A Cognitive Model of Craving – Pt 3

In an earlier blog we asked the question whether  neurobiological or “conditioning” or reinforcement models of craving predict relapse in abstinent alcoholics and addicts?

For us this is the most essential question. How do we explain relapse in those individuals motivated to remain abstinent, especially when they have followed some form of treatment, including 12 step groups.

We have seen that most relapse seems to be prompted by psychological stressors such as interpersonal relationships and the failure to cope with these.

This is very different to conditioning or reinforcement models that simply posit that people relapse because of the lure of alcohol or drug related stimuli, “cues”, or cues in the presence of stress or negative emotions, which we believe does have some affect.

Equally we have shown that in treatment seeking individuals there seems to be an automatic avoidance of cues so attentional bias does not really apply to this group plus there is a negative memory association bias in this abstinent, treatment seeking group also. So why do these people relapse?

What is the craving process prior to relapse for this group? . This is hugely important as neurobiological accounts do not predict relapse, so what does?

Over the next two blogs we will forward a model of craving or “mental obsessing” which we believe more accurately models the mechanisms which lead these individuals committed to staying sober and in recovery to relapse.

Ultimately we believe it may the maladaptive cognitive-affective reaction to naturally occurring  intrusive thoughts about alcohol or drugs (which are also the function of emotional dysregulation) that creates a proliferation of such thoughts, until they become obsessive, and which escalates stress and emotional distress to such an extent that the individual relapses to silence these tortuous obsessive thoughts.

These thoughts may not always be about alcohol or drugs. They may also contain negative perceptions of self, such as low self esteem and negative self schemas as the consequence of abusive early childhoods. These may result in “I am not good enough” thinking or “to hell with it!” relapse which have little to do with an appetitive urge to drink as in some reinforcement models. They are more akin to escape from self.

So models of addiction tend to focus on neurobiological substrates underlying addiction rather than on how affective (and cognitive) processing mediate addictive behaviours (1) although 80% of problem drinkers after outpatient treatment reported drinking episodes aimed at manipulating thoughts or emotions (2), with the majority of treatment clients attributing their relapse to interpersonal stress or negative emotions (3).

Also the involuntary retrieval of drug related thoughts is a hallmark of addicted populations. Over 70% of smokers stated that urges disrupted their thinking or functioning (4). Intensity of obsessive thoughts about alcohol predict relapse rate (5), with addicts motivated to use drugs to “silence” obsessive thoughts (6).  The idea that abstinence automatically decreases alcohol-related thoughts is challenged by research and supported by clinical observation that among abstinent alcohol abusers, alcohol-related thoughts and intrusions are the rule rather than exception (7).

So if emotion regulation difficulties and related intrusive thoughts are so prevalent in recovering abstinent addicts and alcoholics how do we account for this in a satisfactory and comprehensive theory of craving?

One study important to the conceptual framework set out here (6) used heart rate variability (HRV) measures, as a putative index of emotional regulation, to illustrate how craving involves cognitive-emotional processing and how conditioning models may not fully explain  ‘craving’.  This is consistent with the increasing concern in the literature about the applicably of such “one-dimensional” conditioning models explaining the results of cue reactivity studies (6). This study, among various findings, showed a link between HRV and obsessive thoughts,  in simple terms, the greater the emotional dysregulation, the greater the obsessive thoughts about alcohol.

It may even be that these “conditioning” reinforcement models or dopaminergic or stress-based models are describing “urges” rather than craving.  For us “craving” is distinct but interdependent on this “urge” state, it is partly triggered by it, if you like.

As an alternative to such passive “respondent” or “conditioning” models, some researchers have advocated the use of information-processing theory to understand how dependent individuals react in their encounter with “drug-related” cues (external and internal, e.g. stress or negative emotions) (6). Craving may thus be a different phenomenological experience to that of the physiological urges, although one may prompt the other.

According to one ‘info-processing’ view of craving, forwarded by Stephen Tiffany (4), ‘craving’,  occurs only when the automatic approach behaviour commonly seen in addicts in thwarted. This is particularly pertinent to those abstinent, treatment seeking individuals. In addiction, drug use behaviour develops various rituals around the seeking, preparation and consumption of drugs. These habitual procedures become stored in memory, in automatized action schemata  or action plans.

Encoded within these unitized memory systems are prompts such as external events (e.g. sight of a hypodermic syringe,) or internal events such as physical states (e.g. NA). Although activation of these memory structures may not be a sufficient for addicts to respond to ‘urges’, via actual drug seeking, they may stimulate approach behaviours.

