Great post. It raised a question for me: why did evolution give us the pleasure mechanism at all, if the urge mechanism is sufficient to make us do stuff?
I agree that pleasure has something to do with learning, but I don’t see why the “urge” or “desire” mechanism couldn’t help us learn to do rewarding things without the existence of pleasure.
Without pleasure, things could work like this: If X is good for the animal, make the animal do X more often.
With pleasure, like this: If X is good for the animal, make the animal feel pleasure. Make the animal seek pleasure. (Therefore the animal will do X more often.)
So pleasure would seem to be a kind of buffer. My guess is that its purpose is to reduce the number of modifications to the animal’s desires, thereby reducing the likelihood of mistaken modifications, which would be impossible to override.
Awesome answer, thanks! So the “urge” mechanism is for things we know how to do, and the “pleasure” mechanism is for things we don’t? Now I wonder how to test this idea.
I would say the study supports the thesis that they learn, but then aren’t motivated to act:
“A retest after 24 h showed that DD mice can learn and remember in the absence of dopamine, leading to the inference that the lack of dopamine results in a performance/motivational decrement that masks their learning competence for this relatively simple task.”
I really wish I could get into that paper. I’d like to know whether was dopamine precursor given to the rats before the retest to enable eating? If so the learning may have been buffered and acquired through sleep or there is a different method for learning in sleep. I’ll see if I can get to it in the next few days.
I’d agree that some learning did occur without dopamine, the knowledge of where to go was learnt. The brain is to complex to mediate all learning with direct feedback. What we are interested in is learning what should be done. That is the behaviour was learnt but that the behaviour should be performed wasn’t immediately learnt. Or in other words it didn’t know it should be motivated.
A bad experiment specification: it only tests that brains don’t work so well after you damage them. (That is, more detail is absolutely necessary in this case.)
From the link (pretty much the whole content unless you have access)
Dopamine-deficient (DD) mice have selective inactivation of the tyrosine hydroxylase
gene in dopaminergic neurons, and they die of starvation and dehydration at 3-4
weeks of age. Daily injections of L-DOPA (50 mg/kg, i.p.) starting approximately 2
weeks after birth allow these animals to eat and drink enough for survival and growth.
They are hyperactive for 6-9 h after receiving L-DOPA and become hypoactive
thereafter. Because these animals can be tested in the presence or absence of DA,
they were used to determine whether DA is necessary for learning to occur. DD mice > were tested for learning to swim to an escape platform in a straight alley in the
presence (30 min after an L-DOPA injection) or absence (22-24 h after an L-DOPA
injection) of dopamine. The groups were split 24 h later and retested 30 min or 22-24 h after their last L-DOPA injection. In the initial test, DD mice without dopamine
showed no evidence of learning, whereas those with dopamine had a learning curve
similar in slope to controls but significantly slower. A retest after 24 h showed that DD
mice can learn and remember in the absence of dopamine, leading to the inference
that the lack of dopamine results in a performance/motivational decrement that masks
their learning competence for this relatively simple task.
That is: The mice were made so they couldn’t manufacture dopamine naturally due to inability to make a precursor. Some were then given a dopamine precursor suplement they couldn’t make that enabled them to make dopamine. These learnt almost as well as mice that could manufacture dopamine by themselves. So they showed that they could make the mice learn almost as well as those without damage by replacing something that was lost.
If this doesn’t narrow down things enough, what more do you want?
My first thought: redundancy. Having multiple circuits for the same task means there is a higher probability that at least one of them is working. However, this doesn’t explain the differences between the two circuits.
Great post. It raised a question for me: why did evolution give us the pleasure mechanism at all, if the urge mechanism is sufficient to make us do stuff?
The “urge” mechanism does not help us learn to do rewarding things.
I agree that pleasure has something to do with learning, but I don’t see why the “urge” or “desire” mechanism couldn’t help us learn to do rewarding things without the existence of pleasure.
Without pleasure, things could work like this: If X is good for the animal, make the animal do X more often.
With pleasure, like this: If X is good for the animal, make the animal feel pleasure. Make the animal seek pleasure. (Therefore the animal will do X more often.)
So pleasure would seem to be a kind of buffer. My guess is that its purpose is to reduce the number of modifications to the animal’s desires, thereby reducing the likelihood of mistaken modifications, which would be impossible to override.
Awesome answer, thanks! So the “urge” mechanism is for things we know how to do, and the “pleasure” mechanism is for things we don’t? Now I wonder how to test this idea.
If dopamine = urge you can make dopamine deficient mice. They don’t learn so well...
I would say the study supports the thesis that they learn, but then aren’t motivated to act:
“A retest after 24 h showed that DD mice can learn and remember in the absence of dopamine, leading to the inference that the lack of dopamine results in a performance/motivational decrement that masks their learning competence for this relatively simple task.”
http://www.ncbi.nlm.nih.gov/pubmed/14684249
I really wish I could get into that paper. I’d like to know whether was dopamine precursor given to the rats before the retest to enable eating? If so the learning may have been buffered and acquired through sleep or there is a different method for learning in sleep. I’ll see if I can get to it in the next few days.
I’d agree that some learning did occur without dopamine, the knowledge of where to go was learnt. The brain is to complex to mediate all learning with direct feedback. What we are interested in is learning what should be done. That is the behaviour was learnt but that the behaviour should be performed wasn’t immediately learnt. Or in other words it didn’t know it should be motivated.
There is lots of work on dopamine and learning. I’m currently watching another interesting video on the subject.
Do you know of any related to opioids? All I can find some stuff on fear response learning.
My school let me at the full text of the paper; here ’tis.
Thanks. I’ll read it later.
A bad experiment specification: it only tests that brains don’t work so well after you damage them. (That is, more detail is absolutely necessary in this case.)
From the link (pretty much the whole content unless you have access)
That is: The mice were made so they couldn’t manufacture dopamine naturally due to inability to make a precursor. Some were then given a dopamine precursor suplement they couldn’t make that enabled them to make dopamine. These learnt almost as well as mice that could manufacture dopamine by themselves. So they showed that they could make the mice learn almost as well as those without damage by replacing something that was lost.
If this doesn’t narrow down things enough, what more do you want?
The dopamine and opiate mechanisms are rather tangled together in practice:
The following study tests the hypothesis that dopamine is an essential mediator of various opiate-induced responses:
http://www.nature.com/nature/journal/v438/n7069/full/nature04172.html
My first thought: redundancy. Having multiple circuits for the same task means there is a higher probability that at least one of them is working. However, this doesn’t explain the differences between the two circuits.