The incidence of sudden death is higher in epileptic people compared to the general population and sudden unexpected death in epilepsy (SUDEP) is second only to stroke in the years of potential life loss among the major neurological disorders. In the majority of observed human SUDEP cases, respiratory dysfunction post-seizure is shown to be the primary initiating event leading to cardiac asystole and death. During seizures, several neuroactive agents are shown to be released, including serotonin and adenosine. Previous research has shown the effects of these neuroactive agents on seizure and respiratory function independently. A role of adenosine in triggering death post-seizures in a chemically-induced seizure model has been shown, but the mechanism of death is not clear. Studies from our lab have shown the role of fluoxetine (selective serotonin-reuptake inhibitor) in preventing seizure-induced respiratory arrest (S-IRA) in DBA/1 mouse model of SUDEP, but the neuronal networks mediating S-IRA and the brain structures involved in the fluoxetine-mediated blockade of S-IRA are not known. Data from human SUDEP imaging has underlined the role of periaqueductal gray (PAG), which is also implicated in audiogenic seizure (AGSz) network and respiratory modulation in other models. The goal of my dissertation is to understand the mechanisms by which adenosine could cause SUDEP susceptibility, the neuronal networks in the DBA/1 mice that lead to S-IRA and how fluoxetine modulates the neuronal activity at these neuronal network structures to prevent S-IRA. A better understanding of these mechanisms may lead to development of potentially important targeted therapies to prevent SUDEP in future. In the first aim, I have examined the role of adenosine in mediating SUDEP. Genetically epilepsy prone rats (GEPR-9s) exhibit AGSz but the incidence of death post-seizure is very low. I tested whether decreasing adenosine breakdown could increase the incidence of death in GEPR-9s. My study shows that adenosine metabolic blockers, which prevent the metabolism of released adenosine during seizures significantly increased the duration of respiratory dysfunction, post-ictal depression, decreased the peripheral oxygen saturation and subsequently, increased the incidence of death post-seizure in GEPR-9s. These findings on the role of adenosine and role of specific adenosine receptors in SUDEP are required to be validated in another SUDEP model. This formed the core of my second specific aim and since DBA/2 mice are susceptible to AGSz, and after seizures a large percent of these DBA/2 mice show S-IRA, while the rest don’t show S-IRA. Therefore, I tested if adenosine antagonism could prevent S-IRA post AGSz in DBA/2 mice, and found that caffeine a non-selective adenosine antagonist significantly decreased the incidence of S-IRA post AGSz. Administration of adenosine metabolic blockers increased the incidence of S-IRA in DBA/2 mice similar to GEPR-9s. Parallel studies from our lab have shown that administration of selective A2a antagonist but not A1 antagonist also decreased S-IRA incidence in DBA/2 mice. These data from GEPR-9s and DBA/2 mice suggests for a potentially important role of selective adenosine receptors in mediating the susceptibility to SUDEP by acting on respiratory function. In the third specific aim, I have examined the role of subcortical neuronal network structures including the PAG in mediating S-IRA and the quantitative differences in respiratory function elicited by electrical stimulation at PAG between DBA/1 and C57 mice. While the role of neuroactive agents in SUDEP has received attention, the neuronal networks mediating SUDEP in pre-clinical models are not known, specifically in DBA/1 mice an established SUDEP model susceptible to AGSz. The role of subcortical neuronal network structures including PAG in AGSz has been well-studied in other AGSz models. To decipher the neuronal networks that lead to S-IRA in DBA/1 mice, I exposed both DBA/1 mice that show AGSz and S-IRA and C57 mice that are non-susceptible to AGSz to acoustic stimulus and performed an ex vivo manganese-enhanced magnetic resonance imaging (MEMRI). Data analyses revealed the role of several brain structures in auditory, sensorimotor-limbic, respiratory networks and serotonergic raphe nuclei in DBA/1 mice. Of interest the PAG, a region implicated in other models of AGSz, respiratory modulation and human SUDEP has shown a significant increase in MEMRI signal intensity compared to C57 mice. These findings formed the rationale for the fourth specific aim to examine the quantitative differences in PAG-mediated respiratory modulation in response to electrical stimulation between C57 and DBA/1 mice. The threshold of current needed at PAG for a significant increase in respiration in DBA/1 mice is four times greater than C57 mice. Electrical stimulation at amygdala (AMG) showed marginal differences between DBA/1 and C57 mice suggesting the least possible pathological role of AMG in DBA/1 mice to mediate S-IRA. These data support a reduced respiratory function of PAG in DBA/1 mice compared to C57 mice. Taken together, these findings suggest that a reduced respiratory function of PAG in DBA/1 mice could lead to S-IRA and support a potentially critical compensatory role of PAG in DBA/1 mice. In the fifth specific aim, I examined the effect of fluoxetine on the subcortical neuronal network structures in DBA/1 mice that may lead to blockade of S-IRA. Fluoxetine has been shown to prevent S-IRA in DBA/1 mice effectively, but where in the brain does this drug act to prevent the susceptibility to SUDEP in DBA/1 mice is not known. To address this question, I used ex vivo MEMRI in DBA/1 mice that received fluoxetine at a dose which selectively blocks S-IRA but not AGSz. Fluoxetine treated DBA/1 mice that didn’t show S-IRA have shown a potential compensatory increase in activity at several sub-cortical structures including PAG compared to DBA/1 mice that showed S-IRA. In summary, these data suggest the PAG as a critical compensatory structure among the other sub-cortical neuronal network structures identified for SUDEP in this mice model. Differential modulation of these subcortical neuronal network structures by adenosine or serotonin released during seizures could determine the susceptibility to SUDEP.