The Goal-Directed Nature of Relapse

One of the biggest problems with drug addiction is that it is a disorder that is characterized by relapse. You just CAN’T QUIT. It’d be one thing if you got hooked, got sober, there were some initial bad reactions, and then you were ok. But drug addiction isn’t like that. Drug addicts relapse, even when they are completely and totally sure they never want to do the drug again, when they know the drug isn’t worth it. They relapse anyway. And this is one of the biggest problems with trying to treat drug addicts. Scientists have been working for years to determine what triggers relapse to drug taking behavior, what connections in the brain are involved, and how permanent they are. Answers so far: a lot of triggers, lots of connections, and pretty long-lasting.
So when Sci saw this paper in PLoS ONE, she got excited:
ResearchBlogging.org Root et al. “Evidence for Habitual and Goal-Directed Behavior Following Devaluation of Cocaine: A Multifaceted Interpretation of Relapse” PLoS ONE, 2009.


As of right now, there are two major hypothesis of drug relapse: goal-related and habit-related. It is Sci’s personal opinion that the reality probably involves some of both of these hypotheses. The habit hypothesis is that drug-related cues (like drug taking needles, the crack house, or the friend who you do drugs with) cause the start of a habit cycle of behavior. Once you have the cue, habit kicks in, and drug-taking just happens, regardless of the negative consequences and all promises you made to the contrary. On the other hand, there is the goal-related hypothesis, which is the idea that the drug-related cues make you anticipate the positive effects of the drug, and this makes you seek them out, overwhelming all negative consequences and promises you made that you wouldn’t do this anymore.
So is it habit? Or is it goal-directed? In this study, the authors wanted to find out. Using some rats, and some cocaine, and a little compound we like to call lithium chloride.
So how do you do this? Well, the scientists reasoned that, if cocaine taking had become habit, rats that had been on coke, and then abstinent, would relapse when a lever and light were present, showing that cocaine was available. To determine this, they gave a bunch of rats the opportunity to self-administer cocaine. All rats were trained on how to self-administer cocaine using a lever and a cue light.
Once the animals had all learned to self-administer, half of them got cocaine alone every time they pressed the lever. The other half got cocaine paired with lithium chloride (LiCl), a compound that isn’t bad for you, but tastes AWFUL and makes for one very queasy rat. I should note that the coke alone rats also got LiCl, only on a different day and not paired with a cue light, just to control who got what.
Once the animals had all been through training, they were taken off the cocaine for a month, which is a pretty long time for a rat. They then put them back in the chamber with the lever and cue light present (but no cocaine). The coke trained animals of course started banging on the lever for their hit, while the coke-LiCl paired guys banged less. Interestingly, though, when the cocaine WAS present along with the cues, the LiCl rats hit just as hard. The authors concluded that relapse to cocaine-seeking behavior is BOTH habitual and goal-directed (told you so!).
So, here’s what we like:
I like how this study has all animals receiving equal amounts of cocaine, and equal amounts of LiCl. With LiCl, cocaine responding can be reduced, so I like how they controlled for that.
And here’s what we don’t like:
This study is very simple, and does imply possible goal-directed behavior. But honestly, Sci finds this study to be incomplete. It seems like there’s one main set of data, one experiment, which they analyzed the living heck out of. Which is fine, but this is coke self-administration, and there’s a lot more they could have done with what they had.
1) First off, the animals ALL learned to self-administer cocaine before the LiCl group got treated to be queasy. So they could well have formed some habit behavior for the lever and cues right off, which then got erased by the LiCl. This wasn’t tested. It definitely could have been. I want to see them relapse, and then pair with LiCl and “stop” relapse by stopping the goal-directed behavior. And I want to see it more than once, over a long period of time.
2) I want to SEE that LiCl paired with cocaine is decreasing the reward value of the coke stimulus. For that, I’d like to see some conditioned place preference testing, which is considered to be a measure of the rewarding value of stimuli in rodents (basically, give a rat two chambers, with cues to tell them apart, like different smells or something. Pair one side with cocaine, and one with saline. Do this for a few times. Then place the rat in the box and see which chamber they like more. If LiCl has reduced the rewarding value of cocaine, cocaine place preference should be reduced in these rats, and I’d like to see them do this following abstinence), and though they mention that LiCl doesn’t block cocaine conditioned place preference, it doesn’t block EXPRESSION, but giving it concurrently to training, or prior to training, may have a different effect).
Or perhaps, instead of conditioned place preference, some progressive ratio studies (which is a lever pressing task in which the animal has to press once for coke, then twice, then four times, then 8, then 16…etc, until the animal decides its just not worth it. If cocaine is “worth” less to the LiCl treated rats, their progressive ratio values would be lower, because they aren’t willing to “pay”).
3) They talk about both “habitual” and “goal-directed” responding, but it’s very confusing to tell which is which (Sci reads stuff like this all the time, and was still confused). They mention that habitual responding may have been taking place at the beginning of the session, and goal-directed later, but they don’t really go into it. They don’t seem able to really separate the habitual and goal-directed responding in the results or discussion, and I don’t think the experimental set-up can really even tell the difference. Honestly, the DID have results, but were unable to really interpret them without further studies. A few more studies in the rats really could have given them some better ideas. And note that, when the coke was present, all the animals were self-administering, regardless of LiCl, implying that the LiCl treatment wasn’t long-term effective at all, and that the animals could pick up the coke self-administration right where they left off.
4) They conclude that reducing the rewarding effects of cocaine in addicts (by giving them something nasty in conjunction or something) could help prevent relapse. And this is where I need to get a little bit snarky. Oh REALLY?! Tell me something I didn’t know. Pairing chocolate cake with pepper will indeed make me want less chocolate cake. But there’s a big problem with this idea. When I want chocolate cake bad enough, I’m not going to shake pepper on it. And I’m going to have my cake and eat it, too.
We’ve been trying this technique with alcoholics for YEARS. And it doesn’t work well at all. There is a drug in use, called Antabuse, which is the drug disulfuram. This drug prevents elimination of acetylaldehyde from the body. Acetylaldehyde is a breakdown product of ethanol metabolism, and causes some NASTY feelings. Like the worst hangover you’ve ever had. While drinking. So you think, hey yeah, then people won’t want to drink. Well, sure. But try and get them to stay on their medication. Doesn’t happen. When they get the craving for alcohol, they WILL skip their medication, throw it down the drain, and relapse. I seriously doubt that a similar treatment for cocaine would be any different. You have to FORCE them to take it every day. And even then, sometimes they’ll get around it. (Best study had a 50% effectiveness under constant supervision, but I’d love to see where those addicts are now.)
5) Can I see a bar graph of total responses over the extinction session? Any differences?
So there you have it. Perhaps Sci interpreted this paper incorrectly, but it seems the authors were just as confused. When in doubt, a little more data supporting your point is never amiss.
And don’t eat your cake with pepper on it. All it does is make you hate pepper.
Root DH, Fabbricatore AT, Barker DJ, Ma S, Pawlak AP, & West MO (2009). Evidence for habitual and goal-directed behavior following devaluation of cocaine: a multifaceted interpretation of relapse. PloS one, 4 (9) PMID: 19779607

