Emotional salience of the image component facilitates recall of the text of cigarette warning labels

Emotional salience of the image component facilitates recall of the text of cigarette warning labels Abstract Background Graphic warning labels (GWLs) on cigarette packages, that combine textual warnings with emotionally salient images depicting the adverse health consequences of smoking, have been adopted in most European countries. In the US, the courts deemed the evidence justifying the inclusion of emotionally salient images in GWLs insufficient and put the implementation on hold. We conducted a controlled experimental study examining the effect of emotional salience of GWL’s images on the recall of their text component. Methods Seventy-three non-treatment-seeking daily smokers received cigarette packs carrying GWLs for a period of 4 weeks. Participants were randomly assigned to receive packs with GWLs previously rated as eliciting high or low level of emotional reaction (ER). The two conditions differed in respect to images but used the same textual warning statements. Participants’ recognition of GWL images and statements were tested separately at baseline and again after the 4-week repetitive exposure. Results Textual warning statements were recognized more accurately when paired with high ER images than when paired with low ER images, both at baseline and after daily exposure to GWLs over a 4-week period. Conclusion The results suggest that emotional salience of GWLs facilitates cognitive processing of the textual warnings, resulting in better remembering of the information about the health hazards of smoking. Thus, high emotional salience of the pictorial component of GWLs is essential for their overall effectiveness. Introduction Cigarette packages sold in the US carry a brief text-only Surgeon General’s warning messages that occupy less than a quarter of one of packages’ sides. There are four different messages that are rotated quarterly. This design has not changed significantly since it was introduced in the 1965 Federal Cigarette Labeling and Advertising Act. Although the current labels are considered to be ineffective, there is no consensus among stakeholders about the ways to improve them.1 The Family Smoking Prevention and Tobacco Control Act of 2009 (TCA) required graphic warning labels (GWLs) that combined color graphics and textual warning statements to cover 50% of the front and rear panels of the cigarette package. TCA specified the nine textual warnings to be included in the GWLs, leaving the graphic component to the discretion of the Food and Drug Administration (FDA).2 Tobacco companies sued and were granted a permanent injunction from the implementation of the labels by the District of Columbia court. This ruling was affirmed by a US Court of Appeals, which ‘vacated’ the graphic labels requirement alltogether and ‘remanded’ the Final Rule back to the FDA.3 The district court highlighted the lack of justification for the high emotional salience of the images selected by the FDA, while the appellate court was more concerned with the lack of ‘direct evidence showing that enacting the proposed graphic warnings will accomplish the agency’s stated objective of reducing smoking rates’.3 This reversal spurred a regulatory debate on GWLs in the US that continues to this day. The debate revolves around the relative benefits of the graphic and textual components of GWLs to consumers vs. the risk to society from the infringement by such labels on the tobacco companies constitutional right to free speech. While a large body of literature examined the population-level effects of GWLs, less attention has been given to the cognitive mechanisms that link the perceived information with memory and ultimately objective measures of smoking.4 To this end, an experimental demonstration of whether the emotional salience of the graphic component of GWLs aids the retention of the message in the accompanying textual warning is essential to finding an optimal way to improve their effectiveness with minimum infringement on commercial free speech. Cognitive neuroscience studies find that emotional salience of still or moving images increases their memorability.5,6 However, studies disagree about the benefits of images’ emotional salience to the retention of the information they are paired with.7,8 In the context of warning labels, if GWLs are akin to a figure with a caption, the question is whether a more salient image facilitates or detracts from processing the caption. Recognition accuracy is commonly used as an intermediate outcome measure in communication research in general, and to test the efficacy of health messages in particular.9 Although prior studies show that emotion-provoking GWLs are more memorable,5,10,11 it is still unknown whether increased emotional salience of the accompanying graphic component actually improves remembering the textual component. We hypothesized that images that evoke greater emotional reaction (ER) will facilitate remembering the textual component of the GWLs, and that these effects will be reflected in the recognition accuracy of each of the GWLs components (i.e. text statement vs. image). Methods Participants Seventy-three non-treatment seeking smokers (31 females, 31 Caucasian, 28 African American, 8 Asian and 6 mixed race. Five African American and 4 mixed race participants also self-identified as Hispanic) were recruited through advertising between 2014 and 2016. Participants were 30.89 ±10.30 years old, reported smoking 14.18 ±9.23 cigarettes per day, and had Fagerstrom Test for Nicotine Dependence (FTND) scores of 5.55 ±2.89 (mean ±SD). We excluded individuals who smoked less than 5 cigarettes per day in the last 7 days or had urine cotinine levels undetectable by the qualitative enzyme-linked immunosorbent assay test (Reditest Smoke Cassette, Redwood Toxicology Labs, Santa Rosa, CA, USA). Pregnant or breast-feeding women and smokers currently enrolled in a smoking cessation trial were also excluded. All participants gave written informed consent to participate. The study protocol was approved by the University of Pennsylvania Institutional Review Board. Stimuli Nine TCA-mandated warning statements about the harms of cigarette smoking12 served as the text component of the GWLs. The FDA paired each of the nine statements with different smoking-related graphic images, resulting in 36 GWLs. Therefore some GWLs carried the same textual statement but all differed on the graphic image. The FDA commissioned an online survey of 18 000 respondents who evaluated the emotional salience of these 36 composite GWLs using an emotional reaction (ER) scale.13 Based on the ER score of each GWL image, we selected GWLs with the highest ER score for each textual warning statement to form the ‘high ER GWL condition’, and GWLs with the lowest ER scores for each textual warning statement to form the ‘low ER GWL condition’. Thus, high and low ER conditions shared nine identical textual warning statements while differing on the level of emotional salience of the accompanying graphics (figure 1). The ER scores are significantly different between the high and low ER GWL conditions (P < 0.001). Figure 1 View largeDownload slide An illustration of the experimental procedure Figure 1 View largeDownload slide An illustration of the experimental procedure Procedure Study candidates who passed the phone pre-screening were invited to an in-person screening that included informed consent followed psychiatric history and urine drug screens for cotinine