FEATURE: DIY guide to improve the results of Kilosort

[marius10p]

Feature you'd like to see:

We need a DIY guide to help people identify problems in their spike sorting pipelines and take the right steps to fix them (i.e. change some parameters or add a custom preprocessing step). It's very difficult to anticipate everything that could go wrong with different types of data and experiments. But if you know what to look for, it's possible to identify issues and some of the issues might be solvable (computationally or experimentally).

Additional Context

No response