuron and, thereby, allows for influencing and monitoring how targeted neurons operate within the network. In the first project, we studied cultured neocortical neurons by using high-density microelectrode arrays and optical imaging, complemented by the patch-clamp technique, with the aim to correlate a neuron’s responsiveness to extracellular stimulation to the morphology and electrical characteristics of its various subcellular compartments. We developed strategies to electrically identify any neuron in the network, while subcellular-spatial-resolution recording of extracellular AP traces enabled their assignment to the axon initial segment (AIS), axonal arbor and proximal somatodendritic compartments. We proposed analytical approaches to reveal correlations between spatiotemporal features of extracellular APs and effective stimulation voltages, and explored the limits of targeted extracellular stimulation of different neuronal compartments. We found that stimulation at the AIS required low voltages and provided immediate, selective and reliable neuronal activation, whereas stimulation at the soma required high voltages and produced delayed and unreliable responses. Subthreshold stimulation at the soma depolarized the somatic membrane potential without eliciting APs. Approaches established in the first project were used to develop a method to non-invasively detect single action potentials throughout the whole axonal arbor of individual cortical neurons. The method was then used to measure (I) the precision of neuronal activation in response to extracellular stimulation; (II) the precision of AP propagation across multiple axonal branches; and (III) to observe changes in the precision and velocity of AP propagation during a high-frequency regime of neuronal activity. We found that (I) an increase in stimulation voltage magnitude decreases neuronal activation latency and, at the same time, increases temporal precision of the activation itself; (II) the precision of AP propagation decreases with the length of the propagation path; and (III) the high-frequency regime of neuronal activity decelerates AP propagation by 20.7 % and decreases the precision of AP propagation by 21.4 %.