and illicit drugs. Once their eligibility was confirmed, participants were randomly assigned to either the high or low ER GWL condition, in which they were exposed to high or low ER GWLs affixed to the packs of their preferred brand of cigarettes(.) These were supplied by the study team, over a 4-week period. The number of cigarette packs per week was determined by self-reported daily cigarette consumption at screening and baseline assessments. Before distributing the packs to participants, research staff removed the package cellophane and applied the self-adhesive label to cover the top 50% of the front and back panels of the packs, in accordance with the proposed FDA guidelines.2 All the participants attended all four sessions, as they picked up their weekly supply of experimental cigarette packs at these sessions. Participants were encouraged to return unfinished packs, although none brought any back. The average number of packs provided per week did not differ between the high and low ER conditions (P = 0.735). Before and after the 4-week exposure, subjects attended an encoding session, in which they viewed either high or low GWLs, mixed with control stimuli. The control stimuli consisted of nine neutral images from the International Affective Picture System (University of Florida, Gainesville, FL, USA) with text statements unrelated to smoking to match the length and formation of GWLs. Thirty minutes after each encoding, a test of memory for graphic images and text warnings [i.e. Recognition Memory Task, (RMT)] was administered. The RMT consisted of 36 unique text or image stimuli: nine previously seen textual warning statements extracted from the GWLs, nine previously unseen made-up statements related to smoking, nine images extracted from high ER GWLs and nine images extracted from low ER GWLs (figure 1). Thus, for the ‘high ER GWL group’, targets were the nine previously seen textual warnings and nine images from the high ER GWLs, and foils were the nine unseen textual warnings and nine images from the low ER GWLs. Conversely, for the ‘low ER GWL group’, targets were the nine seen textual warnings and nine images from the low ER GWLs, and foils were nine unseen textual warnings and nine images from the high ER GWLs. In each screen, subjects were prompted to respond ‘Yes’ or ‘No’ to the question ‘Have you seen this?’. RMT performance was calculated as percentage of correctly recognized GWL textual warning statements or images respectively, resulting in two scores: Accurate-Text and Accurate-Image. Statistical analysis Statistical analysis was performed using the IBM Statistical Package for the Social Sciences (IBM SPSS version 22). Independent t-tests were used to compare the demographic characteristics of high and low ER GWL groups. To compare the memory of GWL text warnings in two groups at two time points, a two-way repeated-measures analysis of variance (ANOVA) was applied on Accurate-Text using session (encoding session 1 vs. encoding session 2) as the within-subject variable and group (high ER GWL vs. low ER GWL) as the between-subject variable. The same test was repeated on Accurate-Image. Furthermore, to compare memory of GWL text and image within each group across time, a two-way repeated-measures ANOVA was applied in either high or low ER GWL group using component (text vs. image) and session (encoding session 1 vs. encoding session 2) as within-subject variables. Results There were no differences in gender, race, age, education, FTND scores and cigarettes per day between the high and low ER groups at baseline (table 1). Table 1 Subjects characteristics in high and low GWL groups (mean ± SD) GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  Table 1 Subjects characteristics in high and low GWL groups (mean ± SD) GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  ANOVA analysis on the recognition performance of the text warnings revealed a significant effect of group [F(1, 58) = 4.90, P = 0.031] but not session [F(1, 58) = 0.046, P = 0.832] or group × session interaction [F(1, 58) = 0.046, P = 0.832]. That is, the high ER GWL group was better at recognizing text warnings than the low ER GWL group across time (figure 2a). ANOVA analysis of recognition performance of the images revealed a significant effect of group [F(1, 58) = 8.925, P = 0.004], session [F(1, 58) = 13.032, P = 0.001] and group × session interaction [F(1, 58) = 8.925, P = 0.004]. That is, while the high ER GWL group remained better at recognizing images than the low ER GWL group across time, the low ER GWL group improved after the 4-week exposure (figure 2b). In the high ER GWL group ANOVA revealed a significant effect of component [F(1, 29) = 13.486, P = 0.001] but not of session [F(1, 29) = 0.083, P = 0.776] or component × session interaction [F(1, 29) = 0.335, P = 0.567]. That is, subjects were better at recognizing images than text in the high ER GWL group at both timepoints (figure 3a). ANOVA analysis in the low ER GWL group revealed a significant effect of component [F(1, 29) = 18.502, P < 0.001], session [F(1, 29) = 6.032, P = 0.020] and component × session interaction [F(1, 29) = 7.922, P = 0.009]. That is, subjects were better at recognizing text both at baseline and after a 4-week exposure, while their ability to recognize images improved over time (figure 3b). Figure 2 View largeDownload slide (a) Recognition accuracy of the GWL text was significantly higher in the high ER group than in the low ER group across time. (b) Recognition accuracy of the GWL image remained high in the high ER GWL group across time, while it improved in the low ER GWL group Figure 2 View largeDownload slide (a) Recognition accuracy of the GWL text was significantly higher in the high ER group than in the low ER group across time. (b) Recognition accuracy of the GWL image remained high in the high ER GWL group across time, while it improved in the low ER GWL group Figure 3 View largeDownload slide (a) In the high ER group, images of GWLs were better recognized than texts of GWLs over time. (b) In contrast, the low ER group images of GWLs were less recognized than images of GWLs in Session 1, and the gap become narrower after repeated exposure Figure 3 View largeDownload slide (a) In the high ER group, images of GWLs were better recognized than texts of GWLs over time. (b) In contrast, the low ER group images of GWLs were less recognized than images of GWLs in Session 1, and the gap become narrower after repeated exposure Discussion We found that textual warning statements were better remembered when they were presented together with images that evoked stronger emotion, i.e. high ER GWLs. This finding was unchanged after 4 weeks of daily exposure. The image component of the high ER labels was better remembered at both time points, though the gap between the memory of high and low ER images narrowed at the 4-weeks. Finally, in the low ER group, the text component was better remembered than the image, with the difference diminishing after 4 weeks of exposure to the low ER GWLs. To the best of our knowledge, this is the first controlled experimental demonstration of the emotional salience of the image component of GWL enhancing the memorability of its textual component. The finding that images of the high ER GWLs were much better recognized than those of the low ER GWLs is consistent with cognitive research finding that higher emotional salience enhances memorability of negatively valenced images.6,14,15 Similar observations have been made with words16 and narratives.17 However, there