Discussion: P3 targeting/2025-07-15

[severine2305]

Hi everyone! As I mentioned this am, I would like to start a conversation about the use of a 4 shank NP probe in V1. Because it is not a plug and play thing, I would like to have a summary of your opinions before starting changing things in the rig.

My only cons are "need some more work" and also "4 insertions of a 4 shanks in the same hole, same angle, can make damage and is very hard maybe impossible to annotate".

I attach again my presentation here.

ImplantHistoryP3Summary.pptx

[maierav]

I would be very curious what others think.

My opinion is that the cost/benefit is worth it. In fact, it would elevate the data to an entirely new level, which may prove fruitful for the entire field (including future Allen/OpenScope experiments).

4 shanks allow for the investigation of spatial activation patterns that we are blind to otherwise, and V1 seems like the ideal target for that. This can provide a bridge to some of the other data of this project, such as SLAP2. Specifically, 4 shanks allows for CSD - a measure of net synaptic depolarization - to be computed in 3D rather than 2D. The 3D analysis matches the SLAP2 approach, where we image several 2D planes at once, and can answer questions about the role of extracellular electric fields impacting spiking probability (not to get too much into the weeds, but for calculating in vivo field strengths, such as those relevant for ephaptic coupling, it takes 3D CSD - but that has not yet been done in mouse nor primate).

The damage is less of a concern to me. Before Neuropixels, we routinely used much larger diameter microelectrodes at the exact same location. And to our own astonishment, we could not find permanent lesions in subsequent histology. This is why electrolytic lesions became a staple in the field. Without them, electrode tracks could often not be found.

The annotation is also less of a concern to me since much of that information can be recovered in data analysis. Here is a study showing how electrode geometry can be reconstructed from the data (LFP coherence) using multi-dimensional scaling: https://pubmed.ncbi.nlm.nih.gov/12929926/
Briefly, LFP similarity falls off systematically as a function of cortical distance, so arranging electrodes by LFP similarity reveals their original geometry during the experiment.

Would the extra work be worth it if you also gain these benefits for future experiments using these rigs?

[Sarruedi]

Hi Severine,

It would be great to know the exact channel configurations that are possible with the 4-shank probes to decide how to proceed. Does pinpoint show the area targeting taking this into account or simply reflect the anatomical track of the probes?

I haven't used 4-shank probes yet and it would be useful to better understand how we could leverage 4-shanks in the best way possible. I'm not entirely convinced that 4-shanks in V1 is the way to go as it really comes down to the channel configurations that are possible.

[severine2305]

Hi Sarah, here are the configurations possible with the 4 shanks probes. I think pinpoint is stuck on configuration 1 in this probe (you see the channels in red in the coronal slide), but not 100% sure. It's this configuration 1 that I wanted to use.

[maierav]

Configuration 1 might yield the most interesting data, so that is good.