Reversible deactivation techniques have shown that the cerebral networ
k: (1) is dynamic, its functions depending on contemporaneous processi
ng elsewhere in the network; (2) is composed of single nodes that cont
ribute to several behaviors; (3) possesses an inherent plasticity that
tends to minimize lesion-induced deficits; and (4) comprises feedforw
ard and lateral connections that contribute in different ways to netwo
rk operations. The next major advances in understanding network operat
ions will probably be made by applying a combination of behavioral, ne
uron-recording and deactivation techniques. The greatest near-term gai
ns ave likely to be made in understanding the contributions that feedb
ack projections make to cerebral network function.