A “Kindling” Model of the Development of Addiction

Sometimes, while daydreaming in the shower or in the car, an insight hits you out of the blue. That happened last week. It occurred to me that the best neurobiological model of addiction has a serious missing link. Addiction develops. It grows. A successful theory of addiction needs to be a developmental theory – a theory of neural development.

In my view, the neural basis of addiction is best captured by Berridge and Robinson’s model of incentive sensitization. In a nutshell, these researchers show that “wanting” and “liking” are quite independent, subserved by different neurochemicals, and addiction is characterized by “wanting,” not “liking.” That helps explain why addicts keep on craving – and obtaining – their substance of choice long after it stops being pleasureful. The research supporting the model shows that, contrary to a tenacious myth, dopamine does not cause pleasure. Rather, dopamine is critical for the pursuit of goals, including the behaviors required to reach them and – most important– the powerful motivations needed to execute those behaviors. According to Berridge and colleagues, dopamine gets released from the midbrain in buckets when addicts are presented with cues associated with their substance of choice. A sight, sound, or memory, reminiscent of that stuff (e.g., a cramp in the gut, the fleeting glimpse of someone who looks like a drug buddy, a scrap of paper dotted with a few flecks of white powder) will activate dopamine release and send it straight to the nucleus accumbens (NAcc; a major component of the ventral striatum) where it induces goal-oriented behavior (when the stuff is available) or craving (when it’s not).  But the power of cues to elicit the addictive impulse must take time to develop. It’s not present the first time you try drugs or booze, or even necessarily the 20th time. This process is therefore called incentive sensitization; because the cues that trigger drug seeking become sensitized over time.

The trouble is, Berridge and colleagues don’t explain how this sensitization takes hold. The cues must be processed somewhere in the back half of your brain (where “perception” first arises) and they must activate the amygdala, the famous limbic structure that produces emotional feelings on the basis of perception. But how do these perceptual and evaluative processes come to trigger the urge, the thrust, the powerful desire that is the essence of approach motivation? That would have to take place in the frontal cortex and its master motivator, the striatum.

I’d been reading a recent chapter by Berridge and Robinson (2011), in preparation for a class I’d soon be teaching on the neuroscience of addiction. But the paper seemed to be missing something – a mechanism. This bothered me for a few days, and then, out of the blue, I remembered the idea of “kindling” – something I’d read about years before. Kindling in neuropsychology initially meant the tendency for animals to get seizures more and more predictably in response to less and less of the seizure-inducing stimulus. So rats might go into a seizure when exposed to an electric shock, but the amount of electricity needed to evoke the seizure would diminish, session after session, and seizures would eventually occur spontaneously, without any shock.

The kindling model was used at first to understand epileptic seizures, which occur more frequently, with less to trigger them, as people age. But in 1992, Robert Post published a kindling model of depression.  According to Post, depressive episodes follow the same developmental trajectory as epilepsy. An initial depressive episode is triggered by a major stressor, but subsequent bouts of depression are triggered by less and less adversity. That’s why adult depressives get, um, depressed….so easily. In his excellent book, Listening to Prozac, Peter Kramer fleshes out the kindling model of depression. He tells of patients who become more vulnerable to depression with age, despite increasingly minor triggers, as exemplified by a concentration camp survivor whose depression, initially elicited by a horrible experience in his youth, came back to haunt him later in life. More recent work uses the kindling concept to describe the development of bipolar disorder and PTSD. What all these accounts have in common is the notion that sensitivity to certain cues – cues that elicit negative effects or negative affects – increases with development.

Could that be what’s going on in addiction? Could kindling explain incentive sensitization? It could, but so far kindling models have only been applied to the elicitation of negative thoughts, moods, and emotions. Can kindling also apply to the elicitation of actions? Impulsive or compulsive actions like acquiring drugs, gambling away the down payment on your house, or drinking yourself out of your marriage or your job? Instead of taking place in the back half of the brain or the amygdala, kindling would have to arise in the frontal brain, the goal-seeking part, and in particular the ventral striatum or NAcc, the seat of motivated action.

