# Point process model for Precisely Timed Spike Trains

At COSYNE 2009, I presented my work on point process modeling as a poster (abstract doi). I proposed a new counting process model for precisely timed events which are often observed in spike trains when identical stimulation is given. My proposed model is very intuitive and simple, since it models the absolute time of the event themselves as precisely time action potentials generally are found as such. This is because most of the time the preciseness comes from the external world, such as a sharp change in input current or certain input pattern or initiation of movement and so on. It is usually not from the neuron’s intrinsic dynamic (oscillatory) properties. For example, Schreiber et al. showed that frozen noise reliably generates presicely timed spike trains compared to constant current injection in in vitro. My approach was to directly represent the precisely timed events through the precision (jitter distribution of the timing), and the reliability (probability of the event to show up) and add up the counting processes corresponding to each event. Currently, I am working on a paper version of this work, so I’ll post an update about this work soon.