Our findings implicate an excitatory neural population in the generation of rhythmicity. We note that the activity of inhibitory neurons involved in reciprocal inhibition between rhythm-generating centers could also influence the frequency of the motor rhythm. Decreasing inhibition in such populations of inhibitory neurons will phase-delay the switching between half-centers, thereby decreasing the frequency of the locomotor rhythm. This effect is most likely what is observed after ablation of inhibitory En1+ neurons (Gosgnach et al., 2006) suggesting that at least part of this population is responsible for reciprocal inhibition

between rhythm-generating half-centers. In addition to connectivity between Shox2 INs, some Shox2 INs provide direct excitation to commissural neurons. Although we show that Shox2off V2a neurons are necessary for normal left-right alternation (see above and Figure 8A), these are not marked by GFP in the Shox2::Cre; Z/EG. Therefore, Metabolism inhibitor these findings demonstrate that Shox2+ V2a and/or Shox2+ non-V2a INs also project to commissural pathways. We speculate

that Shox2+ non-V2a neurons are likely candidates for these projections. So why is left-right coordination not affected in the Shox2–vGluT2Δ/Δ or Shox2-eNpHR mice? The most selleckchem likely explanation for this is that the Shox2+ non-V2a INs and the Shox2off V2a INs drive commissural pathways active at different speeds of locomotion ( Figure 8B; see also Talpalar et al., 2013). The Shox2off V2a commissural pathway seems to be active at medium to high speeds ( Crone et al., 2009) and it is likely that non-V2a Shox2+ neurons, together with other yet-to-be-identified iEINs, drive left-right alternation at lower frequencies of locomotion. Therefore, left-right alternation

at higher frequencies is supported by Shox2off V2a INs and at lower speeds the other rhythm-generating iEINs are capable of maintaining left-right alternation ( Figure 8B). Transsynaptic virus injections demonstrate that many Shox2 INs are premotor INs and located in a lateral population within the spinal cord. Our findings that ablating Shox2+ V2a neurons in the Shox2-Chx10DTA mice does not affect the locomotor frequency Adenosine but leads to increased variability of locomotor bursts strongly suggests that locomotor-related premotor Shox2 INs are Shox2+ V2a neurons. These Shox2+ V2a neurons would then be downstream of the rhythm-generating kernel (Figure 8B). Flexor dominance was detected both in the firing of rhythmic Shox2 INs as well as connectivity profile analysis to motor neurons. In a comparative analysis, we detected approximately three times more Shox2 INs connecting to flexor (TA) than to extensor (GS) motor neurons. This observation is in line with previous findings showing that premotor neurons provide a much stronger synaptic excitation to flexor motor neurons than to extensor motor neurons during locomotor-like activity (Endo and Kiehn, 2008).