Dear Peter Lund Ovesen

You should better try for specific nervous tissue stains (Cresyl violente, methilene blue, Golgi, Cajal, ... ). 

You can find out at: 

http://www.histology-world.com/stains/stains.htm

Hi Peter,

use NeuN antibody, or indeed Cresyl violet (Nissl staining). The cortical layers are more evident then. H&E is mostly used for analysis of non-neuronal tissue. 

It works well with cresyl violet a simple histochemical staining method which gives good results and allows to distinguish easily between the different layers.

Please note the above comments on staining. They will help.

You did not say if you prefixed (or post fixed) so:  fixation is important  too.  It will improve your nuclear staining, which will help you identify the different cell types. A quick fix which we found to work quite well is: 40 cc methanol, 40 cc acetone, 10 cc  10 percent formalin. Fix chilled for 30 sec. to a min. This will bring out the nuclear staining with all the above stains. Rinse and then go on to stain. Thickness can be a problem some times with the cns. Because you often want to see the whole neuron but for cytoarchitecture you are in the range.  But you can compare your section thickness to those of the current Stereotaxic Atlases on the Rat brain. If you are not seeing what you want, you can adjust your thickness. Good luck

Dear Peter, in what species are you working?

In mouse, for example, identifying cortical layer boundaries is quite subjective and may depend significantly on what markers / stains / illumination methods are used. The Allen Institute's database (or similar resources) can be quite helpful for a gene/antibody-based approach. For a paper demonstrating multiple complementary techniques of laminar identification, please see our recent work (attached).