L.4 - Cognitive Bias Modification Flashcards
1
Q
Learning Objectives
A
- Indicate how an attentional bias, implicit associations (memory bias), and an approach bias play a role in addiction [analyzing]
- Name and describe experimental methods to measure cognitive biases [paraphrasing], and apply them to answer a new research question [independent thinking]
- Explain how an attention bias / approach bias can be reduced by Cognitive Bias Modification (CBM), in general terms and on the basis of a case study [paraphrasing and analyzing]
- Argue - based on empirical evidence - that working memory plays a role in the susceptibility to substance abuse [analyzing]
- Explain why it is important to distinguish between proof-of-principle studies and RCT’s when weighing the empirical evidence for the effectiveness of CBM (and knowing important differences between these studies) [paraphrasing]
- Describe the differences between traditional CBM and ABC training, and the rationale behind those changes [evaluating]
- Based on a case study, propose which antecedent cues, behavioral choices and consequences to incorporate as part of ABC training [analyzing, independent thinking]
2
Q
Overview
A
- Theory & Cognitive Biases
- Cognitive training
- Comparison with medication
- Next steps
- Conclusion
3
Q
Introduction
A
- Central paradox of addiction: adidcted people continue their self-distructive behavior despite knowledge of the consequences
→ imbalance between strong impulsive or associative reactions to drug-related cues and relatively weak reflective or controlled processes
→ this makes the individuals susceptible to sensitized cues triggering action tendencies, leading to addictive behavior
4
Q
what are cognitive biases?
- what are the main ones?
A
Biased pattern of information processing
- Attention
> relatively strong attention (engagement/disengagement) for substance-related stimuli
> things related to addiction grab attention
→ e.g. failure to disengage from attention to beer bottle
- Action-tendencies
> approach-bias vs avoid-bias
→ e.g. tendency to approach beer bottles
- Memory
> relatively easy retrieval substance-related associations (antecedents/effects)
> easier to retrieve in memory the associations that involve substances
! unique prediction of behavior after controlling for explicit cognitions
→ these biases predict unique variances in individual differences in addictive behavior
> e.g. if we measure tendency to engage with alcohol, this can predict unique variance
5
Q
what are some example questions regarding the cognitive biases in addiction?
A
Attention: where do you look?
Memory: first thing that you think of?
Action tendency: what movements would you tend to?
6
Q
what is CBM?
- what is it used for?
A
CBM = cognitive bias modification
1. test the causal role of biases
2. clinical application to reduce maladaptive cognitive biases
> already some studies have shown CBM to be successful in reducing attention bias for social anxiety and addiction
7
Q
what are the two accounts for CBM?
A
Dual process account
> biases are driven by (bottom-up) mental associations (Pavlovian and/or instrumental), and their effect is moderated by (top-down) executive control processes
> simple automatic and associative mental processes
Inferential account (Wiers)
> the effectiveness of CBM depends on inferential processes
> people actively interpret and evaluate information
> CBM works because it changes how people consciously infer meaning from cues or situations
> during CBM, participants engage in reasoning or interpretation (top-down perspective; deliberate and conscious reflection)
8
Q
Attentional bias
- how can it be measured?
+ general idea
A
- form of selective attention that addicts have for drug-associated stimuli
- can be measured via eye movement & Dot probe measure
> motivationally relevant stimuli attract and capture attention
> related to subjective craving (with elaboration in working memory) [depends on your history, etc]
9
Q
what is the Dot probe test?
A
- you get to see two images on computer screen, followed by dots
- press the button corresponding to either one or two dots, depending on how many you saw
> if much quicker in responding when dots are on alcohol side → you most likely like to drink alcohol
! terrible reliability (but can be used for training)
(picture 1)
10
Q
Approach bias
+ how can it be measured?
