One central goal of current neuroscience is to accurately detect communities of physically and functionally coupled neurons in neuronal networks. Due to the reduced amplitude resolution of the network readouts, connectivity mapping is often restricted to statistical inferences from spontaneous activity, which are confounded by the context in which the activity is observed. Accurate connectivity mapping demands methodology advances and their evaluations, as well as probing experimental approaches other than observational. This work aimed to identify which information about the connectome could be reliably extracted from the current readouts and whether a single-neuron perturbation evokes network-level responses that mark the interconnected neurons. Chapter 4 addresses how accurately novel 2D+ electrodes convey information about the signal source and network activity. Paired electrophysiology revealed a pronounced variation of the same neuron’s extracellular waveforms, resulting in erroneous identifications of detected signal sources. While intracellular waveform variations contributed to the variability of 2D+ electrode signals, an additional contributing factor originated from inconsistencies in the signal transfer. The changes in the transfer function were reversible, occurred at a second timescale, and indicated a physically grounded contribution. Unlike the population-level metrics, pairwise correlation was particularly sensitive to misclassifications. Chapter 4 concludes with analytical challenges in high amplitude records and offers solutions at the spike-sorting curation step. Chapter 5 evaluates the consistency and context dependence of the calcium-firing code relation in population-level readouts. Benchmarking of the calcium signals by patch-clamp demonstrated that the calcium trace reliably predicts membrane depolarizing events. Accordingly, network synchrony events associated with strong depolarizations were the most reliably detected via calcium imaging. However, detecting isolated firing events was less consistent and depended on the magnitude of firing-evoking membrane depolarizations. The findings inform current firing prediction models about the single-cell short-term translation inconsistencies and position membrane depolarization predictions as the first step toward accurate firing activity inferences from calcium imaging studies. Chapter 6 investigated the feasibility of topology mapping via single-neuron perturbation. Calcium imaging during the giga-seal patch-clamp revealed that a single-neuron deformation induces calcium influxes in the target that persist through the neurites and propagate to coupled neurons. Calcium influxes were via mechanosensitive ion channels, were further potentiated by L-type calcium voltage-gated channels, and evoked slow membrane depolarizations capable of inducing firing in the target and surrounding neurons. Unlike the single-neuron electrical stimulation, giga-sealing produced robust calcium responses that propagated further in the network and excited the responding ensemble. The findings suggest that targeted deformation of neurons may be a promising approach for the broad mapping of neuronal communities and inform about a biological component to the spatial resolution reduction in deformations. Finally, the potentiation of the responding ensemble following giga-seal informs the optimal time window for reliable activity sampling using patch-clamp.
Bogdana Čepkenović (Wed,) studied this question.