Neural activity often becomes rhythmic during mental processing. But there has been no direct proof that rhythmicity, per se, is important for mental function. We assessed this issue in relation to the contribution of hippocampal theta-frequency rhythmicity to learning in the Morris water maze by blocking theta (and other septal inputs to the hippocampus) and then using electrical stimulation to restore rhythmicity. We injected tetracaine into the medial septal area, and so blocked septal input to the hippocampus in rats throughout 16 consecutive trials in a Morris water maze. Rats with no hippocampal theta also showed no initial learning in the maze. Theta rhythmicity in the supramammillary area remained after septal blockade, and we used this to trigger electrical stimulation of the fornix superior. This substantially restored hippocampal theta-like rhythmicity throughout training at a normal frequency but with abnormal wave forms. This treatment applied throughout training substantially restored initial learning. Fixed frequency (7.7 Hz) stimulation produced rhythmic activity and a brief improvement in learning. Irregular stimulation with an average frequency of 7.7 Hz produced little rhythmicity and little improvement in learning. These results demonstrate that brain rhythmicity, per se, can be important for mental processing even when the detailed information originally carried by neurons is lost and when the reinstated pattern of population firing is not normal. The results suggest that the precise frequency of rhythmicity may be important for hippocampal function. Functional rhythmicity needs, therefore, to be included in neural models of cognitive processing. The success of our procedure also suggests that simple alterations of rhythmicity could be used to ameliorate deficits in learning and memory. (c) 2006 Wiley-Liss, Inc.