A
- tendency to approach appetitive stimuli
- can be measured through Approach Avoidance Task (AAT)
> this is also called “the Irrelevant Feature version” → it is similar to the dot-probe test, but this is specifically for approach bias
11
Q
what is the Irrelevant Feature version of the dot probe measure?
A
- the format of the experiment determines the action
> if picture is horizontal/(tilted left) → you push
> if picture is vertical/(tilted right) → you pull - they also did it with neutral images as control conditions (e.g. picture of beer or soft drink)
→ if it’s easier/faster to pull than to push a picture of alcohol, it means that that is your tendency towards the object (beer glasses)
(picture 2)
12
Q
what was the problem with the irrelevant feature version and how was it resolved?
A
- people had different ideas of what was the reference point (me or the glass?)
→ they made the picture become bigger when pulled, and smaller when pushed
13
Q
STUDY!
what are the results of the Irrelevant feature version experiment?
A
- significant difference in light vs heavy drinkers (heavy drinkers are faster to approach alcohol)
- there is specific related risk allele
> OPRM1 G-allele (mu-opioid receptor gene)
→ people with this allele showed stronger approach biases
> also related to cue-induced craving in people with obesity
! one single gene only explains less than 1% of the variance
(picture 3)
14
Q
what is the Relevant feature paradigm?
A
- other version of same experiment where participants in one group have to pull for alcohol picture, and push for other pictures, and viceversa for other group
15
Q
Relevant vs Irrelevant feature paradigm
A
Relevant fp: reliability is higher
Irrelevant fp:
→ don’t instruct anything (more implicit)
→ same for modification
16
Q
what is the dual probe?
what is the advantage?
A
→ measures attentional and approach biases
- you see two videos at the same time, e.g. alcohol and coke ad
- then the videos switch, and you see a specific probe in both videos
- you need to indicate which probe you see, and this shows where your attention was drawn
✓ very good reliability
17
Q
if you drink a lot, do you always have attentional/approach bias?
+ clinic studies
A
- no, it is relative
- on average there is an approach tendency, but not all heavy drinkers show these biases
- it is at group level, not individual level
In recovery clinic
> 1/2 have clear approach bias for alcohol
> 1/3 no approach bias
> 1/6 negative approach bias
18
Q
where else can attentional and approach bias be found?
A
- across many addictions:
> cannabis
> gambling, …
19
Q
Summary of attentional and approach biases
A
- related to substance abuse
- measurement issues related to “irrelevant feature method”
= people respond to something else rather than content (e.g. format of picture, AAT) - generally unreliable (much better for dual probe)
- good for research into relatively automatic mechanisms (group-level, not for individual diagnosis)
20
Q
what is the clinical relevance of the approach and attention bias?
A
- existing treatments (mainly CBT) are not effective for all patients
> 50% relapse within 6 months
> 70% within three years
→ CBM might reduce risk of relapse
21
Q
How can memory biases be assessed?
A
- Reaction Time test (e.g. implicit association test - IAT)
1. on the left side you have “active” and “alcohol”, on the right side you have “neutral” and “softdrinks”
> you get given words like “fun”, or pictures of drinks
2. in second phase you have on one side “neutral” and “alcohol”, and “active” and “softdrinks”
22
Q
what do you measure with a IAT?
A
- you calculate the difference between the reaction time for phase one and phase two
→ this allows us to calculate how easy it is for us to associate different words to different drinks (e.g. alcohol & active)
= found stronger association between alcohol & active in heavy drinkers, compared to alcohol and neutral
23
Q
What have different studies found regarding IATs?
A
- related to drinking, also after controlling for explicit expectancies
> we measure something a bit different from what you ask in a questionnaire - not a pure measure of association
> e.g. EC artefacts (if you’re better at switching, you get shorter reaction time) - alcohol-arousal associations related to heavy drinking, and this could be related to sensitization
> you get a stronger and stronger response if you use drug more often - more recent works highlights the relevance of me-drinker associations
> if you identify yourself as a drinker, this predicts drinking patterns for even 5 years later
24
Q
what is another way to study memorhy bias?
