Aim: 

The role of synaptic plasticity in neuronal circuits is current object of debate. Whereas it is usually assumed that long-term changes in synaptic efficacy are needed to learn activity patterns forming memories in the brain, the impact of these changes on circuit computation remains obscure in most cases. Here we propose an hypothesis on how synaptic plasticity controls computation and spike timing in the cerebellar network.

Methods: 

The methods encompass patch-clamp, multi-electrode array and voltage-sensitive dye imaging in cerebellar slices, as well as computational modelling. A detailed description is reported in the original papers.

Results: 

In the cerebellum granular layer, mossy fiber - granule cell LTP and LTD are expressed presynaptically. The main consequence is that of regulating the dynamics of EPSP temporal summation in the postsynaptic neuron. The impact of this mechanism emerges when the granular layer circuit is considered as a whole. Granular layer activation occurs in short bursts causing granule cell excitation, which is soon (in about 5 ms) interrupted by inhibition through the feed-forward Golgi cell loop. Thus, synapses in the LTP state will force spikes to occur in the permissive time-window, whereas synapses with LTD will generate slow depolarizations and will be unable to generate spikes before inhibition arrives. This effect is spatially organized in a center-surround geometry.

Conclusion: 

LTP and LTD, together with circuit inhibition, implement a filter mechanism determining the spatio-temporal organization of signals transmitted through the cerebellum. This mechanism is potentially capable of causing the millisecond precise adaptations required to learn and store complex motor sequences in the cerebellar cortex.

References: 

Nieus et al. (2006) J Neurophysiol 95:686–699; Mapelli and D'Angelo (2007) J. Neurosci. 27:1285–1296; D'Angelo (2008) Frontiers in Cell Neurosci, in press.