That was the thought that came unbidden last week. Hadn’t I read about kindling in the striatum – somewhere? I went to my computer, and the first paper I looked up was a little-known chapter by Don Tucker — Tucker, D. M. (2001). Motivated anatomy: a core-and-shell model of corticolimbic architecture. Handbook of Neuropsychology, 2nd  Edition, Volume 5, Gainotti (Ed.). Elsevier. All I remembered was that the chapter had knocked me out when I first read it, and Tucker has long been one of my favorite brain theorists. I hit the jackpot! In previous work, Tucker had explained how depression could indeed be kindled in the limbic system – most probably in the amygdala. After all, it’s the amygdala’s job to get sensitized, so that stimuli evoke immediate emotional responses based on previous associations. But in this chapter, Tucker went on to describe kindling in the striatum, the source of voluntary behavior.

So, here’s how it might work: The amygdala, which lies at the gateway of perception, gets sensitized by stimuli. As an addict, you get more and more moved by drug-related cues – the empty pill bottle, the email address of your dealer, the buzzing neon sign in front of the liquor store. And the NAcc gets more and more sensitized to a specific goal – getting the substance or doing the activity — and to the series of moves you can make, you must make, to achieve that goal. Sure enough, there’s a dense pathway of axons leading from the amygdala to the striatum, a one-way street from biased perception to biased action.  So incentive sensitization can take place within and between these two systems, both of which are highly programmable, plastic, modifiable, and loaded with synapses that get molded by experience. Both the amygdala and the striatum serve as hubs at the center of extensive cortical systems. Together they form a macrosystem sometimes referred to as the extended amygdala.  The shaping of synaptic networks within this region may be the mechanism by which meaningful perceptions and meaningful actions converge over development, thereby colonizing one’s emotional memory and cementing one’s emotional habits.

I see addiction as developing in exactly that way. The difference from “normal” development is that the kindling that leads to addiction is focused on goals that are so desirable, so very attractive, that they set in motion a feedback loop between “wanting” and doing. Then, other goals become less and less salient, and that’s a serious problem.

12 thoughts on “A “Kindling” Model of the Development of Addiction”

  1. Would this pathway only be feasible with chemical addictions, or could it also be the same way other addictions develop? For example, addiction to your cellphone.

    I personally notice myself incessantly checking my cellphone for no reason, even though I acknowledge it’s stupid and pointless, I seem to do it practically instinctively. Not as bad as a drug habit, but still unsettling…


    1. Mitch, I’m having the same issue. And it makes me feel slightly out of control. (Just slightly, thus far.) I see where you’re going–and though I’m no neuroscientist, I do believe that all addiction behavior involves dopamine and motivation toward a goal. If that’s true, then cell phone addiction counts. So does video game addiction. And anything else that you do more frequently with fewer triggers as the kindling turns to a full fire (in my opinion).


  2. Very nice explanation. I spoke with Robert Post years ago about his kindling model, and he was wary about extending that mechanism to addiction. However, James Halikas at the University of Minnesota ran with the idea with respect to cocaine addiction for awhile, and it seems to make some sense, but doesn’t seem to have ever gotten all that much traction in the research community…


    1. Hi Dirk. I wanted to see if others would reply — but, maybe it’s not political enough! Anyway, I’m very intrigued that Post was reluctant to see addiction as kindling.

      I realize, in thinking about this, that kindling really isn’t an explanation. Rather, it’s a description. You might be interested in Tucker’s exploration of these issues. But we have a long way to go. What are the actual mechanics of “sensitization” — whether we call it incentive sensitization a la Berridge or “kindling”? Are we just referring to feedback between synaptic proliferation and perceptual bias? I mean, that might be enough. More synapses dedicated to seeing things a certain way, or doing things a certain way….feeds back to experience (perceiving, doing) which continues to strengthen like-minded synapses. Is that it? Or is something more going on?

      Only the shadow knows….. for now.


  3. Interesting article. I like the incentive sensitization model as well. I also agree that this model explains more the late stages of drug use once the addictive state has been reached, at a time when the process has perhaps become more glutamate based than dopamine based.

    How exactly does this happen or develop? At what stage does a person move from being a user to an addict in the neurological sense? Does this correspond with psycho-social and/or behaviour changes, or is there a lag or precursor? These are questions that interest me.

    We see changes in delta fosB, which some are suggesting could be an indicator for this shift to addiction, but are not 100% sure of these implications. (Interestingly this is noted in endogenous addictions as well). Is it, perhaps, that the repeated bursts of dopamine have some form of toxic effect that brings about the changes that move one towards addiction?