- NON-RT tasks
A
Open-ended memory associations:
- Outcome-behavior associations
> present outcome, assess spontaneously generated behaviors
> top-of-mind awareness test
> e.g. “having fun: _”
- Cue-behavior associations
> present a word or picture cue or context, assess spontaneously generated behaviors
> “friday night: _”
- first associations to ambiguous words which can be alcohol related or not
> e.g. “draft”
! predict addicted behavior over time
! tests do not ask for introspection or recollection, just ask for first word or behavior that comes to mind
25
what is the importance of Cognitive Control processes?
- study hypothesis
- explicit C.C.P. (e.g. working memory) moderate the effect of implicit cognitive process
- hypothesis:
> if you score higher on working memory
→ rationality, pros vs cons, reasoning, ... are all better predictors of smoking, drinking, etc
> if you score lower on working memory
→ implicit cognitive processes are better predictors ("atuomated pilot")
~ 88 adolescents
~ low level of education
26
How was working memory tested in the study?
Self-Ordered Pointing Task (SOPT)
- task is to point at different object every time (images are shuffled every time you press one)
- we must remember our order of objects in order not to press same button twice
(picture 5)
27
what did the SOPT show?
- adolescents with low score on memory task: implicit positive arousal cognitions predict alcohol use & associated problems
> automated pilot determines whether they drink or not
- adolescents with high score on memory task: explicit expectancies predict alcohol use & associated problems
> the more they expect positive vs negative things, the more they drink
(picture 6)
28
Grenard study - set up
- 145 high school students
- measure:
> spontaneous memory associations (e.g. draft, having fun, ...)
> SOPT
> alcohol/drug use questionnaire
29
what are the results of the Grenard study?
(relationship between W.M., Alcohol-related implicit associations & alcohol use)
- the higher the working memory score, the lower the prediction of these associations of the behavior
> if you have a great working memory, the alcohol-related memory associations do not influence much the frequency of alcohol use
> if you have a low working memory, implicit cognitive processes lead to higher alcohol use
= poor executive functions increase the vulnerabilty to appetitive impulses (e.g. usign substances)
(open-ended associations→ relationship is not biased by test format)
(picture 7)
30
what is more evidence regarding the role of working memory?
- relatively automatic processes predict alcohol/drug use in individuals with relatively low executive control
- similar pattern in other behaviors where the impulsive and reflective processes may clash: aggression, aggression after alcohol, eating while being on a diet, sexual distraction
> the lower the working memory, the more these automatic associations predict the behavior
31
Conclusions on cognitive processes
1. Cognitive Biases are related to substance use
2. Biases in attention, action tendencies,
cognitions (memory)
3. Cognitive control processes (e.g., working
memory) moderator: stronger influence
of automatically activated cognitions in
individuals with relatively weak control
capacity
= the stronger the cognitive moderators, the weaker the influence of implicit cognitions on behaviors
! both within and between subjects (e.g. if tired your own implicit associations will lead you to drink more)
32
2. Cognitive training
- what types of cognitive trainings do we have?
Without addiction-relevant cues:
- General ability training
> working memory training
> general inhibition training
> self-control training
With addiction-relevant cues:
- Cognitive bias modification
> attentional bias retraining
> approach bias retraining
> selective inhibition training (memory)
(picture 8)
33
Training General Abilities
- what is the main idea behind it?
- pros and cons
- if training executive control & working memory, people develop less issues in the future
✓ W.M. training showed positive results in children with ADHD and other problems
X problems with generalization
> people get better in that specific task, but sometimes results do not generalize to similar tasks (close generalization)
> many time results do not translate to real life (far generalization)
> this is a crucial in training research!
34
How can working memory be trained?
- computer screen with different patterns to memorize, and it gets increasingly harder (more squares which order you have to remember)
- in control conditions, you only always get three squares to memorize, so you don't get better
35
what were the results of the studies on the training?
