Single pulse responses in cultured neuronal networks to describe connectivity

Synaptic connections between neurons play a crucial role in cognitive processes like learning and memory. In recent work we developed a method, using conditional firing probability (CFP analysis), to estimate functional connectivity in terms of strength and latency, and here we further explored on this method. CFP analysis estimates functional connectivity between pairs of neurons by calculating the probability that neuron j will fire at t=τ, given that neuron i fired at t=0. Because neuron i often fires more than once in the analysis interval of 500 ms CFPs don´t indicate the average response to a single action potential in neuron i. Rather, CFPs are biased by the probability that neuron i will fire again in the analysis interval (CFP_i,i, or autocorrelation). We developed a method to estimate single pulse responses (SPRs), by deconvolving CFP_i,i from the probability curve. We investigated the performance of this deconvolved measure in experiments with cholinergic network activation of cultured cortical networks. Ideally, acetylcholine should affect only the dynamic behavior of the system, but not the described (glutamatergic) connections in cortical networks. We found that changes in SPRs under different dynamic behavior were much smaller than those in CFPs. However, changes were still considerable, most likely reflecting the non-linear nature of synaptic transmission.