Subliminal Affective Priming of Implicit Motivations

Winkielman et al (2005): Summary of experimental results

In a widely cited and theoretically provocative paper on ‘unconscious affect’, Winkielman, Berridge & Willbarger (2005) investigated  the influence of subliminally presented happy or  angry face primes on (i) the amounts of pouring and consumption of a sweet beverage, (ii) the perception of the beverage’s value and (iii) reports of conscious feelings during the experiment.

Participants first rated their level of thirst and hunger. Subliminal expressions of the same valence (first neutral and then either happy or angry) were then imperceptibly embedded in an eight trail  “gender-classification” task, in which participants identified the gender of easily visible backward masking neutral faces.

Following this task, the participants poured, consumed and rated a sweet beverage (drinking task) and rated their mood and arousal level (mood task). (See paradigm diagram below). Results indicated there was a strong effect of subliminal primes on beverage consumption – but only for thirsty participants; there was no happy vs angry effect of subliminal primes on nonthirsty participants – or hungry participants. Thirsty participants poured more than twice the amount of the beverage after happy primes than angry primes, and after pouring, they drank 171% more of the beverage after happy primes than after angry primes. Despite the large impact of the affective primes on thirsty participants, there were no observed differences in mood or arousal ratings before and after priming: subjective experience did not change.

Follow up experiments showed that subliminal primes also strongly affected  participants willingness to pay for a drink and their ratings of how much they wanted to drink. Thirsty participants who rated the drink after priming were willing to pay three times the price after happy primes than after angry primes. A similar result was found for ratings by thirsty – but not nonthirsty participants of how much they wanted to consume more of the drink after sampling it. Once again, in these experients there were no change in subjective experience, even immediately after priming, despite an opportunity to report on 20 nuances of mood.  Interestingly, in none of the experiments was there an effect of priming on how much the participants reported liking beverage immediately after consumption on ‘how delicious is the drink’ ratings, although thirst level did modulate baseline liking ratings, prior to exposure to the primes.

Authors’ interpretation of the data

According to Winkielman, Berridge & Willbarger, the results of their study can be explained by an ‘incentive value account’ from the field of biopsychology: basic affective stimuli act via unconscious, automatic processes to directly alter the assessment of the drink’s ‘hedonic value and motivated behaviour towards its consumption’. Thirst enhanced the initial incentive value of the drink  via physiological alliesthesia (scaling of how much the drink is liked as a function of need state/thirst), and this effect is reflected  in thirsty participants’ higher baseline ratings of the deliciousness of the drink. This elevated baseline value was then momentarily multiplied up or down by subliminal happy and angry expressions unconsciously. “For participants who were not thirsty,” they state, “the initial incentive value of the drink was closer to zero, so the subliminal multiplication of evaluative reaction by emotional expressions was negligible” (p.132). The changes in incentive value, they propose, led to differences in consumption behaviour and ratings of price and wanting.

A challenge to the Winkielman, Berridge & Willbarger explanation

The authors state their incentive value explanation “is consistent with the finding that priming did not simultaneously influence ratings of more standard hedonic and sensory dimensions of the beverage, such as deliciousness or sweetness” (p. 132). But the authors do not explain why. Incentive value is (in large part) determined by the hedonic (pleasure) of the drink, so why was this not affected by the subliminal primes if they raised or depressed the drink’s incentive value?

An alternative explanation: ‘Wanting’ vs ‘Liking’

Biopsychological research has revealed two consecutive and functionally dissociable phases or aspects of implicit motivational processes: 1. A motivation phase during which the organism works instrumentally to attain a reward or avoid a punishment. This motivated action could be as simple as taking steps to a water hole and starting to drink, or as complex as socially coordinated hunting of an elusive prey in a tree canopy. This phase of motivation is mediated by the mesolimbic dopamine system. 2. A consummation phase in which there is an evaluation of the hedonic (pleasure-pain or ‘affective’) qualities or ‘goodness’ that accompany the consumption of – or contact with – the incentive or disincentive (Berridge, 1996). This system is mediated by the limbic frontal cortex (particularly the orbitofrontal cortex). Berridge has labelled these dissociable aspects of motivation wanting and liking respectively.

