Strange 'chunks' in TraceView linked possibly to Nblocks and temp_wh.dat

[ross-folkard]

Hi there,
I'm using 64-channel single shank probes from CN, and sometimes in the phy visualisation of the traces I get these strange 'chunks' after spike-sorting with KS3 or KS2.5:

They do not obviously change when altering the ops.ntbuff, but do change when altering the Nblocks for registration. They are absent when Nblocks = 0, and the nature of the chunks changes when Nblocks changes:

Nblocks = 0

Nblocks = 1

Nblocks = 2

Nblocks = 5

I can't quite tell if this phenomenon is more frequent in periods of low spiking activity, or if it's just more noticeable then.

What also seems to matter is whether the dat_path is 'temp_wh.dat' or 'recording.dat'. The chunks only seem present in the 'temp_wh.dat' visualization.

'temp_wh.dat':

'recording.dat':

I have to say this is not something I've noticed in many recordings. Nevertheless I'm wondering if this may underlie some deeper issue that may affect the validity of the units that I get from recordings where this is not so obvious.

In general, I have 3 questions:

What do you think might be happening here?

How likely is it that this obvious chunking may affect the clustering say before and after these time periods?

Might it affect clusters in recordings where Nblocks is used during spike-sorting but the chunking is not as obvious or seemingly absent?

Any help would be greatly appreciated, and thank you in advance for your time!