    We see that users become “programmed” by the expectation of the effects the drug will have. They experience the high the second they make the decision to use, even though the actual using is followed by little relief, and is often almost disappointing, particularly in late stage addiction.

    To me the kindling theory is, as you point out in your comment, more of a description rather than an explanation of the processes that take place, but it is still useful. I am hoping that one day I’ll be standing in the shower and the relationship between the three axes of addiction {(neurobiology:neurochemistry) (thought:behaviour)(Macrosystem:Microsystem)} will become instantly clear!


    1. What a thoughtful comment! I’m afraid my lightning strike wasn’t sufficient to see the whole picture, but you raise some very interesting possibilities that might help.

      Is the shift into addiction from just “using” continuous or discontinuous? You mention several candidates for this shift: change in fosB, dopamine-induced insensitivity, and the switch to glutamate- from dopamine-based synaptic change (which must parallel a shift of center stage from the striatum to the cortex). You also mention a behavioral correlate: the diminution of the “high” and its becoming triggered by anticipation more than by the drug itself.

      Can any of these clues tell us whether the shift is gradual versus sudden? Or — if we discover the profile of change first, can it point us to one mechanism over the others? I don’t know. When I think more about the parallel with the development of depression, with “kindling” as the common denominator, I still don’t know. Depression sometimes seems to lock in quite rapidly, usually in response to a definite trigger.

      It seems that kindling, whatever else it involves, is primarily recursive. It builds on itself. This kind of self-organizing process implies an exponential curve, in other words, a period of rapid growth or acceleration of the process. So….that would look more discontinuous than continuous — more like something that emerges suddenly than something that emerges gradually.

      I’m just thinking out loud, as you are, and we’re a long way from an explanation. Still! Yet this kind of thinking can hopefully lead us to an explanation sooner rather than later. In other words, I guess what I’m saying is that macroscopic modeling is important, and it should complement the microscopic accumulation of data that makes up the meat and potatoes of our science.


      1. Exactly on the same page. What frustrates me so often in this field is the lack of integration of thinking between the behavioural and neuro sciences. To try and approach addiction from an either/or standpoint is very limiting.

        Hopefully as we start examining the endogenous or “behavioural” addictions from a neuro science point of view we will discover more about the switch to addiction.

        I am also worried about the DSM move towards a “spectrum”, because clinically that is not what I see. I see people who use (often successfully), people who abuse (not so successfully, but stop fairly easily once underlying causal issues are dealt with) and then addiction.

        I’ve diverged a bit from your kindling idea. I still think it creates a helpful mental picture. I am working on a cumulative model of addiction which is trying to answer the questions you pose, and certainly your thoughts have helped me expand on some of my thinking.


      2. Maybe the kindling model only really works for the last category you mention — long-term/hard-core addicts. And I agree about the importance of studying behavioural addictions… To me, the best argument against the disease model, at the level of behaviour, is that drug addiction and gambling or sex addiction have very similar profiles. Sooo…I’m planning on attending a conference on behavioural addictions (http://icba.mat.org.hu/) to be held this March in Budapest (of all places).


  4. I think people seeking a contemporary balance between behavior and neuroscience are prone to forgetting that behavior ruled the addiction field exclusively for most of the 20th Century. The move to an emphasis on biochemistry is understandable and long overdue, in my opinion. We’ve had a hundred years of cognitive therapy, and by and large it hasn’t helped addicts. In the future, it’ll probably settle out as a workable mix betweeen the two. But for now, the emphasis on the neurocircuitry involved in addiction is appropriate, in my opinion.


    1. Agreed. But not to the exclusion of behavioural therapies. Clinically I see very little benefit in CBT for heroin addicts in early recovery, but in conjunction with pharmacotherapies I see psychodynamic therapies become effective, but only after the primary addiction has been addressed. It appears I’m not the only one experiencing this:
      <a href="http://download.journals.elsevierhealth.com/pdfs/journals/0002-9343/PIIS0002934312006353.pdf
      " title="CBT and buprenorphine".


  5. Hear! Hear!

    One approach to treatment that may lead the pack in integrating cognitive/psych approaches with neuroscience is mindfulness/meditation. These guys (e.g., http://www.mindandlife.org/research-initiatives/) have carried out hundreds of experiments, looking at neural differences contingent on meditation and related practices. If we can hook this up with neural changes related to addiction, we’ll have built a critically important bridge, made up of both cognitive and neural planks.


Comments are closed.