[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] The pyloric network of the crustacean stomatogastric ganglion (STG) is a central pattern generator that requires descending modulation for normal ongoing rhythmic activity. However, the pyloric rhythm is capable of functional recovery after removal of descending inputs. We used the STG to determine whether or not correlated mRNA ion channels are dependent on neuromodulation. Our hypothesis is that relationships between ion channels are dependent on neuromodulation, not activity. To investigate this, we first measured mRNA expression levels of three calcium channels (Ca1A, Ca1D and T-type-related channel) and two potassium channels (shal and shab), of PD cells to investigate how channel transcription may be modified to influence recovery of burst activity. We collected single PD neurons from both recovered and time-matched control preparations and quantified channel transcript levels with quantitative real-time RT-PCR. There was widespread correlation between all three calcium channels and the two potassium channels in PD cells from intact networks. Specifically, the strongest relationships were between all three calcium channels and the shal channel, which carries an A-type transient potassium current (p[less-than]0.005; R2[greater-than]0.5). Furthermore, our results show that following recovery, there are no significant changes in overall mRNA abundance across all channel types. However, there was a striking lack of any correlation between measured channel types in PD cells following recovery. These results indicate that recovered, decentralized networks do not regain rhythmicity simply by increasing or decreasing mRNA expression for a given channel or channels. In order to determine whether ion channel correlations are dependent on neuromodulation or activity, we decoupled neuromodulatory and activity inputs. We found that preparations with neuromodulatory inputs maintained relationships between mRNA channels while activity input alone did not. Further, addition of pilocarpine, the muscarinic agonist and modulator, to decentralized preparations maintained the same correlations as those found in preparations that only had neuromodulatory input. To determine whether loss of correlations affected network function, we compared the pyloric burst frequency of the different conditions. We found that the pyloric burst frequency decreased under conditions that lost correlations between ion channels due to the removal of neuromodulation. Together, these results indicate that neuromodulation maintains ion channel correlations, which are important to proper network function. They also suggest a possible novel role of neuromodulation in the regulation of gene expression.