I'm doing whole-cell recordings (VC and CC) from cerebellar granule cells while trying to evoke antidromic action potentials with a bipolar stimulating electrode placed in the molecular layer. I find that the spike amplitude decreases over the course of 10-15 minutes which is too short a time frame for me to carry out my experiments. I don't think that the neuron is going downhill on me as the holding current and input resistances are holding steady. Has anybody else seen this and what is the most likely cause? Thanks.