Significant bump/drop in spikes and count differences when moving in the negative contrast vs positive contrast direction

[ren-e1011]

I ran a toy stimulus where half is valued 255 and half at 125 on a uniformly-0 background such that there is one contrast edge and all of the spikes are at that contrast edge (or the borders)

I am using a wrapper for v2e.v2ecore.emulator EventEmulator which converts the frame to greyscale, calls EventEmulator.generate_events(), updates the time (0.02s between frames), and returns the new events

Non-default parameters I am passing to the EventEmulator are leak_rate_hz = 0.0, shot_noise_rate_hz=0.0, pos_thres=0.4, neg_thres=0.4, refractory_period_s=0.0

I moved the stimulus consistently “W” with the brighter leading edge or “E” with the darker leading edge –

Before preprocessing, the resulting spikes at each step look as expected with positive spikes at the leading edge, negative spikes at the leaving edge and spikes at the contrast edge in the center

Removing the outer border, the internal contrast edge holds all the spikes

However, there is a significant bump or drop in spike count when I move the stimulus “W” with the brighter leading edge vs “E” with the darker leading edge

Do you have an explanation for this jump in spikes with a consistent motion of a single edge at 0.0 refractory period
and why there might be a difference when the contrast edge is negative (“W”) or positive (“E”) when the positive threshold==negative threshold?