Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

cousin_it comments on Are wireheads happy? - Less Wrong

108 Post author: Yvain 01 January 2010 04:41PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (95)

You are viewing a single comment's thread. Show more comments above.

Comment author: cousin_it 01 January 2010 06:40:23PM *  2 points [-]

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.

Comment author: whpearson 01 January 2010 07:11:15PM 6 points [-]

If dopamine = urge you can make dopamine deficient mice. They don't learn so well...

Comment author: timtyler 01 January 2010 08:22:01PM 4 points [-]

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."

Comment author: whpearson 01 January 2010 09:01:48PM 0 points [-]

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.

Comment author: Alicorn 01 January 2010 10:21:54PM 5 points [-]

My school let me at the full text of the paper; here 'tis.

Comment author: whpearson 02 January 2010 01:18:11AM 0 points [-]

Thanks. I'll read it later.

Comment author: Vladimir_Nesov 01 January 2010 07:42:14PM *  3 points [-]

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.)

Comment author: whpearson 01 January 2010 07:58:01PM *  4 points [-]

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?

Comment author: timtyler 01 January 2010 08:18:32PM 2 points [-]

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