Recording stability

[LibiF]

I have a few questions about handling some issues with recording stability:

1. How do you recommend treating clusters that are active only in specific periods of a recording?
   In this context, if a cluster appears to be centered on a specific channel, but then as a result of drift a similar (in terms of wave-forms) cluster appears centered on a nearby channel - do you recommend merging?
2. When we discussed ways to improve recording stability the possibility of covering cortex with agar was discussed. I am interested in whether this affects probe cleaning and how comfortable are you with using this and not harming the probe.
   If there are any other tips for improving recording stability I would love to know...
   Thanks

[nsteinme]

1. It depends on your scientific question, and also depends on how confident you feel that they are the same neuron, just shifted. If you see many neurons moving 'up' the probe at the same time, then maybe you feel more confident. If the waveform has some really distinctive feature, then maybe you feel more confident. I don't have a criterion to make a certain answer though :(
2. The only concern with agar is that sometimes it can dry on the shank and get stuck to it. If that happens, then as it continues to try it can bend or break the shank (this hasn't happened to me but it has to other people in the lab). So it's good to try to keep the agar moist. But in general this is a standard procedure for extracellular ephys. I can't promise you'll never break a probe with it though.

The other main stability trick is making a smaller craniotomy. But you can't ever solve the problem, I don't think, since it's mostly the brain itself that is moving. We're working on spike sorting updates that should help.