We test whether the complexity of the cardiac interbeat interval time serie
s is simply a consequence of the wide range of scales characterizing human
behavior, especially physical activity, by analyzing data taken from health
y adult subjects under three conditions with controls: (i) a "constant rout
ine" protocol where physical activity and postural changes are kept to a mi
nimum, (ii) sympathetic blockade, and (iii) parasympathetic blockade. We fi
nd that when fluctuations in physical activity and other behavioral modifie
rs are minimized, a remarkable level of complexity of heartbeat dynamics re
mains, while for neuroautonomic blockade the multifractal complexity decrea
ses.