An explanation for the Winkielman et al study, consistent with this motivational ‘wanting’ vs consummation ‘liking’ distinction, is the following: The subliminal primes didn’t affect the drink’s hedonic value for thirsty participants – i.e. how much it was liked. If they did, the happy face primes would have caused higher ratings of ‘deliciousness’  than angry faces, but there was no difference in experience of the hedonic qualities of the drink. Rather, when the drink was encountered by subliminally primed participants who were thirsty, their mesolimbic systems may have reacted with greater activation leading them to want the drink more, give it higher ‘willingness to pay’ and ‘wanting to drink more’ ratings as well as more pouring and drinking.

Experimental prediction

Evidence strongly suggests that extraverts have a greater capacity for mesolimbic dopamine ‘wanting’ system activation – whether naturally stimulated by incentive signals or artificially induced through dopamine agonists (Depue and Collins, 1999). People high in extroversion respond with greater wanting to the same incentives than people low in extraversion.

Experiment 1

A clear and novel experimental prediction is, therefore, that the subliminal priming effects of happy vs angry faces will be significantly greater for extraverts than for introverts. There is no difference in ‘liking’ responses between extraverts and introverts, so the two systems can be dissociated with this population.

Endorsement of this intepretation and proposed experiment by the paper authors

I have suggested my interpretation of the data and the experimental prediction to the first author, Professor Winkielman (a cognitive psychologist at the University of California, San Diego),  and the second author, Professor Berridge (a biopsychologist at the University of Michigan), and they both agree that the interpretation is a good one, and the prediction that follows from it is worth doing.

Kent Berridge (e-mail correspondence):

I completely agree with your interpretation that the subliminal happy faces may have primed the mesolimbic DA ‘wanting’ system into greater reactivity.  Then, as you say, when the drink was encountered by subliminally primed participants who were thirsty, their mesolimbic systems may have reacted with greater activation leading them give it higher money value ratings and to pour and drink more….It seems quite possible that your prediction of greater subliminal priming effects in extraverts is correct, at least to the extent that extraversion truly does reflect greater mesolimbic DA activity, which I agree is plausible.  That would be a great experiment for someone to run.

Pi Winkielman (e-mail correspondence):

Hi Mark,

Your prediction makes sense to me too.

Additional experiments

In addition to the extraversion/introversion experiment, the following experiments will also contribute to our understanding of processes underlying implicit motives more generally.

Experiment 2. ‘Lower Order’ Motivational Specificity

Based on what is known about the specificity of implicit motives including hunger and thirst, a prediction is that affective priming of food will not be influenced by thirst, but by hunger. Consistent with this idea, the priming effects obtained in the Winkielman et al. (2005) study did not depend on depend on hunger, suggesting motivational specificity (Rolls, 2000).

Experiment 3. ‘Higher Order’ Implicit Motivational Specificity

Biopsychological research suggests that implicit motives include not just basic physiological needs such as food and water, but social affiliation and dominance/power motives shared by all higher mammals. Individuals differ on affiliation and dominance implicit motivations (Schultheiss, 2008). Our model of the motivational mechanism predicts  that ‘high affiliation’ individuals will be significantly affected by subliminal affective primes for affiliation related incentives, but not not dominance related incentives, while ‘high dominance/power’ individuals will be  affected by subliminal affective primes for dominance related incentives, but not affiliation related incentives. An affiliation incentive could be a framed photograph of friends of family; a dominance incentive could be a certificate giving a competitive privilege for having participated in the experiment.  A prediction would also be that there should be no effect of subliminal primes on ‘liking’ ratings of those incentives, only wanting them. A strong prediction would also be that there will be no effect of subliminal priming on ‘liking’ ratings of photographs of facial expressions either expressing affiliation of submission (or the opposite).

