It has been proposed that cortical networks are inherently able to process temporal information because information about the recent input history is inherently captured by time- dependent changes in the state of the network (Buonomano & Merzenich 1995, Buonomano 2000, Maass et al. 2002).

Although this model failed in many of its key properties, it showed how the connectivity of the cerebellar cortex could represent the time since the onset of a stimulus with subsets of different granule cells that become active at different times (Figure 9A). This time-varying stimulus representation was similar in many respects to the activity assumed in certain of the spectral timing models described above. The key mechanistic difference was that this activity was the natural consequence of the sparse, distributed, and recurrent connectivity of the cerebellar cortex.