Tiffany (1990) proposed “urges”, or what we call craving, are said to be associated with conscious efforts to inhibit the operation of drug use action plans (e.g. prevention of relapse or suppressive reaction to intrusive using-related thoughts). In abstinence, these “urges” involve non-automatic (i.e. conscious, effortful) cognitions that compete with automatic (unconscious effortless) drug use related plans. Thus, relapse may occur under two circumstances: when the action plan operates autonomously and when conscious processes to inhibit the action plan (thought suppression) backfire and are unsuccessful.



We agree with Tiffany’s (1990) assertion that, like other stereotypic motor acts, some aspects of the drug-use ritual are susceptible to automatization. In fact in relation to automatic using schemas it is only the “nonautomatic” processing where cognitive resources are consciously devoted to disrupting the course of a perceive threat of relapse and prior experience of these self same affective states in the context of use that activate drug- and alcohol related memories (Bradizza, Stasiewcz, & Maisto, 1994.) and cause “craving”.

The “exhaustive and effortful” effects of “urges” (craving)  in abstinent addicts (Tiffany, 1990 p.158) may reflect consciously trying to inhibit these by thought suppression. Whereas, drug using schemas are firmly established and neurally embedded and require few resources to operate, the “abstinence plan” is poorly established and demands vigilance (i.e. attention) and effort to maintain. It is also a relatively new internal voice and not as familiar to the addict. Therefore, it not surprising that many addicts take the path of least resistance and relapse.

Addictive behaviours thus become increasingly automatic or compulsive in the addiction cycle, which supports Tiffany’s ‘cognitive’ model of automatic action plans. ‘Cravings’ are generated, in our model, by non-automatic, cognitive processes which are invoked to thwart (or interfere) with these drug use action plans.

For example, in abstinent addicts, internal stress/emotional distress provokes automatic action plans (and accompanying intrusive thoughts). These individuals then use non-automatic processing i.e. cognitive control/thought suppression) to ‘fight’ these threatening (naturally occurring) automatic thoughts.

The anterior cingualte cortex (ACC) acts a gateway between what is known as explicit (hippocampal) memory (remembered knowledge about things – e.g. where we drank, with whom, how it felt, noises, smells, atmosphere and ourselves in those situations etc) and implicit (dorsal striatal)  memory (the procedural, how to do memory-  the habitualised procedure of Tiffany’s automatic addiction action plans). The dorsal medial striatum (DMS) plays an important part orchestrating the switching between these memories through a “hippocampal-to-striatal pathway” passing through the ACC (41). It may be ACC hypofunctioning, under extreme stress, which aids transition between explicit and implicit memory networks (42).

Addiction severity is suggested as being represented by a shift in reward processing from ventral stiatum (VS) to DS (28) with this marked by an emergence of automatic thoughts, which the authors suggested as the cognitive thoughts and images of automatized drug action schemata (Tiffany 1990). As addiction escalates there appears to be a greater reliance on implicit rather than explicit (hippocampal) memory too. Also emotional distress is known to recruit the DS region also. So in effect the DS becomes involved in memory, reward and affect in later addiction.  So emotional dysregulation will not only provoke intrusive thoughts, but activate automatic approach behaviour, i.e. will prompt a movement towards getting and consuming drugs and alcohol.

Modell et al, 1992, distinguished between  intrusive thoughts – and memories – in a cognitive component to craving and in compulsions, which is more motoric and action component –  the cognitive component may be governed by the dorsal medial which has connections with the ‘associative’ PFC and lateral DS which is more involved in habitual motor activity As we have already discussed, addiction severity corresponds with the extent of obsessive thoughts as measured by the Obsessive Compulsive Drinking Scale (OCDS) which suggest that as the severity of this illness progresses, so does the intensity of the obsessive thoughts about alcohol and the compulsive behaviours to use alcohol. Kranzler et al. (1999) showed relapsers who scored higher in ‘obsessions’ craving measured by the OCDS predicted relapse in the 12 months after treatment completion. It is tempting to ad that emotional dysregulation also worsens as addiction becomes more severe(  ).

‘Cravings’ are thus generated by non-automatic, cognitive processes which are invoked to thwart (or interfere) with these drug use action plans.  The DMS may be very important in the relapse mechanism we are about to explore.

The DMS may have a potential role in cognitive control of behaviour flexibility and mediating behaviours by hippocampal guidance. As such the DMS and DLS may either compete (Misumori, Yeshenko, Gill and Davis, 2004) or cooperate (Devan, MacDonald, White 1999) under different conditions.  For example, DMS may be activated when a reversal of a previously reinforced response, i.e. habitual response, is required (Eichenbaum et al, 1989). Thus in attempting to inhibit stimulus response, i.e. the automatic alcohol approach behaviour of the DLS,    the DMS activates action-outcome pathways

Thus the ‘cognition and imagery of automatized schema’ becomes increasingly obsessive as the consequence of the anterior cingulate cortex (ACC) detecting conflict between memory intrusions and alerting the prefrontal cortex (PFC) to actively suppress unwanted thoughts (169). This only serves to intensify these thoughts as thought suppression ‘rebounds’ unwanted thoughts more intensely and prolifically into consciousness (170).