Experiment 4. Implicit vs Explicit Motivation Mechanisms

According to the theory developed by Schultheiss (2001; 2008), we can be motivated by both implicit motives and explicit motives under the control of executive processes. Implicit motives respond to nonverbal cues and incentives and have an impact on behaviour and processes that are not controlled by, or accessible to, the individual‘s verbally represented goals or self-concept. Implicit motives are revealed using non-declarative measures that include physiological autonomic responses (e.g., changes in blood pressure, hormone release, muscle tone), and acquisition of new stimulus-stimulus associations and goal-directed behaviors through Pavlovian and instrumental learning of the incentive-motivation type that is explored in the Winkielman and colleagues (2005) paper. Explicit motives, by contrast, respond to verbal symbolic cues and influence measures that tap into a person‘s verbally represented sense of self and the attitudes, judgments, decisions, and goals that are associated with it. Declarative measures for this include valence judgments, decision making behavior, assessments of self-regulatory control, and reports of personal goals. Implicit motives are generally inaccessible to introspection, while explicit motives are accessible to introspection and conscious control.

The mechanisms by which implicit motives interface with explicit motives is not known. The ‘wanting’ (mesolimbic dopamine system) and ‘liking’ (orbitofrontal system) ‘incentive motivation’ system may be confined to implicit motives (investigated in experiments 1 and 2) or it is possible that the same ‘wanting’ and ‘liking’ systems play a role in both implicit and explicit motives. We can also conduct an exploratory study, in which we set up an explicit motivation (with a goal that is not related to implicit motivations) – such as the goal of solving an intellectual puzzle or learning a list of words to later be tested. The question here is whether subliminal happy vs angry faces will have an effect on the desire to solve the puzzle or learn the list of words. We may find no effect on these types of explicit (executive function) goals, showing a dissociation of implicit vs explicit processes, or we may find that there is a priming effect, suggesting a common underlying motivational mechanism.


Wider significance of the proposed experiments to the field

First, the experiments explore the provocative thesis that emotional/affective processes can be unconscious – that affective reactions can be subliminally triggered, change behavior dramatically, yet still remain inaccessible to introspection. This is an interesting philosophical as well as psychological issue.

Second, following Winkielman et al, these studies connect the existing cognitive psychological literature on affect  with contemporary biopsychological research and theory in which affect and motivated behaviour combine in a single explanatory ‘incentive-motivation’ framework. Typically biosychological accounts of human behaviour and cognitive psychological accounts of human behaviour do not communicate with each other.

Third, our proposed reasearch explores mechanisms underlying incentive learning based implicit vs explicit motives. Goal directed behaviour is a growing area of research in cognitive psychology, and the ‘implicit’ vs ‘explicit’ framework from biopsychology, integrating both learning theory in comparative psychology and executive control theories in cognitive psychology, provides a fruitful approach to studying goal directed behaviours.


Apparatus for the experiments

  • The subliminal priming paradigm is freely available online on Professor Winkielman’s website.
  • E-prime experimental design software
  • The IPIP-NEO ‘Big 5’ personality inventory for a valid measure of extraversion.
  • The PSE implicit motive inventory – a valid and reliable measure of implicit motives (Schultheiss, 2008).


Ashton Smith, M.  (2010). Functions of Unconscious and Conscious Affect and Emotion in the Regulation of Implicit and Explicit Motivated Behaviour: A Dual Process Account. In Didem Gökçay and Gülsen Yıldırım (Eds.), Affective Computing and Interaction: Psychological, Cognitive and Neuroscientific Perspectives, IGI Global.

Berridge, K. C. (1996). Food reward: Brain substrates of wanting and liking. Neuroscience and Biobehavioral Reviews, 20, 1-25.

Depue, R. A., & Collins, P. F. (1999). Neurobiology of the structure of personality: Dopamine, facilitation of incentive motivation, and extraversion. Behavioral and Brain Sciences, 22, 491-569.

Rolls, E. T. (2000). The orbitofrontal cortex and reward. Cerebral Cortex, 10(3), 284-294.

Schultheiss, O. C. (Ed.). (2001). An information processing account of implicit motive arousal. ( Vol. 12). Greenwich, CT: JAI Press.

Schultheiss, O. C. (2008). Implicit motives. In O. P. John, R. W. Robins & L. A. Pervin (Eds.), Handbook of Personality: Theory and Research (3 ed., pp. 603-33). New York: Guilford.

Winkielman, P., Berridge, K. C., & & Wilbarger, J. (2005). Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality and Social Psychology Bulletin, 31, 121-135.