The best methods paper I've found on the matter is by a couple of folks clearly frustrated by the problem over at MIT: http://www.ncbi.nlm.nih.gov/pubmed/25023312 who also host the frequently-linked site http://www.brainslicemethods.com.
The short list of precautions are
*Rapid brain removal in ice-cold aCSF
*Vibratome slicing at slowest advance speed, high frequency, and medium amplitude (I've found a displacement of .75-1mm to be quite good), making sure to minimize Z-displacement
*Allowing recovery (though not necessarily slicing) for 10-12m in aCSF that replaces NaCl with NMDG a la the linked Ting et al. paper
*Preparing all solutions with as little advance time as possible before procedure - day of is optimal, but day before typically works for me. Anything >3 days is basically a gamble/waste of time in my experience.
*Keeping anything that comes in contact with the tissue sterilized/clear of heavy metals as possible. You can use a detergent like alconox or citronox for this. Obviously metal tools during dissection are necessary, but keeping any glassware clean and using a teflon coated vibratome blade has been very fruitful for me.
*Providing a nutrient source in your aCSF, we use Na-Pyruvate, but this varies.
Above all, the thing I've found most guarantees healthy slices is being very fast from decapitation to recording - I work with amphibian cells which are relatively resilient to anoxia and temperature changes, but trying to read (whole-cell at least) anywhere after ~4-5hrs from decapitation is essentially impossible.
Providing a more detailed picture of your prep would help with troubleshooting! Let me know if you have any questions on the above, and if anyone has any corrections to my answer I'd love to hear them :)
edit: This is something I've investigated with much less rigor, but the increased calcium concentrations in aCSF that can sometimes come from disrupting pockets of Ca++ during dissection can cause unnecessary cell mortality. Keeping a constant flow of aCSF diminishes this, and I've also recently tried 'spiking' the dissection chamber with an EGTA-rich solution when things get especially hairy.