The mammalian neuronal activity during sleep consists of two main states: non-rapid-eye-movement (NREM), also called slow-wave, and rapid-eye-movement (REM), also called active or paradoxical. The mechanisms underlying this ultradian sleep rhythm are not well understood but likely depend on reciprocal interactions between multiple populations of neurons in the brainstem. Studies have shown variability in sleep patterns across different species, complicating the identification of common mechanistic principles [Siegel, 2008]. Recent evidence in non-mammalian amniotes suggests that similar sleep rhythms might have a common ancestry. Two main hypotheses explain mammalian NREM-REM alternation [Weber, 2017], relying on interactions between neurons with antiphasic activity: through self-inhibitory connections resulting in limit-cycle solutions [Hobson et al, 1975], or through reciprocal circuits and an idealized "REM pressure" process that builds up and discharges regularly [Booth & Behn, 2014]. Although different in detail, these models agree on the need for reciprocally inhibitory circuits to generate sleep-state alternation and rely on neural connections or processes that have not been confirmed experimentally. Circuits with alternating outputs, often called central pattern generators (CPGs), are well known in motor systems [Marder, 2012]. To test the hypothesis that Pogona vitticeps' ultradian sleep rhythm is driven by a CPG, we exploited the regularity and high frequency of its biphasic sleep rhythm [Shein-Idelson, Ondracek et al, 2016; Fenk et al, 2023]. Using perturbation experiments and computational phase analysis, we probed the properties of Pogona's ultradian rhythm. Light pulses induced phase-dependent resets and entrainment, revealing a critical transition from phase delay to phase advance in the middle of NREM. The ultradian rhythm frequency could be modulated within limits by light pulse entrainment, with REM duration being shortened and NREM more flexibly dilated. In awake animals, alternating light/dark epochs induced a brief ultradian-like brain rhythm. In sleeping animals, a unilateral light pulse reset the ultradian rhythm in the opposite hemisphere, with spontaneous resynchronization within a few cycles. Together our findings support the hypothesis that CPGs may also organize the ultradian sleep rhythm in an amniote vertebrate. This insight could advance our understanding of the evolution, development, and mechanisms of sleep.