- working memory does become better in training condition, and remains stable in control condition
- even at follw-up the training still holds
→ but does it impact their drinking?
(picture 9)
36
does W.M. training impact participants' drinking?
- more reduction in weekly alcohol use in training group
- no significant difference between control and training group
> when measured with IAT, people with initial stronger implicit associations and lower working memory significantly reduced their alcohol use post training
(picture 10)
37
now, Cognitive Bias Modification (CBM)
- is there causality?
- cognitive bias are related to psychopathology and health problems (cross-sectionally and prospectively)
> do they play a causal role?
> if we directly change bias, would this affect psychopathology/healthy behaviors?
38
So, is training general abilities ultimately effective?
- now many studies have shown that there is no evidence for an effect on addictive behavior
- there are still some promising effects:
> reduced delay discounting in stimulant addiction (e.g. cocaine)
→ e.g. studies getting one dollar now or 5 in a month: after training people get more future-oriented
> increased episodic future thinking in those with low working memory
→ ability to imagine how you'll tackle problem in the future, in a lot of details (e.g. think about steps to take)
- WM training might be helpful in recovery process, when better integrated
> participants get more confidence, which helps in recovery
39
what does the first experimental test regarding CBM come from?
- "proof-of-principle" studies
- participants: moderately anxious students
> one gourp was trained towards threat pictures, and the other away
→ what is the effect on behavior? do we get an effect on stress responses in next task?
Yes!
40
Does it work with alcohol as well?
- they created a Dot probe task
~ at the beginning, probe replaces problem category (alcohol) and control category equally often
~ experimental group: probe replaces control category most (or all) of the time
~ control group: continued assessment / no training / different task
> split design: moderate drinkers and train half towards and half away from alcohol
41
Proof-of-Principle studies vs Clinical trials
Proof-of-Principle studies
- typically conducted in a lab usually with healthy volunteers
- reveal psychological mechanisms underlying human behavior
RCTs trials
- clinical setting with patients
- test the efficacy of an intervention in a clinical sample
- effect of treatment compared with control
- patients are randomly assigned to experimental and control groups
42
Proof of Principle study on alcohol
- binge-drinking students
> recruited for experiment on alcohol and reaction times
> including taste-test (participants may have to taste a bit of alcohol)
- "50-50" control condition: continued assessment
→ half the time dots are behind alcohol, half the time behind neutral stimulus
- experimental condition: trained away from alcohol pictures (after a bit, dots behind only non-acohol)
- they tested the pictures used in training and other similar pictures
(+ check for generalized effects)
43
what are the results of the study?
- on trained pictures, there is a negative attentional bias
> react more quickly to coke compared to beer
- on untrained pictures, there is a decrease in attentional bias, but not significant (close generalization)
- no effect on drinking pattern
(picture 11)
44
so, is single-session attention re-training successful?
- it is possible to train heavy drinking students towards alcohol
- it is possible to train heavy drinking students away from alcohol, but:
> there is no generalization to new stimuli
> no effects on behavior
→ do we just need more training sessions with new pictures every time (to foster generalization)?
45
Schoenmaker → Clinical effectiveness of attentional bias modification training in abstinent alcoholic patients
- main idea
- this is the first small RCTs in alcohol-dependent patients
- selective attention bias to drug cues is hypothesized to have a causal effect on substance abuse, addiction development and maintentance
→ expectations: ABM will decrease the speeded detection of old and new alcohol stimuli, and decrease the difficulty to disengage from those stimuli
> study of ABM on craving, relapse and overall treatment success
46
what does the study explain about Attentional Bias? (AB)
- according to the incentive sensitization theory, AB results from repeated pairing of alcohol cues with direct effects of alcohol, leading to a sensitized reaction to alcohol cues which causes them to become highly salient
- AB possibly related to craving, severity of addiction, poor treatment outcome and relapse following treatment
47
what is the advantage of this study?
