Timing misalignment with 776270

Adding an issue to record what @maierav reported [here](https://github.com/AllenNeuralDynamics/openscope-community-predictive-processing/discussions/105)

*Context:*
For synchronization we run the SLAP2 microscope in ~20s intervals and then send a quick stop and start pulse and record the harp times for these. We can retrieve an estimate of the timing by getting the slap2 start and stop times and linearly interpolating along with the number of recorded slap2 measurements. So far, we had a bug in Bonsai where it unreliably sends/records an additional stop pulse (or 2) beyond what is recorded in our slap2 trials. To combat this we need to intelligently determine when to remove these initial stop trials pulses.

*The issue here:*
I was deleting one additional stop pulse, so the stop trial times were shortly before the start trial times, so when interpolating the timestamps it generated negative timestamps.

*The fix:*
For this session I had to add a parameter to manually slice off the unexpected trial records. @jeromelecoq expects this is fixed in Bonsai for future sessions. It may be the case that 803496 has a similar bug, I will need to verify the trial times are correct and repackage with this parameter if they are not.