is no consensus on whether emotion enhances or inhibits memory of concurrent verbal information.18,19 Some theories predict that information co-occurring with emotion will be better remembered, while others propose that the emotion would divert attention and processing resources away from the co-occurring verbal or textual information.8,20 Our results concur with the former theory, as we demonstrate that emotional salience not only serves to enhance the memory of images, but also increase the likelihood that viewers will remember the text warnings associated with these images in the composite GWLs. This finding suggests a facilitation effect between the emotional salience of images and processing of textual warning statements. A number of studies found that the negative emotional content of stimuli increases the likelihood of memorizing a contextual detail of the stimulus.16,21 For example, subjects are more likely to correctly report the color of the font in which a negative word was presented than the color of the font of a neutral word.16 Similarly, emotional congruence between images and text may also have an important role in enhancing the memory of a warning statement. Since eight out of nine TCA-mandated text warnings describe extreme health consequences of smoking such as cancer and death, images that are proportionally negative emotionally, are more congruent with those warning statements than the less negative or positive images would be. Indeed, prior studies show that congruent information items are processed faster and ‘better’, as shown by the shorter reaction time and more accurate task performance.22 However, the likely need for continuous development of new GWLs and the possible differential impact of positive vs. negative valence of text and images on the audience warrant future studies that would manipulate the emotional valence and congruence of the GWL text and images.23,24 The enhancing effect of emotion on memory has been attributed to an increase in the interaction between the amgydala and the hippocampus during the encoding and consolidation stages of processing.14,25 The critical role of the amgydala in information processing has been confirmed by studies of emotional memory in patients with amygdalar lesions, who showed impaired recognition of emotional words, pictures and stories.26 In the context of GWLs, this interpretation is consistent with the previously reported greater amygdala response to high ER vs. low ER GWLs.5 According to the Tobacco Labeling Resource Centre and the International Tobacco Control Policy Evaluation Project (The ITC Project), 105 countries or jurisdictions worldwide, including most European countries, require GWLs on cigarette packages although not all have implemented them. Cross-sectional population studies reported positive effects of GWLs on smoking-related outcomes, such as reduced smoking rates in Canada, increased awareness of smoking harm on children and pregnant women in Singapore, increased intention to quit in Thailand, Brazil and Australia and reduced smoking initiation and progression in adolescents in Greece.27 GWL effectiveness is thought to increase with size.28 Thus, countries require GWLs to cover between 30% and 90% of the pack. Nepal and Vanuatu require the largest GWL in the world at 90% of the package, front and back, followed by India (85%) and Thailand (85%). Although the use of graphic images is also recommended,1,28 the knowledge of which aspects of the image should be more heavily weighted, or how image and textual components interact is still limited. In the present study, we showed that emotionally salient images are associated with greater cognitive processing of textual warning messages, suggesting that including images with greater emotional salience could increase the effectiveness of GWLs. Commercial free speech is given greater protection in the USA than in Europe29 setting a higher bar for the evidence required to justify any potential encroachment on the first (free speech) Amendment to the US Constitution. An additional distinction between the European and US medico-legal environment is the degree of emphasis on individual autonomy in making health-related decisions. Thus, an argument that the purpose of cigarette warning labels is to inform consumers rather than to deter them from smoking outright is more likely to be accepted in the US.30 The available data were sufficient for adoption of GWL legislation in many regulatory environments including the European, South East Asian and Americas regions. However, few of the available international studies used objective outcomes or psychological and neuroscientific experimental methodology, creating a dearth of what would be considered incontrovertible objective evidence by the US legal systems. This higher bar is reflected in the legal opinion currently governing the GWL in the USA, that dismissed the link that FDA made between the attitudes and intentions about quitting reported in international studies and the actual smoking as a ‘mere speculation’.3 Indeed, prospective studies using a randomized controlled trial design are better suited to meeting the higher bar set by the courts. Another factor in addressing the gap in tobacco research pointed out by the courts is using reliable outcome measures. The majority of studies of the effects of GWLs on smoking have relied on self-reported outcomes such as changes in attitudes and intentions, quit attempts(,) or self-reported cigarettes smoked per day, rather than direct and objective outcomes such as continuous monitoring of smoking itself or of the levels of nicotine metabolites in bodily fluids.31 A smaller number of trials, including our present study, relied on objective but indirect experimental measures such as recall, eye tracking and behavioral or brain correlates of cue-reactivity.32–34 Memory has been shown to be a useful proxy of subsequent behavior change, suggesting that better memory of the textual messages could be an indicator of subsequent behavioral changes.35,36 Specifically for tobacco, Leas et al reported that better recall of anti-tobacco video ads predicted smoking cessation at follow-up.37 By experimentally demonstrating that emotionally salient graphic component of GWLs enhance the memorability of warning messages, the present study provides an insight about the features of an effective GWL. However, our findings do not directly link the messages’ memorability and their effectiveness at reducing smoking. Establishing such direct causal relationship described in the opinion of the US Court of Appeals would require a prospective randomized-control trial using laboratory measures of nicotine intake as one of its outcome measures.3,32,38 Furthermore, the adoption of GWLs does not signal the end of the need for ongoing revision and updating of GWLs. While our results show no reduction in recall of images or text after 4 weeks of exposure, other studies indicate that GWLs effects may be subject to desensitization over longer periods of exposure.39 Our study offers an approach for testing the GWL components in a controlled fashion that could be used in the ongoing process of refining existing GWLs or evaluating new ones. Future studies may consider using even more objective variables such as neuroimaging, or a longer exposure period, with the same platform. Lastly, our findings may generalize to other public health topics where product labeling exists or is being considered, such as food and alcohol. For example, alcohol consumption in Europe is the highest