| Summary: | Feed-forward inhibitory circuits are common building blocks in the mammalian brain and lead to excitatory input also activating inhibitory input to a common postsynaptic neuron. Such circuits are important for regulating neuronal excitability and timing of activity in the brain. In this thesis I have explored the mechanisms and consequences of feed-forward inhibition in the rat cerebellar cortex, which is known to be involved in coordination and timing of movement. Voltage clamp recordings from Purkinje cells in cerebellar slices exhibit a biphasic current waveform in response to stimulation of parallel fibres, consisting of an excitatory postsynaptic current (EPSC) followed by an inhibitory postsynaptic current (IPSC). The latency difference between the two components - only 1.4 ms - and the complete block of the biphasic response by glutamate receptor antagonists confirmed the second component as feed forward inhibition (FFI). The rapid onset of FFI shortens EPSPs, which enhances spike precision and limits summation of independent inputs. Next, I showed that the latency of FFI does not change with distance along active parallel fibres. This suggests that desynchronisation of action potentials travelling along the parallel fibres is insufficient to cause feed-forward inhibition to arrive ahead of excitation, a theory previously used to explain the observed lack of 'beams' of active Purkinje cells along the parallel fibres. Instead, it is argued that this may result from spatial or temporal spread of activity in the granule cell layer leading to early arrival of inhibition. Both excitation and inhibition in Purkinje cells are subject to plastic changes induced by climbing fibre activation. In a feed-forward network, what is the net effect of this plasticity on the output of the cerebellar cortex First I showed that both inhibition and excitation undergo long-term depression (LTD) to a similar extent when paired with climbing fibre input. This plasticity was reflected in corresponding changes in Purkinje cell spike output triggered by independent inhibitory and excitatory inputs: parallel fibre LTD reduced, and LTD of inhibition increased the number of spikes evoked by the respective inputs. To examine the net effect of simultaneous plasticity of inhibition and excitation on Purkinje cell output with a feed-forward input, I simulated synaptic inputs with dynamic clamp and systematically changed the ratio of excitation and inhibition as well as the amplitude of both components. Depressing both components as observed when pairing the isolated components with the climbing fibre, reduced spike output for feed-forward inputs with small inhibitory components, while for inputs with stronger inhibition the spike output increased. Finally, I showed that pauses after spike bursts evoked by strong parallel fibre inputs in the absence of inhibition scaled with input strength. A classical climbing fibre LTD protocol reduced these pauses, which thus encode information stored by synaptic plasticity for downstream neurons. These findings are discussed in the context of classical theories of cerebellar learning, which are concluded to require revision or refinement.
|
|---|
