In the rhythmic brain stem slice preparation, spontaneous respiratory activity is generated endogenously and can be recorded as output activity from hypoglossal XII rootlets. Here we combine these recordings with measurements of the intrinsic optical signal (IOS) of cells in the regions of the periambigual region and nucleus hypoglossus of the rhythmic slice preparation. The IOS, which reflects changes of infrared light transmittance and scattering, has been previously employed as an indirect sensor for activity-related changes in cell metabolism. The IOS is believed to be primarily caused by cell volume changes, but it has also been associated with other morphological changes such as dendritic beading during prolonged neuronal excitation or mitochondrial swelling. An increase of the extracellular K(+) concentration from 3 to 9 mM, as well as superfusion with hypotonic solution induced a marked increase of the IOS, whereas a decrease in extracellular K(+) or superfusion with hypertonic solution had the opposite effect. During tissue anoxia, elicited by superfusion of N(2)-gassed solution, the biphasic response of the respiratory activity was accompanied by a continuous rise in the IOS. On reoxygenation, the IOS returned to control levels. Cells located at the surface of the slice were observed to swell during periods of anoxia. The region of the nucleus hypoglossus exhibited faster and larger IOS changes than the periambigual region, which presumably reflects differences in sensitivities of these neurons to metabolic stress. To analyze the components of the hypoxic IOS response, we investigated the IOS after application of neurotransmitters known to be released in increasing amounts during hypoxia. Indeed, glutamate application induced an IOS increase, whereas adenosine slightly reduced the IOS. The IOS response to hypoxia was diminished after application of glutamate uptake blockers, indicating that glutamate contributes to the hypoxic IOS. Blockade of the Na(+)/K(+)-ATPase by ouabain did not provoke a hypoxia-like IOS change. The influences of K(ATP) channels were analyzed, because they contribute significantly to the modulation of neuronal excitability during hypoxia. IOS responses obtained during manipulation of K(ATP) channel activity could be explained only by implicating mitochondrial volume changes mediated by mitochondrial K(ATP) channels. In conclusion, the hypoxic IOS response can be interpreted as a result of cell and mitochondrial swelling. Cell swelling can be attributed to hypoxic release of neurotransmitters and neuromodulators and to inhibition of Na(+)/K(+)-pump activity.