We examined the synaptic connections from upper cervical inspiratory neurons to phrenic and intercostal motoneurons in decerebrate rats using cross-correlation. Upper cervical inspiratory neurons (n = 79) were recorded from the C1 and C2 segments of the spinal cord in 38 vagotomized, paralyzed, ventilated, and decerebrate rats. The neurons were identified by their inspiratory firing pattern and antidromic activation from the ipsilateral spinal cord at C7. Whole-nerve recordings were made using bipolar electrodes from the central cut ends of the C5 phrenic nerve and the external and internal intercostal nerves at various thoracic levels. Cross-correlation histograms were computed between these recordings to detect short time-scale synchronizations indicative of synaptic connections. The 55 cross-correlation histograms computed between the upper cervical inspiratory neurons and the ipsilateral phrenic nerve showed seven (13%) narrow peaks (mean half-amplitude width +/- SD, 1.09 +/- 0.15 ms) at short latencies (mean latency +/- SD, 1.29 +/- 0.26 ms) suggestive of monosynaptic excitation, and four (7%) broader peaks (mean half-amplitude width +/- SD, 1.50 +/- 0.17 ms) at short latencies (mean latency +/- SD, 1.40 +/- 0.24 ms) suggestive of oligosynaptic excitation. Another 14 (25%) cross-correlation histograms displayed a central broad peak (mean half-amplitude width +/- SD, 1.59 +/- 0.23 ms) suggestive of common activation. The eight cross-correlation histograms computed between the upper cervical inspiratory neurons and the contralateral phrenic nerve were featureless. The 77 cross-correlation histograms computed between the upper cervical inspiratory neurons and the internal and external intercostal nerves at various thoracic levels (T2-8) showed no peaks suggestive of synaptic connections. We conclude that some upper cervical inspiratory neurons make monosynaptic and paucisynaptic connections to phrenic motoneurons but not to intercostal motoneurons.