Sites of Plasticity in Eyeblink Conditioning Flashcards
What are two implications from the finding that the cerebellar is very uniform?
- Apply general cerebellar theories to explain classical conditioning
- Use data from classical conditioning to test general predictions
What do models predictis the key synapse for plasticity?
Synapses between parallel fibres and purkinje cells
The evidence on sites of plasticity can be split into two areas, what are they?
- Cerebellum vs brainstem
- Cerebellar cortex vs deep cerebellar nuclei
What does it mean that evidence shows conditioning occurs in decerebrate preparation?
Site of plasticity must be in either the cerebellum or brainstem- it has to be one or another
What is the pro-brainstem argument?
Both sides agree that lesions of the anterior interpositus nucleus stop CRs
Disagree on the mechanism of this effect:
-One side claims this is evidence for plasticity in the cerebellum
-The other claims that the lesions work because they affect the brainstem, causing a ‘performance deficit’
What do we know about the interpositus nucleus?
That it tonically excites the red nucleus so there is always a background excitation on brainstem pathways
How could this tonic activation affect learning?
Learning in brainstem area X produces a command for the CR but that can only affect motor neurons if there is sufficient background excitation from interpositus via the red nucleus
How can we test this involvement of the tonic activation?
Lesion the interpositus nucleus to remove tonic excitation from the motor neurons
Thid could lead to much smaller or undetectable CRs
E.g. if the CR= 30 spikes/second and the tonic activation = 50 spikes/second, the motorneuron threshold = 50 spikes/second
SO removing the 50 spikes/second from the tonic activation coming down means the CR command won’t make it through the motorneuron
What are the anti-brainstem arguments? Removing tonic excitation
Evidence FOR this view is weak
Removal of tonic excitation should affect UCRs but effects on UCRs are weak and unreliable: UCRs still present, even when CRs NOT present
What are the anti-brainstem arguments? Sites
- There’s never been a clear brainstem site proposed for this plastic site so we don’t know where to look
What are the anti-brainstem arguments? Cortical lesions
Evidence AGAINST this view is strong
It is known that the cerebellar cortex inhibits tonically all the time the interpositus nucleus which excites the red nucleus
Unilateral cortical lesions decrease CR amplitude
BUT
Unilateral cortical lesions increase UCR amplitude (slightly)
This cannot be explained by simple tonic effect on motorneurons- must be SPECIFIC effect on CRs
What are the anti-brainstem arguments? Reversible inactivation studies
Evidence AGAINST this view is strong
Inject Muscimol, a GABA antagonist (acts as an inhibitory agent stopping neurons from firing in the target) into the red nucleus
Carry out CC
At the end of training when we test the next day after the muscimol has worn off, there are CRs
- this means performance has been affected by this treatment but NOT LEARNING
However, if the same thing is done in the interpositius nucleus (cerebellum), there are no CR responses after muscimol wears off
- so both performance and LEARNING are affected
We can conclude that the major site of plasticity for rabbit NMR and eyeblink conditioning is within the cerebellum
What is the issue with localising the site of plasticity in the cerebellum?
There are TWO CANDIDATE SITES of plasticity in the cerebellum where both CS and UCS information are available
- cerebellar cortex OR deep cerebellar nuclei (second does not come from a model prediction but from anatomy which is also controversial)
What happens if we lesion the cortical cerebellar cortex?
Unilateral cortical lesions (in right place) impair conditioned responses
bilateral lesions abolish them
cortical lesions block conditioning
Necessary evidence, but not sufficient.
What can we do to inactivate the cerebellar cortex?
We know the parallel fibre-purkinje cell synapses use glutamate as a transmitter
Simple spikes are generated from the activation of AMPA receptors, a subtype of glutamate receptor
We can use the reversible blocking agent CNQX to block AMPA receptors
What did Attwell, Rahman & Yeo (2001) do and what did they find when using CNQX to block AMPA receptors?
Used radioactive labelling to show where the drugs went in the brain
Leave the animal for 6 weeks then section it to show where radioactivity was
Can show the injection here affected the cortex but not the deep nuclei
There were no CRs when the drug was active so no learning of CRs
There was also no CRs the next day after the drug had worn off
What is a caveat in Attwell, Rahman & Yeo’s (2001) study?
CNQX may not be acting just on the synapses between parallel fibres and Purkinje cell
The synapses between mossy fibres and granule cells could also have been affected
However, either of these sites indicates the importance of cerebellar cortex
What did Jirenhead et al. (2007) do and find with cortical electrophysiology?
Done in a decerebrate ferret
Recorded from the cortical eyeblink area in HVI and recorded what happened during conditioning
The first trial had effectively no effect on the CS
But once learning is complete, he showed for the first time that the relevant cortical purkinje cells ceased firing, as predicted
This was reflected in the electrophysiology responses as an absence of firing of relevant purkinje cell nuerons during the acquisition of CRs
What is a caveat of Jirenhead et al.’s (2007) study?
These results were obtained in the decerebrate preparation (ferret) and while preliminary data suggest Purkinje cells behave similarly in intact rabbits and mice, there is still some problems with identification of relevant region of cortex
What is a caution regarding the cerebellar cortex in learning?
Current evidence supports key role for cerebellar cortex in initial learning
It is possible that the changes in Purkinje cell firing subsequently produce learning in the deep nuclei
Evidence currently very confusing, e.g. Boele et al (2013)
If it’s the cerebellar cortex that control initial learning, why is deep nuclear inactivation also effective at blocking learning?
The suggestion is that the interpositus nucleus projects to the inferior olive, completing a LOOP
Cerebellar deep nuclei inactivation disrupts this loop so the system no longer works
So inactivating the interpositus nucleus disinhibits the inferior olive so it fires more, and increases the complex spikes in the cortex
-this means fewer simple spikes = an unintentional inactivation of the relevant area of cerebellar cortex
Describe the evidence supporting this loop theory. Zucca et al. (2016)
Electrophysiological evidence =
If you increase inferior olive firing rates, you abolish simple spikes in cerebellar cortex
- this suppresses conditioned eyeblink responses
What don’t we know still?
Increasing inferior olive firing abolishes simple spikes and suppresses conditioning
but we don’t know whether inactivating the anterior interpositus nucleus has similar effects- has yet to be conclusively demonstrated
But at least we have a reasonable explanation of the effects of deep nucleus activation