→ what three factors does this study focus on?
- first randomized controlled experimental study on the effectiveness of a visual probe based ABM in a clinical sample of (alcohol) dependent patients
- there are three factors hypothesized to increase the effectiveness of ABM interventions:
1. motivating participants to improve training performance and control over own attention
2. presentation of a large number of different stimuli in the training
> generalization towards new stimuli has only been found when there are several stimuli already in the training
3. performing multiple training sessions
> shown to have more profound effects than single training sessions
48
Study layout
- 5 sessions, different alcohol pictures every time
- 43 participants, from treatment center
- 2 conditions: ABM group and control group
- pre-, post- and 3 months follow-up assessment
Measures:
- attentional bias → visual probe task
> fixation cross in middle of computer screen
> simultaneous quick showing of alcoholic and neutral drink
> quick showing of arrow pointing up or down, following either drink
> participants had to report where the arrow was pointing
= faster responses to arrows following alcoholic drink indicate AB
Conditions:
- ABM training condition
> modified visual probe task
> probes replaced neutral pictures every time
- Control condition
> same pictures, but have to be sort out in IAT
- get the same motivating feedback
49
What were the results of the study?
- control group increases in attentional bias
> if you don't do anything and you are in a clinic, attentional bias increases (incubation)
- ABM intervention led to a generalized decrease in difficulty to disengage attention from alcohol-related stimuli (one component of AB)
> generalized effect has important clinical relevance
- ABM had no effect on speeded detection or subejective training
- ABM effect found 3-4 days after intervention
- relapse was delayed by one month in training condition
- discharge was anticipated by one month in training condition
(picture 12)
50
what are two important collateral findings of the study?
- no reduction in speeded detection suggests that participants first scan pictures, before avoiding them
→ this implies that control is needed to disengage attention (reduction in AB not fully automatic)
- ABM affects vulnerability to respond emotionally instead of affecting conscious emotional states (e.g. subjective craving) directly
51
approach bias training
- is approach bias causal?
- what were the effects?
- measured with Irrelevant feature version
- hazardously drinking students
- one group trained to approach alcohol (90% pull alcohol pictures), the other group trained to avoid alcohol (90% push alcohol)
Generalized effects on:
> untrained pictures, after a single session (never found before in attentional training)
> Implicit association test
> behavior (taste-test)
52
Retraining automatic action tendencies study - layout
German clinic with alcoholic patients
4 conditions:
- relevant training (push alcohol away) [training]
> explicit instruction of pushing joystick when alcohol picture, and pulling when soft-drink picture
- irrelevant training (portrait-landscape) [training]
> no explicit instructions (implicit)
> all alcohol pictures were presented in the format they were avoiding, and all softdrink pictures in the format they had to approach
- sham-training
> assessment control (50-50) [control]
- no training [control]
+ abstinence-oriented CBT
+ 1 year follow-up questionnaire
! participants had to correct errors
! 4 20-minute sessions
53
what was this study measuring?
- changes in action tendencies (AAT)
> difference between mean reaction time for pushing and pulling → standardized scores
- alcohol-approach associations (IAT)
- subjective craving (Likert scale)
54
What were the specific results?
- no difference in pretest between training and control group
- no effect in control group
- strong effect in training condition (pushing away alcohol images) even in IAT words association test
! really strong generalization (from picture to word test, it works with both)
→ what is the clinical relevance?
(picture 13)
55
What are the general results?
Brief CBM intervention:
- changed patient's approach bias for alcohol (→ avoidance bias)
> generalized effect across stimuli and measures
- improved treatment outcome one year later
- did not work for attentional retraining (only for close generalization to untrained stimuli)
→ retraining action tendencies has stronger effect than retraining attentional bias
> maybe because action tendencies relate to motivational state
56
what is the clinical relevance of the experiment?