in the world and there is a discussion of reducing harmful use of alcohol by introducing health labels on alcohol products.40 Indeed, labels provide a unique opportunity for health authorities to communicate the alcohol-related health risks at points of sale and consumption. Since data on effective alcohol labeling for public health purposes are limited or lacking, data from studies like ours could provide the initial basis for developing and testing alcohol warning labels. Conclusions We demonstrated that emotional salience of the image component of GWLs enhances the processing of its textual component, resulting in better memory of the warning message. This finding suggests that highly emotionally salient pictures increase the GWLs effectiveness. Our study offers an approach to the evaluation of GWL effectiveness and contributes to a better understanding of cognitive processing of health warnings that may be applied to the development of health warning labels in other public health domains. Funding This work was supported by the National Institute on Drug Abuse (R01DA036028) and National Institute of Child Health and Human Development (K99HD084746). Conflicts of interest: None declared. Key points Textual warning statements were better remembered when paired with more emotionally salient smoking-related images. The effect of the image component on the memory of the textual component did not decline after 4 weeks of repetitive daily exposure. High emotional salience of the image component of graphic warning labels is essential for their overall effectiveness. References 1 Hammond D. Health warning messages on tobacco products: a review. Tob Control  2011; 20: 327– 37. Google Scholar CrossRef Search ADS PubMed  2 FDA. Required warnings for cigarette packages and advertisements. Final rule. Fed Regist  2011; 76: 36628– 777. 3 R.J. Reynolds Tobacco Company v. Food and Drug Administration. F3d 1205 US Court of Appeals for the DC Circuit ( 2012). 4 Noar SM, Hall MG, Francis DB, et al.   Pictorial cigarette pack warnings: a meta-analysis of experimental studies. Tob Control  2016; 25: 341– 54. Google Scholar CrossRef Search ADS PubMed  5 Wang AL, Lowen SB, Romer D, et al.   Emotional reaction facilitates the brain and behavioural impact of graphic cigarette warning labels in smokers. Tob Control  2015; 24: 225– 32. Google Scholar CrossRef Search ADS PubMed  6 Hamann SB, Ely TD, Grafton ST, Kilts CD. Amygdala activity related to enhanced memory for pleasant and aversive stimuli. Nat Neurosci  1999; 2: 289– 93. Google Scholar CrossRef Search ADS PubMed  7 Anderson L, Shimamura AP. Influences of emotion on context memory while viewing film clips. Am J Psychol  2005; 118: 323– 37. Google Scholar PubMed  8 Langleben DD, Loughead JR, Ruparel KR, et al.   Reduced prefrontal and temporal processing and recall of high “sensation value” ads. Neuroimage  2009; 46: 219– 25. Google Scholar CrossRef Search ADS PubMed  9 Houts PS, Witmer JT, Egeth HE, et al.   Using pictographs to enhance recall of spoken medical instructions II. Patient Educ Couns  2001; 43: 231– 42. Google Scholar CrossRef Search ADS PubMed  10 Shi Z, Wang AL, Emery LF, et al.   The importance of relevant emotional arousal in the efficacy of pictorial health warnings for cigarettes. Nicotine Tob Res  2017; 19: 750– 5. Google Scholar PubMed  11 Klein EG, Quisenberry AJ, Shoben AB, et al.   Health warning labels for smokeless tobacco: the impact of graphic images on attention, recall, and craving. Nicotine Tob Res  2017; 19: 1172. Google Scholar CrossRef Search ADS PubMed  12 Orentlicher D. The FDA's graphic tobacco warnings and the first amendment. N Engl J Med  2013; 369: 204– 6. Google Scholar CrossRef Search ADS PubMed  13 Nonnemaker J, Farrelly M, Kamyab K, Busey A, Mann N. Experimental Study of Graphic Cigarette Warning Labels. Research Triangle Park, NC 27709: RTI International, 2010: 1– 3. 14 Dolcos F, LaBar KS, Cabeza R. Interaction between the amygdala and the medial temporal lobe memory system predicts better memory for emotional events. Neuron  2004; 42: 855– 63. Google Scholar CrossRef Search ADS PubMed  15 Blake TM, Varnhagen CK, Parent MB. Emotionally arousing pictures increase blood glucose levels and enhance recall. Neurobiol Learn Mem  2001; 75: 262– 73. Google Scholar CrossRef Search ADS PubMed  16 Kensinger EA, Corkin S. Memory enhancement for emotional words: are emotional words more vividly remembered than neutral words? Mem Cognit  2003; 31: 1169– 80. Google Scholar CrossRef Search ADS PubMed  17 Cahill L, McGaugh JL. A novel demonstration of enhanced memory associated with emotional arousal. Conscious Cogn  1995; 4: 410– 21. Google Scholar CrossRef Search ADS PubMed  18 Tambini A, Rimmele U, Phelps EA, Davachi L. Emotional brain states carry over and enhance future memory formation. Nat Neurosci  2017; 20: 271– 8. Google Scholar CrossRef Search ADS PubMed  19 Bisby JA, Horner AJ, Horlyck LD, Burgess N. Opposing effects of negative emotion on amygdalar and hippocampal memory for items and associations. Soc Cogn Affect Neurosci  2016; 11: 981– 90. Google Scholar CrossRef Search ADS PubMed  20 Riegel M, Wierzba M, Grabowska A, et al.   Effect of emotion on memory for words and their context. J Comp Neurol  2016; 524: 1636– 45. Google Scholar CrossRef Search ADS PubMed  21 Doerksen S, Shimamura AP. Source memory enhancement for emotional words. Emotion  2001; 1: 5– 11. Google Scholar CrossRef Search ADS PubMed  22 Sachs O, Weis S, Zellagui N, et al.   How different types of conceptual relations modulate brain activation during semantic priming. J Cogn Neurosci  2011; 23: 1263– 73. Google Scholar CrossRef Search ADS PubMed  23 Brown MG, Gold RS. Cognitive responses to positively and negatively framed health messages: a thought-listing study. Psychol Health Med  2014; 19: 724– 9. Google Scholar CrossRef Search ADS PubMed  24 Glock S, Unz D, Kovacs C. Beyond fear appeals: contradicting positive smoking outcome expectancies to influence smokers' implicit attitudes, perception, and behavior. Addictive behaviors  2012; 37: 548– 51. Google Scholar CrossRef Search ADS PubMed  25 LaBar KS, Cabeza R. Cognitive neuroscience of emotional memory. Nature reviews Neuroscience  2006; 7: 54– 64. Google Scholar CrossRef Search ADS PubMed  26 Adolphs R, Cahill L, Schul R, Babinsky R. Impaired declarative memory for emotional material following bilateral amygdala damage in humans. Learn Mem  1997; 4: 291– 300. Google Scholar CrossRef Search ADS PubMed  27 Vardavas CI, Connolly G, Karamanolis K, Kafatos A. Adolescents perceived effectiveness of the proposed European graphic tobacco warning labels. Eur J Public Health  2009; 19: 212– 7. Google Scholar CrossRef Search ADS PubMed  28 Bansal-Travers M, Hammond D, Smith P, Cummings KM. The impact of cigarette pack design, descriptors, and warning labels on risk perception in the U.S. Am J Prev Med  2011; 40: 674– 82. Google Scholar CrossRef Search ADS PubMed  29 Garde A. Freedom of commercial expression and public health protection in Europe. Cambridge Yearb Eur Legal Stud  2010; 12: 225– 56. Google Scholar CrossRef Search ADS   30 R.J. Reynolds Tobacco Co. v. United States FDA. 845 Fd Supp 2d 266 U.S. Dist. LEXIS 26257 (D.D.C., 2012). 31 Brewer NT, Hall MG, Noar SM, et al.   Effect of pictorial cigarette pack warnings on changes in smoking behavior: a randomized clinical trial. JAMA Intern Med  2016; 176: 905– 12. Google Scholar CrossRef Search ADS PubMed  32 Strasser AA, Tang KZ, Romer D, et al.   Graphic warning labels in cigarette advertisements: recall and viewing patterns. Am J Prev Med  2012; 43: 41– 7. Google Scholar CrossRef Search ADS PubMed  33 Wang AL, Romer D, Elman I, et al.   Emotional graphic cigarette warning labels reduce the electrophysiological brain response to smoking cues. Addict Biol  2015; 20: 368– 76. Google Scholar CrossRef Search ADS PubMed  34 Wang AL, Ruparel K, Loughead JW, et al.   Content matters: neuroimaging investigation of brain and behavioral impact of televised anti-tobacco public service announcements. J Neurosci  2013; 33: 7420– 7. Google Scholar CrossRef Search ADS PubMed  35 Knapp ML, Stohl C, Reardon K. Memorable messages. J Commun  1981; 31: 27. Google Scholar CrossRef Search ADS   36 Smith SW, Ellis JB. Memorable messages as guides to self-assessment of behavior: an initial investigation. Commun Monogr  2001; 68: 14. 37 Leas EC, Myers MG, Strong DR, et al.   Recall of anti-tobacco advertisements and effects on quitting behavior: results from the California smokers cohort. Am J Public Health  2015; 105: e90– 7. Google Scholar CrossRef Search ADS PubMed  38 Benowitz NL, Hukkanen J, Jacob P3rd. Nicotine chemistry, metabolism, kinetics and biomarkers. Handb Exp Pharmacol  2009; 192: 29– 60. Google Scholar CrossRef Search ADS   39 Schneider S, Gadinger M, Fischer A. Does the effect go up in smoke? A randomized controlled trial of pictorial warnings on cigarette packaging. Patient Educ Couns  2012; 86: 77– 83. Google Scholar CrossRef Search ADS PubMed  40 WTO Regional Office for Europe. Alcohol Labeling: A Discussion Document on Policy Options. 2017 (http://www.euro.who.int/__data/assets/pdf_file/0006/343806/WH07_Alcohol_Labelling_full_v3.pdf?ua=1). © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The European Journal of Public Health Oxford University Press

Emotional salience of the image component facilitates recall of the text of cigarette warning labels

Loading next page...
 
/lp/ou_press/emotional-salience-of-the-image-component-facilitates-recall-of-the-WkDTUVCe2J
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.
ISSN
1101-1262
eISSN
1464-360X
D.O.I.
10.1093/eurpub/cky059
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Graphic warning labels (GWLs) on cigarette packages, that combine textual warnings with emotionally salient images depicting the adverse health consequences of smoking, have been adopted in most European countries. In the US, the courts deemed the evidence justifying the inclusion of emotionally salient images in GWLs insufficient and put the implementation on hold. We conducted a controlled experimental study examining the effect of emotional salience of GWL’s images on the recall of their text component. Methods Seventy-three non-treatment-seeking daily smokers received cigarette packs carrying GWLs for a period of 4 weeks. Participants were randomly assigned to receive packs with GWLs previously rated as eliciting high or low level of emotional reaction (ER). The two conditions differed in respect to images but used the same textual warning statements. Participants’ recognition of GWL images and statements were tested separately at baseline and again after the 4-week repetitive exposure. Results Textual warning statements were recognized more accurately when paired with high ER images than when paired with low ER images, both at baseline and after daily exposure to GWLs over a 4-week period. Conclusion The results suggest that emotional salience of GWLs facilitates cognitive processing of the textual warnings, resulting in better remembering of the information about the health hazards of smoking. Thus, high emotional salience of the pictorial component of GWLs is essential for their overall effectiveness. Introduction Cigarette packages sold in the US carry a brief text-only Surgeon General’s warning messages that occupy less than a quarter of one of packages’ sides. There are four different messages that are rotated quarterly. This design has not changed significantly since it was introduced in the 1965 Federal Cigarette Labeling and Advertising Act. Although the current labels are considered to be ineffective, there is no consensus among stakeholders about the ways to improve them.1 The Family Smoking Prevention and Tobacco Control Act of 2009 (TCA) required graphic warning labels (GWLs) that combined color graphics and textual warning statements to cover 50% of the front and rear panels of the cigarette package. TCA specified the nine textual warnings to be included in the GWLs, leaving the graphic component to the discretion of the Food and Drug Administration (FDA).2 Tobacco companies sued and were granted a permanent injunction from the implementation of the labels by the District of Columbia court. This ruling was affirmed by a US Court of Appeals, which ‘vacated’ the graphic labels requirement alltogether and ‘remanded’ the Final Rule back to the FDA.3 The district court highlighted the lack of justification for the high emotional salience of the images selected by the FDA, while the appellate court was more concerned with the lack of ‘direct evidence showing that enacting the proposed graphic warnings will accomplish the agency’s stated objective of reducing smoking rates’.3 This reversal spurred a regulatory debate on GWLs in the US that continues to this day. The debate revolves around the relative benefits of the graphic and textual components of GWLs to consumers vs. the risk to society from the infringement by such labels on the tobacco companies constitutional right to free speech. While a large body of literature examined the population-level effects of GWLs, less attention has been given to the cognitive mechanisms that link the perceived information with memory and ultimately objective measures of smoking.4 To this end, an experimental demonstration of whether the emotional salience of the graphic component of GWLs aids the retention of the message in the accompanying textual warning is essential to finding an optimal way to improve their effectiveness with minimum infringement on commercial free speech. Cognitive neuroscience studies find that emotional salience of still or moving images increases their memorability.5,6 However, studies disagree about the benefits of images’ emotional salience to the retention of the information they are paired with.7,8 In the context of warning labels, if GWLs are akin to a figure with a caption, the question is whether a more salient image facilitates or detracts from processing the caption. Recognition accuracy is commonly used as an intermediate outcome measure in communication research in general, and to test the efficacy of health messages in particular.9 Although prior studies show that emotion-provoking GWLs are more memorable,5,10,11 it is still unknown whether increased emotional salience of the accompanying graphic component actually improves remembering the textual component. We hypothesized that images that evoke greater emotional reaction (ER) will facilitate remembering the textual component of the GWLs, and that these effects will be reflected in the recognition accuracy of each of the GWLs components (i.e. text statement vs. image). Methods Participants Seventy-three non-treatment seeking smokers (31 females, 31 Caucasian, 28 African American, 8 Asian and 6 mixed race. Five African American and 4 mixed race participants also self-identified as Hispanic) were recruited through advertising between 2014 and 2016. Participants were 30.89 ±10.30 years old, reported smoking 14.18 ±9.23 cigarettes per day, and had Fagerstrom Test for Nicotine Dependence (FTND) scores of 5.55 ±2.89 (mean ±SD). We excluded individuals who smoked less than 5 cigarettes per day in the last 7 days or had urine cotinine levels undetectable by the qualitative enzyme-linked immunosorbent assay test (Reditest Smoke Cassette, Redwood Toxicology Labs, Santa Rosa, CA, USA). Pregnant or breast-feeding women and smokers currently enrolled in a smoking cessation trial were also excluded. All participants gave written informed consent to participate. The study protocol was approved by the University of Pennsylvania Institutional Review Board. Stimuli Nine TCA-mandated warning statements about the harms of cigarette smoking12 served as the text component of the GWLs. The FDA paired each of the nine statements with different smoking-related graphic images, resulting in 36 GWLs. Therefore some GWLs carried the same textual statement but all differed on the graphic image. The FDA commissioned an online survey of 18 000 respondents who evaluated the emotional salience of these 36 composite GWLs using an emotional reaction (ER) scale.13 Based on the ER score of each GWL image, we selected GWLs with the highest ER score for each textual warning statement to form the ‘high ER GWL condition’, and GWLs with the lowest ER scores for each textual warning statement to form the ‘low ER GWL condition’. Thus, high and low ER conditions shared nine identical textual warning statements while differing on the level of emotional salience of the accompanying graphics (figure 1). The ER scores are significantly different between the high and low ER GWL conditions (P < 0.001). Figure 1 View largeDownload slide An illustration of the experimental procedure Figure 1 View largeDownload slide An illustration of the experimental procedure Procedure Study candidates who passed the phone pre-screening were invited to an in-person screening that included informed consent followed psychiatric history and urine drug screens for cotinine and illicit drugs. Once their eligibility was confirmed, participants were randomly assigned to either the high or low ER GWL condition, in which they were exposed to high or low ER GWLs affixed to the packs of their preferred brand of cigarettes(.) These were supplied by the study team, over a 4-week period. The number of cigarette packs per week was determined by self-reported daily cigarette consumption at screening and baseline assessments. Before distributing the packs to participants, research staff removed the package cellophane and applied the self-adhesive label to cover the top 50% of the front and back panels of the packs, in accordance with the proposed FDA guidelines.2 All the participants attended all four sessions, as they picked up their weekly supply of experimental cigarette packs at these sessions. Participants were encouraged to return unfinished packs, although none brought any back. The average number of packs provided per week did not differ between the high and low ER conditions (P = 0.735). Before and after the 4-week exposure, subjects attended an encoding session, in which they viewed either high or low GWLs, mixed with control stimuli. The control stimuli consisted of nine neutral images from the International Affective Picture System (University of Florida, Gainesville, FL, USA) with text statements unrelated to smoking to match the length and formation of GWLs. Thirty minutes after each encoding, a test of memory for graphic images and text warnings [i.e. Recognition Memory Task, (RMT)] was administered. The RMT consisted of 36 unique text or image stimuli: nine previously seen textual warning statements extracted from the GWLs, nine previously unseen made-up statements related to smoking, nine images extracted from high ER GWLs and nine images extracted from low ER GWLs (figure 1). Thus, for the ‘high ER GWL group’, targets were the nine previously seen textual warnings and nine images from the high ER GWLs, and foils were the nine unseen textual warnings and nine images from the low ER GWLs. Conversely, for the ‘low ER GWL group’, targets were the nine seen textual warnings and nine images from the low ER GWLs, and foils were nine unseen textual warnings and nine images from the high ER GWLs. In each screen, subjects were prompted to respond ‘Yes’ or ‘No’ to the question ‘Have you seen this?’. RMT performance was calculated as percentage of correctly recognized GWL textual warning statements or images respectively, resulting in two scores: Accurate-Text and Accurate-Image. Statistical analysis Statistical analysis was performed using the IBM Statistical Package for the Social Sciences (IBM SPSS version 22). Independent t-tests were used to compare the demographic characteristics of high and low ER GWL groups. To compare the memory of GWL text warnings in two groups at two time points, a two-way repeated-measures analysis of variance (ANOVA) was applied on Accurate-Text using session (encoding session 1 vs. encoding session 2) as the within-subject variable and group (high ER GWL vs. low ER GWL) as the between-subject variable. The same test was repeated on Accurate-Image. Furthermore, to compare memory of GWL text and image within each group across time, a two-way repeated-measures ANOVA was applied in either high or low ER GWL group using component (text vs. image) and session (encoding session 1 vs. encoding session 2) as within-subject variables. Results There were no differences in gender, race, age, education, FTND scores and cigarettes per day between the high and low ER groups at baseline (table 1). Table 1 Subjects characteristics in high and low GWL groups (mean ± SD) GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  Table 1 Subjects characteristics in high and low GWL groups (mean ± SD) GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  GWL group  High  Low  P-value  Gender  17 F (45.9%)  14 F (38.9%)  –  Race  14 AA (37.8%) 18 Caucasian (48.6%) 5 Asian (13.5%)  16 AA (44.4%) 17 Caucasian (47.2%) 3 Asian (8.3%)  –  Age (year)  32.86 ± 11.30  28.86 ± 8.87  0.097  Education (year)  13.89 ± 2.62  13.61 ± 2.09  0.615  Fagerstrom  5.35 ± 3.16  5.75 ± 2.60  0.559  ANOVA analysis on the recognition performance of the text warnings revealed a significant effect of group [F(1, 58) = 4.90, P = 0.031] but not session [F(1, 58) = 0.046, P = 0.832] or group × session interaction [F(1, 58) = 0.046, P = 0.832]. That is, the high ER GWL group was better at recognizing text warnings than the low ER GWL group across time (figure 2a). ANOVA analysis of recognition performance of the images revealed a significant effect of group [F(1, 58) = 8.925, P = 0.004], session [F(1, 58) = 13.032, P = 0.001] and group × session interaction [F(1, 58) = 8.925, P = 0.004]. That is, while the high ER GWL group remained better at recognizing images than the low ER GWL group across time, the low ER GWL group improved after the 4-week exposure (figure 2b). In the high ER GWL group ANOVA revealed a significant effect of component [F(1, 29) = 13.486, P = 0.001] but not of session [F(1, 29) = 0.083, P = 0.776] or component × session interaction [F(1, 29) = 0.335, P = 0.567]. That is, subjects were better at recognizing images than text in the high ER GWL group at both timepoints (figure 3a). ANOVA analysis in the low ER GWL group revealed a significant effect of component [F(1, 29) = 18.502, P < 0.001], session [F(1, 29) = 6.032, P = 0.020] and component × session interaction [F(1, 29) = 7.922, P = 0.009]. That is, subjects were better at recognizing text both at baseline and after a 4-week exposure, while their ability to recognize images improved over time (figure 3b). Figure 2 View largeDownload slide (a) Recognition accuracy of the GWL text was significantly higher in the high ER group than in the low ER group across time. (b) Recognition accuracy of the GWL image remained high in the high ER GWL group across time, while it improved in the low ER GWL group Figure 2 View largeDownload slide (a) Recognition accuracy of the GWL text was significantly higher in the high ER group than in the low ER group across time. (b) Recognition accuracy of the GWL image remained high in the high ER GWL group across time, while it improved in the low ER GWL group Figure 3 View largeDownload slide (a) In the high ER group, images of GWLs were better recognized than texts of GWLs over time. (b) In contrast, the low ER group images of GWLs were less recognized than images of GWLs in Session 1, and the gap become narrower after repeated exposure Figure 3 View largeDownload slide (a) In the high ER group, images of GWLs were better recognized than texts of GWLs over time. (b) In contrast, the low ER group images of GWLs were less recognized than images of GWLs in Session 1, and the gap become narrower after repeated exposure Discussion We found that textual warning statements were better remembered when they were presented together with images that evoked stronger emotion, i.e. high ER GWLs. This finding was unchanged after 4 weeks of daily exposure. The image component of the high ER labels was better remembered at both time points, though the gap between the memory of high and low ER images narrowed at the 4-weeks. Finally, in the low ER group, the text component was better remembered than the image, with the difference diminishing after 4 weeks of exposure to the low ER GWLs. To the best of our knowledge, this is the first controlled experimental demonstration of the emotional salience of the image component of GWL enhancing the memorability of its textual component. The finding that images of the high ER GWLs were much better recognized than those of the low ER GWLs is consistent with cognitive research finding that higher emotional salience enhances memorability of negatively valenced images.6,14,15 Similar observations have been made with words16 and narratives.17 However, there is no consensus on whether emotion enhances or inhibits memory of concurrent verbal information.18,19 Some theories predict that information co-occurring with emotion will be better remembered, while others propose that the emotion would divert attention and processing resources away from the co-occurring verbal or textual information.8,20 Our results concur with the former theory, as we demonstrate that emotional salience not only serves to enhance the memory of images, but also increase the likelihood that viewers will remember the text warnings associated with these images in the composite GWLs. This finding suggests a facilitation effect between the emotional salience of images and processing of textual warning statements. A number of studies found that the negative emotional content of stimuli increases the likelihood of memorizing a contextual detail of the stimulus.16,21 For example, subjects are more likely to correctly report the color of the font in which a negative word was presented than the color of the font of a neutral word.16 Similarly, emotional congruence between images and text may also have an important role in enhancing the memory of a warning statement. Since eight out of nine TCA-mandated text warnings describe extreme health consequences of smoking such as cancer and death, images that are proportionally negative emotionally, are more congruent with those warning statements than the less negative or positive images would be. Indeed, prior studies show that congruent information items are processed faster and ‘better’, as shown by the shorter reaction time and more accurate task performance.22 However, the likely need for continuous development of new GWLs and the possible differential impact of positive vs. negative valence of text and images on the audience warrant future studies that would manipulate the emotional valence and congruence of the GWL text and images.23,24 The enhancing effect of emotion on memory has been attributed to an increase in the interaction between the amgydala and the hippocampus during the encoding and consolidation stages of processing.14,25 The critical role of the amgydala in information processing has been confirmed by studies of emotional memory in patients with amygdalar lesions, who showed impaired recognition of emotional words, pictures and stories.26 In the context of GWLs, this interpretation is consistent with the previously reported greater amygdala response to high ER vs. low ER GWLs.5 According to the Tobacco Labeling Resource Centre and the International Tobacco Control Policy Evaluation Project (The ITC Project), 105 countries or jurisdictions worldwide, including most European countries, require GWLs on cigarette packages although not all have implemented them. Cross-sectional population studies reported positive effects of GWLs on smoking-related outcomes, such as reduced smoking rates in Canada, increased awareness of smoking harm on children and pregnant women in Singapore, increased intention to quit in Thailand, Brazil and Australia and reduced smoking initiation and progression in adolescents in Greece.27 GWL effectiveness is thought to increase with size.28 Thus, countries require GWLs to cover between 30% and 90% of the pack. Nepal and Vanuatu require the largest GWL in the world at 90% of the package, front and back, followed by India (85%) and Thailand (85%). Although the use of graphic images is also recommended,1,28 the knowledge of which aspects of the image should be more heavily weighted, or how image and textual components interact is still limited. In the present study, we showed that emotionally salient images are associated with greater cognitive processing of textual warning messages, suggesting that including images with greater emotional salience could increase the effectiveness of GWLs. Commercial free speech is given greater protection in the USA than in Europe29 setting a higher bar for the evidence required to justify any potential encroachment on the first (free speech) Amendment to the US Constitution. An additional distinction between the European and US medico-legal environment is the degree of emphasis on individual autonomy in making health-related decisions. Thus, an argument that the purpose of cigarette warning labels is to inform consumers rather than to deter them from smoking outright is more likely to be accepted in the US.30 The available data were sufficient for adoption of GWL legislation in many regulatory environments including the European, South East Asian and Americas regions. However, few of the available international studies used objective outcomes or psychological and neuroscientific experimental methodology, creating a dearth of what would be considered incontrovertible objective evidence by the US legal systems. This higher bar is reflected in the legal opinion currently governing the GWL in the USA, that dismissed the link that FDA made between the attitudes and intentions about quitting reported in international studies and the actual smoking as a ‘mere speculation’.3 Indeed, prospective studies using a randomized controlled trial design are better suited to meeting the higher bar set by the courts. Another factor in addressing the gap in tobacco research pointed out by the courts is using reliable outcome measures. The