, 2013b). Finally, the phase shifts of extra-SCN oscillators in the OB and SN but not in the CPU were accelerated by the SCN lesion in parallel with the phase shift of the activity band of the MAP-induced behavioral rhythm. Although the circadian rhythm in the CPU was not significantly phase-shifted by R-MAP as compared with that by R-Water, this does not necessarily indicate that MAP did not affect the circadian oscillator in this structure. As R-Water affected the circadian oscillation

in the CPU in the absence of the SCN, R-Water might be inappropriate as a control for R-MAP. When compared with the circadian phases under ad lib feeding and drinking (Natsubori et al., 2013a), a small but statistically significant Proteasome inhibitor phase-advance was detected in the CPU

by R-MAP. Thus, R-MAP could also influence the circadian oscillation in the CPU. The above considerations lead us to the hypothesis that MAO is a complex or population oscillator consisting of multiple extra-SCN circadian oscillators (Fig. 9). Chronic MAP treatment reorganises the networks of these extra-SCN oscillators to build-up MAO. The circadian oscillators in the OB, PC, see more SN and probably CPU are important components but the involvement of these in other parts of the brain is not excluded in MAO (Model 1). The structures examined in the present study are the major components of the brain dopaminergic system, and it is highly possible that these circadian oscillators in some of these structures are directly affected by MAP treatment, as MAP is an antagonist of the dopamine transporter and activates the dopaminergic system in the brain.

Alternatively, the extra-SCN circadian oscillators in the OB and SN are not components of MAO but slave oscillators located downstream of MAO (Model 2). MAO is located somewhere else. This alternative is less probable because the extent and direction of phase shifts by R-MAP were different among the extra-SCN brain oscillators. Feedback effects from behavior on phasing of the extra-SCN oscillators are possible but also less likely, because the phase responses were different depending on the area examined and the treatment given (Natsubori et al., 2013a) Farnesyltransferase even though MAP-induced behavior enhancement was not much different among them. On the other hand, ad-MAP revealed behavioral rhythms in the R-Water group when the bilateral SCN was lesioned. The behavioral rhythms started to free-run from the phase immediately after the daily water supply (Fig. 2), indicating that R-Water induced behavioral rhythms in the absence of the SCN circadian pacemaker. The free-running period was close to 24 h and significantly different from that of R-MAP-induced behavioral rhythm (Fig. 4B). The period was rather similar to FEO (Yoshihara et al., 1997).