Dear Amila,

I usually choose soma, primary dendrites (close to soma-until 3rd branching order), and distal dendrites. Choosing dendrites may also depend on the cell type and their specific inputs: such as dorsal, ventral, medial  and lateral dendrites may be considered seperately if they receive specific and differential inputs. Since you are studying in primary cultures, then for you spine measurements in the soma, primary and distal dendrites should be sufficient. The density and shape of spines can vary with their location such as proximal vs distal dendrites, and therefore you need to be consistent where you select to measure/count spines. Please see examples in my following papers.

Best wishes,

Refik Kanjhan

Article Developmental changes in the morphology of mouse hypoglossal...

Article Emerging Roles of Filopodia and Dendritic Spines in Motoneur...

If you are looking at spines then you want it to be random. Decide how many neurons you want (ie.10) per slide and capture those images at random. If you are choosing them then you are biasing your data. Hope this helps.

based on experience from Golgi staining, spin density in basal dendrites is increasing until 50um away from soma, and then keep steady at ~0.6-1.2/um (depend on cell type) after that point.

You should also make a correction for spines occluded by dendrite thickness.  This correction has been published years ago.  Apologies for forgetting the specific reference.