This, in abstinent addicts, appears to make the situation worse leading to greater stress reactivity and  need to further inhibit habitual response which activates even more action-outcome memory, e.g. the automatic activation of mental representations in associated memory networks of what course of action has normally been followed to affect the outcome of reducing this distress, i.e. which normally has been to drink.

Whereas the DMS normally in adaptive processes competes with the DLS to resolve a situation, for the abstinent addict, it only increases the problem by suggesting solutions which in fact make the situation more acutely adverse.

For the addict attempting to maintain abstinence, declarative memory and controlled processes may often be “corrupted” in service of promoting or rationalizing drug use. This will occur because negative affect is aversive and intrinsically primes escape and avoidance strategies.

Thus the ACC in recruiting explicit memory to counteract the automatic alcohol related thoughts of the DS may unwittingly be increasing memories of drinking and explicit prompts to drink as this is what has normally been the course of action in such situations of negative emotions.

The best and most well-intentioned efforts to remain sober/clean threaten sobriety most; producing a mnemonic ‘Hydra Effect’ whereby attempts to cut off this terrible flowering of intrusive thoughts leads to increased proliferation of these thoughts and accompanying emotional distress.


This, we posit, is what occurs in the mind of a recovering/abstinent alcoholic and is more akin to the “mental obsession” of the Big Book that purely neurobiological/physiological urge states.

Equally it should be noted that craving or mental obsession does not suggest that the alcoholic or addict in recovery/abstinence is actually motivated or even wants to relapse to former use. One can engage in this “mental obsession” or cognitive craving simply via a maladaptive emotional dysregulation whereby a defective emotional strategy such as thought suppression of threatening intrusive thoughts can set up a chain of reactions which lead to an unfortunate proliferation of thoughts and memories which promote alcohol and drug use to relieve escalating emotional distress which leads to relapse even if the alcoholic or addict in recovery did not even wish it! What else is this other than a craving beyond one’s mental (cognitive) control!

Relapse can happen to an alcoholic or addict if he does not manage his underlying condition of emotional dysregulation, in other words.



References (to follow)

1. Cheetham, A., Allen, N. B., Yücel, M., & Lubman, D. I. (2010). The role of affective dysregulation in drug addiction. Clinical psychology review30(6), 621-634.

2. Sanchez-Craig, M., Annis, H. M., Bronet, A. R., & MacDonald, K. R. (1984). Random assignment to abstinence and controlled drinking: evaluation of a cognitive-behavioral program for problem drinkers. Journal of consulting and clinical psychology52(3), 390.

3. Lowman, C., Allen, J., Stout, R. L., & Group, T. R. R. (1996). Replication and extension of Marlatt’s taxonomy of relapse precipitants: overview of procedures and results. Addiction91(12s1), 51-72.

4. Tiffany, S. T. (1990). A cognitive model of drug urges and drug-use behavior: role of automatic and nonautomatic processes. Psychological review97(2), 147.

5. Bottlender, M., & Soyka, M. (2004). Impact of craving on alcohol relapse during, and 12 months following, outpatient treatment. Alcohol and Alcoholism39(4), 357-361.

6. Ingjaldsson JT, Laberg JC, Thayer JF. Reduced heart rate variability in chronic alcohol abuse: relationship with negative mood, chronic thought suppression, and compulsive drinking. Biological Psychiatry. 2003;54(12):1427–1436.

7.  Hoyer, J., Hacker, J., & Lindenmeyer, J. (2007). Metacognition in alcohol abusers: How are alcohol-related intrusions appraised?


. Bradizza, C. M., Stasiewicz, P. R., & Maisto, S. A. (1994). A conditioning reinterpretation of cognitive events in alcohol and drug cue exposure. Journal of Behavior Therapy and Experimental Psychiatry, 25, 15 – 22

Modell, J. G., Glaser, F. B., Cyr, L. & Mountz, J. M. (1992) Obsessive and compulsive characteristics of craving for alcohol in alcohol abuse and dependence. Alcoholism: Clinical and Experimental Research, 16, 272–274.

. Kranzler, H. R., Mulgrew, C. L., Modesto-Lowe, V. and Burleson, J. A. (1999) Validity of the obsessive compulsive drinking scale (OCDS): Does craving predict drinking behavior? Alcoholism: Clinical and Experimental Research 23108–114.

(Misumori, Yeshenko, Gill and Davis, 2004)

(Devan, MacDonald, White 1999)

(Eichenbaum et al, 1989).

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