The spinal circuitry underlying the generation of basic locomotor synergies has been described in substantial detail in lampreys and the cellular mechanisms have been identified. The initiation of locomotion, on the other hand, relies on supraspinal networks and the cellular mechanisms involved are only beginning to be understood. This review examines some of the findings relative to the neural mechanisms involved in the initiation of locomotion of lampreys. Locomotion can be elicited by sensory stimulation or by internal cues associated with fundamental needs of the animal such as food seeking, exploration, and mating. We have described mechanisms by which escape swimming is elicited in lampreys in response to mechanical skin stimulation. A rather simple neural connectivity is involved, including sensory and relay neurons, as well as the brainstem rhombencephalic reticulospinal cells, which act as command neurons. We have shown that reticulospinal cells have intrinsic membrane properties that allow them to transform a short duration sensory input into a long-lasting excitatory command that activates the spinal locomotor networks. These mechanisms constitute an important feature for the activation of escape swimming. Other sensory inputs can also elicit locomotion in lampreys. For instance, we have recently shown that olfactory signals evoke sustained depolarizations in reticulospinal neurons and chemical activation of the olfactory bulbs with local injections of glutamate induces fictive locomotion. The mechanisms by which internal cues initiate locomotion are less understood. Our research has focused on one particular locomotor center in the brainstem, the mesencephalic locomotor region (MLR). The MLR is believed to channel inputs from many brain regions to generate goal-directed locomotion. It activates reticulospinal cells to elicit locomotor output in a graded fashion contrary to escape locomotor bouts, which are all-or-none. MLR inputs to reticulospinal cells use both glutamatergic and cholinergic transmission; nicotinic receptors on reticulospinal cells are involved. MLR excitatory inputs to reticulospinal cells in the middle (MRRN) are larger than those in the posterior rhombencephalic reticular nucleus (PRRN). Moreover at low stimulation strength, reticulospinal cells in the MRRN are activated first, whereas those in the PRRN require stronger stimulation strengths. The output from the MLR on one side activates reticulospinal neurons on both sides in a highly symmetrical fashion. This could account for the symmetrical bilateral locomotor output evoked during unilateral stimulation of the MLR in all animal species tested to date. Interestingly, muscarinic receptor activation reduces sensory inputs to reticulospinal neurons and, under natural conditions, the activation of MLR cholinergic neurons will likely reduce sensory inflow. Moreover, exposing the brainstem to muscarinic agonists generates sustained recurring depolarizations in reticulospinal neurons through pre-reticular effects. Cells in the caudal half of the rhombencephalon appear to be involved and we propose that the activation of these muscarinoceptive cells could provide additional excitation to reticulospinal cells when the MLR is activated under natural conditions. One important question relates to sources of inputs to the MLR. We found that substance P excites the MLR, whereas GABA inputs tonically maintain the MLR inhibited and removal of this inhibition initiates locomotion. Other locomotor centers exist such as a region in the ventral thalamus projecting directly to reticulospinal cells. This region, referred to as the diencephalic locomotor region, receives inputs from several areas in the forebrain and is likely important for goal-directed locomotion. In summary, this review focuses on the most recent findings relative to initiation of lamprey locomotion in response to sensory and internal cues in lampreys.