9 Responses

  1. OK, so I admittedly didn’t read the article, just your recap, but I think I understand. What seems problematic for me is in the final session, where there IS cocaine being administered and they found that both groups lever pressed equally…is this surprising? There’s no LiCl! The paired group is probably freaking out that they’ve got their “old” cocaine back (I’d imagine it would only take one or two presses to figure this out), while the unpaired group is just happy not to be on the wagon anymore.
    I agree with your point that we need to see that the paired group resists bar pressing over a long period of time, on multiple occasions (of course, I’d imagine this would only happen if the animals don’t experience LiCl-free cocaine in the interim). I don’t think the results necessarily suggest that the LiCl effects aren’t long lasting, but when they give the animals cocaine without it, I’d imagine that the positive experience would override the previous negative memory. Like you mention in the cake analogy, if there’s not pepper in your cake, you’re going to eat it.

  2. Not only is it likely that human relapse involves a combination of goal-directed and habitual motivators, but the proportion of the two is likely to be different in different people. For example, I have known some potheads who would get very anxious if they ran out of weed, and would go to great lengths to secure some more. And once they bought the weed, they would admire it smell it, carefully break it apart and clean it, and roll a fancy-ass doobie.
    And then they would frequently just put the doobie down and seem quite at peace, until one of the other potheads would be all like, “Dude, what the fuck are you waiting for?! Spark that motherfucker UP!!”

