The purpose of this study was to analyze differences in the activity of medullary respiratory neurons in the unanesthetized, intact cat during wakefulness and non-rapid-eye-movement (NREM) sleep. We studied single respiratory neurons located within a 1-2 mm deep, 8-10 mm long zone that followed, and included in its dorsal aspect, the retrofacial and ambiguus nuclei. The analysis of variance was used to detect respiratory activity, and cycle-triggered histograms were plotted. The respiratory signal strength and consistency of the respiratory activity were quantified with the eta 2 statistic. We determined for each breath in wakefulness and NREM sleep the average discharge rate during the active phase of the cell, the number of action potentials during the active phase of the cell, and durations of both the cycle and inspiration. Differences in discharge rates and in the number of discharges between wakefulness and NREM sleep were tested with the t test. A bimodal distribution of eta 2 values for the population of neurons indicated there were two groups of respiratory cells: those with eta 2 values less than 0.3 and those with values greater than 0.3. The former we call weak respiratory cells; the latter, strong respiratory cells. Strong and weak cells were classified further as inspiratory or noninspiratory on the basis of the shape of their cycle-triggered histograms. Within the class of strong inspiratory cells, those with the highest eta 2 values 1) reached their peak discharge rate early, 2) discharged at high rates throughout inspiration, and 3) were inactive during expiration. The values of these variables diminished progressively in inspiratory cell groups with lower eta 2 values. Most cells were less active in NREM sleep than in wakefulness. Similar proportions of weak and strong cells and inspiratory and noninspiratory cells were affected by sleep. The reduction in sleep of the activity of strong inspiratory cells was consistent with a general relationship between this activity and the duration of inspiration. Lower discharge rates were associated with longer breaths; higher rates with shorter breaths. This relationship existed within both NREM sleep and wakefulness, and the plot of the relationship across these states formed a continuous function. The reduction in discharge rate in sleep was greater for weak than for strong inspiratory cells: the correlation coefficient between percent change in rate and eta 2 values was -0.636 for inspiratory cells, but it was not significant (-0.265) for noninspiratory cells.(ABSTRACT TRUNCATED AT 400 WORDS)