- one-year follow-up:
> 13% less relapse in training condition
= adding CBM to CBT results in (minor) decrease in relapse likelihood
(picture 14)
57
what are the crucial ingredients of CBM?
1- pushing motion is crucial → embodied emotion and cognition theory
2- not movements, but associations between the object and the concepts of approach and avoidance are crucial
= still unanswered whether moving or zooming feature of CBM is most crucial
58
Replication study
(500 participants, training vs no training)
- average: 10% less relapse
- mediation effect
> if you are in training condition, you have lower relapse rates mediated by change in alcohol approach bias
→ this shows that we are tackling relevant mechanism
- only if you have strong approach bias you can see the results (some people don't have approach bias, so training is useless)
(picture 15)
! it only works on top of therapy, it doesn't replace it
59
do biases show just an automatic, unconscious process?
- not really
- the processes are not totally automatic
> e.g. when you do the training, you get to figure out that you are pushing away alcohol
> when you stop impulse, it is a conscious process
> etc...
60
are there neural effects of CBM?
- in the amygdala and nucleus accumbens there is strong reactivity to alcohol before the training, which then decreases
- after training, there is an increase in the salience of the alternative (!)
(picture 16)
61
so, is CBM effective in addiction?
- meta-analysis
- there is an effect of CBM on the bias, but no clinical effects
→ all clinical studies showed effects, so how is this possible?
- this is because they analyzed completely different studies (some looking at causality)
62
what types of study were included in the meta-analysis?
- experimental lab studies:
> causality looked in students, no motivation for treatment
> small temporary effects (if bias changed)
> 20/25 studies
- online RCTs:
> volunteers, aware that might receive intervention
> motivated to change behavior, problem drinkers
> reduction in all conditions (which is good, but there is no specific effect for each training)
- Clinical RCTs:
> if you change bias in people motivated to change their behavior, there is an effect
> consistent add-on effects
> on top of real treatment, you see 10% less relapse (differential effect)
(picture 17)
63
CBM for behavior change in alcohol and smoking addiction: meta-analysis
- 14 studies, +2000 patricipants
- small effect on cognitive bias
- small effect on relapse training
- no significant effecton use
64
after meta-analysis, they did 3 RCTs because in need of more data
- what were the results?
+ discussion
(picture 18)
Training is overall better than no training
- replicated beneficial training effects
- initial expectation: best outcomes in combined training
- probably too short for most (3+3)
- median number of sessions needed in AAT training is 6 (with much variation)
65
for whom does CBM work best?
- addiction has lots of comorbidity with other disorders (anxiety, depression, ...)
→ does CBM work better or worse if you have comorbid internalizing conditions?
- better!
= good group to train is people with other internalizing problems
66
Conclusion CBM
so, is CBM effective?
- CBM is effective as add-on in the treatment of AUD when people are motivated to change
(but have problems succeding due to cue-reactivity, strong bias and strong impulsivity)
67
for whom does CBM not work?
- binge-drinkers who are not motivated to change
- online training with people who want to reduce
68
Comparison with medication
- what does the meta-analysis show?
> medication vs CBM
- NNT=12 (number needed to treat)
> acamprosate (increasing abstinence)
> naltrexone (reducing alcohol consumption)
→ 12 people need to be treated for one person to benefit (small but meaningful effect)
- NNT = 12 also for adding CBM to TAU (treatment as usual)
= CBM has comparable benefit to pharmacological options
! medication sometimes works better for people with whom therapy doesn't work
! CBM only works as an add-on to treatment
69
how did researchers try to improve cognitive training?
1- tried to gamify treatment
> more playful, but not necessarily better
→ participants did not take it seriously
→ was still quite boring
2- more personalized alternative goals
> idea: get motivating alternatives
> started off to reduce smoking
> if you don't drink alcohol, you can drink anything else; but if you don't smoke, what's the alternative?