  3. HAH!!! That is totally me CPP… I used to get so fucking cranky and angst ridden, when I ran out of herb. Might be a full day or more after I got some, before I actually smoked it. But I was totally creeped out if I didn’t have a packed pipe in my happy tin.
    My ex, on the other hand, was more like I imagine you to have been, back in the day…
    Sci –
    I think that disulfuram was one of those things that occurred to someone who had no experience with actual – you know – addicts. To me, it is a perfect example of the disconnect between lab science and clinical treatment, that I have bitched about once or twice. I mean it is one thing to use it on halfway house and group home residents, where the intake can be monitored. But who the hell was thinking this would be very successful outside those parameters?
    Although a junkie friend of mine had a great suggestion, assuming it might be possible – add it to benzos or clonidine and help folks dry out with those. Not necessarily the healthiest way to go, but when we’re talking people who have alcohol problems that are killing them anyways, it is a total step up…

  4. I’m exactly like DuWayne describes, except it’s cigarettes for me. I can go hours without smoking if I have an extra pack laying around, but as soon as I’m down to two or three cigarettes… I begin to panic.

  5. There’s also the question of whether reinstatement of lever pressing is really a good model of “relapse.” Rats will start working for cocaine after seeing a cocaine-paired stimulus. That isn’t the same thing as liking cocaine and being unable to quit because the government has threatened you with violence and confinement. It’s not the same thing as liking cocaine but being unable to cut back because it’s interfering with the attainment of other goals, or because it’s causing health effects that you know could kill you, intellectually.
    In what sense do the rats in reinstatement models wish to reduce their cocaine intake yet fail to inhibit lever pressing?
    Addiction is an interaction between biological factors and social contexts. Someone might not have any desire to cut back on their drug use whatsoever, except for the fact that they need to pass a drug test. Such a difficulty arises in part from dysfunction of impulse control systems. It also arises, I would think, because drug use is at least as old as agriculture. There’s archeological evidence for the use of betel nut 13,000 years ago, and coca 5000-7000 years ago. Rock art from the Stone Age depicts the types of grids, wavy lines, etc. people see when taking certain drugs (or in other sorts of trances).
    Thus, drug use is an ancient behavior, commonly found in non-Christian religious contexts. As part of the important missionary work of eradicating indigenous cultures, it had to be decided that Drugs Are Bad. In a way, telling people not to use drugs is as ridiculous as telling them not to plant seeds or draw pictures.

  6. Sci,
    I have read some of your work and suggestions before without making too many comments. I respect the idea of expressing opinion regarding science and think it fosters the type of Socratic discussion necessary to solve problems.
    In this case though, I feel the need to speak out.
    Now, unlike the first few comments, I took a few minutes to read this article and generate my own thoughts (part of the beauty of an open source journal). In brief, I feel that although it is a great idea to post scientific ideas and discuss them in a way that might draw in anyone who cares to listen, with this task comes the responsibility of reporting things with a greater degree of accuracy than I see here.
    Your description of the article’s paradigm is off. Now, This might be fine if the inaccuracies were in a way that switched a tone with a light, or a lever with a nose poke response. However, in this case your inaccuracies paint a picture that is far enough off to detract from science that –although I agree is imperfect in some ways– has some validity.
    Furthermore, in the beginning of your blog, you state “It is Sci’s personal opinion that the reality [of addiction] probably involves some of both of these hypotheses.”
    I find this to be a major pivot point around which the authors center the whole discussion.
    I enjoy your writing, but in this case I feel like the excitement of a new blog drew a number of premature, and in some cases inaccurate, conclusions.

  7. SWresearcher: Thanks for your comments. I actually re-read what I wrote in the light of day today, and I agree with you that I misrepresented the paradigm. The stimulus was paired with a tone and cue light, and only when the tone was present was cocaine administered. And I’m sorry that I presented the paradigm in such a way as to make it appear invalid, because I did not find that to be the case, there were actually several things I liked about the paradigm, but I’m still not sure that it accurately reflects a division between the goal-directed and habitual portions of relapse.
    And yeah, you’re right, I said I thought it was both, and so did they, and I only acknowledged it with an “I told you so!” at the end. I should have made that more clear. More clarifications will follow shortly.

  8. no one has to be a junkie to become addicted to something. ^^’… i see that alot w/ my adhd n anti-depressants… (0_0′)

  9. Chocolate cake with pepper on it – my favourite! Great analogy, but it’s left me with a hell of a chocolate craving.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: