The aim of this study was to determine if episodic hypoxia evokes persistent increases of genioglossus muscle (GG) activity, termed long-term facilitation (LTF), in neonatal rats in vivo. Experiments were performed on anaesthetized, spontaneously breathing, intubated neonatal rats (postnatal days (P) 3–7), divided into three groups. The first group (n = 8) was subjected to three 5-min periods of hypoxia (5% O2–95% N2) alternating with 5 min periods of room air. The second group (n = 8) was exposed to 15 min of continuous hypoxia. The third (n = 4) group was not exposed to hypoxia and served as a control. GG EMG activity and airflow were recorded before, during and for 60 min after episodic and continuous hypoxic exposure. During hypoxia, GG EMG burst amplitude and tidal volume (VT) significantly increased compared to baseline levels (episodic protocol: mean ±s.e.m.; 324 ± 59% of control and 0.13 ± 0.007 versus 0.09 ± 0.005 ml, respectively; continuous protocol: 259 ± 30% of control and 0.16 ± 0.005 versus 0.09 ± 0.007 ml, respectively; P < 0.05). After the episodic protocol, GG EMG burst amplitude transiently returned to baseline; over the next 60 min, burst amplitude progressively increased to levels significantly greater than baseline (238 ± 40% at 60 min; P < 0.05), without any significant increase in VT and respiratory frequency (P> 0.05). After the continuous protocol, there was no lasting increase in GG EMG burst amplitude. We conclude that LTF of upper airway muscles is an adaptive respiratory behaviour present from birth.