> better link to CBT (ABC training)
3- personalized learning parameters
4- training based on more reliable assessments
5- training after reactivation (disruption of memory consolidation)
6- add neurostimulation (tDCS, rTMS)
> speeds up the progress, but at the end not really a difference
70
from CBM to ABC
General notes
- CBM is more effective in clinical samples than in PoP studies with non-clinical volunteers
- online CBM less effective than in clinical context
CBM:
- rooted in dual-process models and developed to target distortions in automatic mental processes (linked to associative representations)
> CBM would change dominant associations underlying mental disorders
→ results do not fit well with associative explanations!
71
Why are associative explanations not really valid?
- repeated avoidance of addiction-related stijmuli in the lab does not always translate to changes in behavior
> the effects depend on important moderators (e.g. beliefs about implications of learnt associations→ avoiding alcohol will help me refrain from drinking)
- approach-avoidance effects can be based on verbal instructions rather than extensive training
- change requires awareness of relevant contingencies
~ these findings are not in line with associative account, but are more in line with inferential account
72
Conclusion of ABC training
- ABC-training is theory-based new variety of cognitive training aimed at targeting actions based on automatic inferences (expectancies) rather than associations
- Promising initial results in volunteers
- Next step: RCTs in clinical samples
73
what do these findings show?
- CBM works not by replacing associations (alcohol approach to alcohol avoidance), but through changes in propositional representations containing information about how concepts are causally related (inferrence)
- CBM invokes propostitions about the contingencies between stimuli (e.g. alcohol), responses (e.g. avoidance) and outcomes (e.g. positive effects), which translate to behavior
74
what do these findings lead to?
Inferential perspective
> behavior reflects goal-driven inferences that are learnt and evoked on the basis of beliefs about their instrumental relevance to people's goals
(behavior is shaped by people’s goals and what they believe will help them reach those goals)
> contextual cues might automatically trigger approach tendencies in heavy drinkers because of expected desirable outcomes
→ CBM might be more effective if designed to automatize adaptive goal-directed predictions
1> e.g. during alcohol-avoidance training, participants may learn new inferences about action tendencies (e.g. alcohol avoidance) that would result in valued outcomes (e.g. abstinence or recovery)
2> the inferences can facilitate the implementation of similar actions (e.g. avoid) when participants are confronted with similar cues
3> when well practiced (automatized) they will be more readily available and translated into behavior
75
what is the best new development in treatment?
- ABC training
- originally: dual process models
> idea: you have impulsive models (attention, approach bias, working memory, ...) and if executive control capacity is not strong enough, then automatic mechanism kicks in
- now, we add unique component of training device → ABC training
76
what is ABC training?
- it is an alternative account mechanism
> automatic inferential propositional mechanism (rather than associations)
- PoP studies showed:
> some CBM effects can be generated by instruction only (proof contrary to associative account)
> CBM effects occur only with conscious awareness (different from idea that we only have automatic associative mechanisms)
→ therefore, it is important to include consequence in the training
> background is automatic inferences
> brain is continuously predicting whatever is next
> we want to create another scenario where we can achieve something that is not drinking (...)
77
what are the studies ABC training is based on?
1. PoP with irrelevant feature joystick training
(a) participants in pushing the alcohol picture condition were faster to associate alcohol-related words to avoidance in IAT
(b) heavy drinking students that had been successfully training to avoid alcohol drank less in a subsequent taste test, compared to other students who were trained to approach alcohol
2. patients trained to avoid alcohol on top of standard treatment
(a) 10% less relapse after one year
3. CBM effects are stronger when done during detox
78
what are the steps of ABC training?
- participants choose personally-relevant event
A: antecedent context
> e.g. coming home stressed
B: behavioral alternatives relevant in that context
> e.g. go for a walk rather than smoke/drink (especially important for other addictions than alcohol, where there is universal alternative)
C: consequences
> e.g. better health, save money, ...
79
Change 1: goal-relevant alternative behavior (B)
- in CBM, people are trained to choose between alcoholic/non-alcoholic drinks (relevant choice)
- however, with other substances it is not so straighforward (e.g. smoking vs ?)
> visually matched alternatives were used (e.g. holding a pen)
> increased effectiveness when behavioral alternatives were personalized (e.g. running to reduce stress instead of smoking)
! include a goal-relevant behavioral choice
→ behaviors leading to personally relevant desirable outcomes should be trained
80
Change 2: Personally relevant consequences (C)
- behavioral choices have consequences, and CBM might require learning those consequences
> patient should be able to experience the effectiveness of the alternative behavior to accomplish his/her goals
> CBM more effective if including real-life goal-relevant consequences
Experiment
- participants should maximize the health of the avatar (fridge choice)
> approaching unhealthy foods → negative consequences
= participants who learnt consequences were able to apply learnt behavior in real life, compared to standard CBM training or control condition
! adding a relevant consequence might significantly improve the effectiveness of CBM
! consequences should be tailored to patient's own goals
81
Change 3: Personally relevant antecedent contexts (A)
- Participants completed the CBM task in a simulated real-life context in which they were standing in front of a refrigerator
> incorporating these real-life context cues (A: refrigerator) in association with the behavioral choice (B) and the action consequences (C) might facilitate transfer to real life
- real-life antecedents can be simple stimulus or high-risk situation
! multiple relevant antecedent contexts (and relevant behavioral alternatives) can be identified and trained as part of the intervention
82
ABC - implementation of all changes
To summarize, our proposed ABC training represents a novel, theory-based variety of CBM that involves training goal-relevant behavioral choices (B) triggered by antecedent cues (A) and followed by positive or negative action consequences (C) for the pursuit of specific goals
- ABC training aims to automatize behavioral choices relevant to an individual’s goals in specific contexts
83
ABC vs new therapies
- ABC training might enhance effectiveness of CBT and other treatments
> it combines these two approaches
> it targets automatization of adaptive inferences and related behaviors
→ which may be crucial in revising habitual behavior that is difficult to change
1. ABC training involves practice that aims to automatize behavioral choices and to reduce the amount of effortful control required
> patients have difficulties doing their homework
> can be delivered online (lots of premature drop-out)
2. Compared with traditional CBM (often experienced as meaningless by patients) personalized ABC training would seem a more meaningful complement to therapy
84
ABC vs CBT
- ABC is directly related to CBT
> both focus on patient's personal long-term goals
> original CBM was often found to be rather meaningless to patients and unrelated to their therapeutic goals
- ABC adds systematic and controllable personalized training rather than homework people have difficulty with
85
example of ABC training
- what are the three phases?
- created avatars and participants pick one
- they are e.g. in a stressful situation and have to choose whether to take a walk or drink
- avatars then display action and consequences
1. forced choice to learn consequences (continued in sham training)
2. open choice with consequences
3. speeded open choice with consequences, to foster automatization
86
what have PoP studies shown about ABC training?
Prolific randomized over ABC training vs sham-control
! - negative alcohol expectancies after ABC training increased
- stronger increase in self-efficacy after ABC
- heavier drinkers reduced their drinking more after ABC training
87
abstinence challenge
- extra evidence
ABC vs CBM (push away alcohol) vs sham-CBM as add-on to NoThanks! challenge (dry january)
→ ABC does better
- stronger increase successful in abstinence in ABC training group
- no differential effect in drinking during two weeks after challenge (just in january, all effects gone by february)
→ supports goal but not long-term
88
Overall conclusions
- Cognitive Biases play a role in addiction
- Cognitive training to change biases can be
of use in the treatment of addictions:
> General ability training some promise, but long; probably only inpatient recovery
> CBM, effective as add-on to inpatient abstinence-oriented treatment for alcohol use disorders (small but reliable effect, in clinical guidelines now in Germany; Aus; NL)
- Room for improvement: theory-based ABC-training
! does not imply that nonvoluntary processes play no role in addiction