majority of studies of the effects of GWLs on smoking have relied on self-reported outcomes such as changes in attitudes and intentions, quit attempts(,) or self-reported cigarettes smoked per day, rather than direct and objective outcomes such as continuous monitoring of smoking itself or of the levels of nicotine metabolites in bodily fluids.31 A smaller number of trials, including our present study, relied on objective but indirect experimental measures such as recall, eye tracking and behavioral or brain correlates of cue-reactivity.32–34 Memory has been shown to be a useful proxy of subsequent behavior change, suggesting that better memory of the textual messages could be an indicator of subsequent behavioral changes.35,36 Specifically for tobacco, Leas et al reported that better recall of anti-tobacco video ads predicted smoking cessation at follow-up.37 By experimentally demonstrating that emotionally salient graphic component of GWLs enhance the memorability of warning messages, the present study provides an insight about the features of an effective GWL. However, our findings do not directly link the messages’ memorability and their effectiveness at reducing smoking. Establishing such direct causal relationship described in the opinion of the US Court of Appeals would require a prospective randomized-control trial using laboratory measures of nicotine intake as one of its outcome measures.3,32,38 Furthermore, the adoption of GWLs does not signal the end of the need for ongoing revision and updating of GWLs. While our results show no reduction in recall of images or text after 4 weeks of exposure, other studies indicate that GWLs effects may be subject to desensitization over longer periods of exposure.39 Our study offers an approach for testing the GWL components in a controlled fashion that could be used in the ongoing process of refining existing GWLs or evaluating new ones. Future studies may consider using even more objective variables such as neuroimaging, or a longer exposure period, with the same platform. Lastly, our findings may generalize to other public health topics where product labeling exists or is being considered, such as food and alcohol. For example, alcohol consumption in Europe is the highest in the world and there is a discussion of reducing harmful use of alcohol by introducing health labels on alcohol products.40 Indeed, labels provide a unique opportunity for health authorities to communicate the alcohol-related health risks at points of sale and consumption. Since data on effective alcohol labeling for public health purposes are limited or lacking, data from studies like ours could provide the initial basis for developing and testing alcohol warning labels. Conclusions We demonstrated that emotional salience of the image component of GWLs enhances the processing of its textual component, resulting in better memory of the warning message. This finding suggests that highly emotionally salient pictures increase the GWLs effectiveness. Our study offers an approach to the evaluation of GWL effectiveness and contributes to a better understanding of cognitive processing of health warnings that may be applied to the development of health warning labels in other public health domains. Funding This work was supported by the National Institute on Drug Abuse (R01DA036028) and National Institute of Child Health and Human Development (K99HD084746). Conflicts of interest: None declared. Key points Textual warning statements were better remembered when paired with more emotionally salient smoking-related images. The effect of the image component on the memory of the textual component did not decline after 4 weeks of repetitive daily exposure. High emotional salience of the image component of graphic warning labels is essential for their overall effectiveness. References 1 Hammond D. Health warning messages on tobacco products: a review. Tob Control  2011; 20: 327– 37. Google Scholar CrossRef Search ADS PubMed  2 FDA. Required warnings for cigarette packages and advertisements. Final rule. Fed Regist  2011; 76: 36628– 777. 3 R.J. Reynolds Tobacco Company v. Food and Drug Administration. F3d 1205 US Court of Appeals for the DC Circuit ( 2012). 4 Noar SM, Hall MG, Francis DB, et al.   Pictorial cigarette pack warnings: a meta-analysis of experimental studies. Tob Control  2016; 25: 341– 54. Google Scholar CrossRef Search ADS PubMed  5 Wang AL, Lowen SB, Romer D, et al.   Emotional reaction facilitates the brain and behavioural impact of graphic cigarette warning labels in smokers. Tob Control  2015; 24: 225– 32. Google Scholar CrossRef Search ADS PubMed  6 Hamann SB, Ely TD, Grafton ST, Kilts CD. Amygdala activity related to enhanced memory for pleasant and aversive stimuli. Nat Neurosci  1999; 2: 289– 93. Google Scholar CrossRef Search ADS PubMed  7 Anderson L, Shimamura AP. Influences of emotion on context memory while viewing film clips. Am J Psychol  2005; 118: 323– 37. Google Scholar PubMed  8 Langleben DD, Loughead JR, Ruparel KR, et al.   Reduced prefrontal and temporal processing and recall of high “sensation value” ads. Neuroimage  2009; 46: 219– 25. Google Scholar CrossRef Search ADS PubMed  9 Houts PS, Witmer JT, Egeth HE, et al.   Using pictographs to enhance recall of spoken medical instructions II. Patient Educ Couns  2001; 43: 231– 42. Google Scholar CrossRef Search ADS PubMed  10 Shi Z, Wang AL, Emery LF, et al.   The importance of relevant emotional arousal in the efficacy of pictorial health warnings for cigarettes. Nicotine Tob Res  2017; 19: 750– 5. Google Scholar PubMed  11 Klein EG, Quisenberry AJ, Shoben AB, et al.   Health warning labels for smokeless tobacco: the impact of graphic images on attention, recall, and craving. Nicotine Tob Res  2017; 19: 1172. Google Scholar CrossRef Search ADS PubMed  12 Orentlicher D. The FDA's graphic tobacco warnings and the first amendment. N Engl J Med  2013; 369: 204– 6. Google Scholar CrossRef Search ADS PubMed  13 Nonnemaker J, Farrelly M, Kamyab K, Busey A, Mann N. Experimental Study of Graphic Cigarette Warning Labels. Research Triangle Park, NC 27709: RTI International, 2010: 1– 3. 14 Dolcos F, LaBar KS, Cabeza R. Interaction between the amygdala and the medial temporal lobe memory system predicts better memory for emotional events. Neuron  2004; 42: 855– 63. Google Scholar CrossRef Search ADS PubMed  15 Blake TM, Varnhagen CK, Parent MB. Emotionally arousing pictures increase blood glucose levels and enhance recall. Neurobiol Learn Mem  2001; 75: 262– 73. Google Scholar CrossRef Search ADS PubMed  16 Kensinger EA, Corkin S. Memory enhancement for emotional words: are emotional words more vividly remembered than neutral words? Mem Cognit  2003; 31: 1169– 80. Google Scholar CrossRef Search ADS PubMed  17 Cahill L, McGaugh JL. A novel demonstration of enhanced memory associated with emotional arousal. Conscious Cogn  1995; 4: 410– 21. Google Scholar CrossRef Search ADS PubMed  18 Tambini A, Rimmele U, Phelps EA, Davachi L. Emotional brain states carry over and enhance future memory formation. Nat Neurosci  2017; 20: 271– 8. Google Scholar CrossRef Search ADS PubMed  19 Bisby JA, Horner AJ, Horlyck LD, Burgess N. Opposing effects of negative emotion on amygdalar and hippocampal memory for items and associations. Soc Cogn Affect Neurosci  2016; 11: 981– 90. Google Scholar CrossRef Search ADS PubMed  20 Riegel M, Wierzba M, Grabowska A, et al.   Effect of emotion on memory for words and their context. J Comp Neurol  2016; 524: 1636– 45. Google Scholar CrossRef Search ADS PubMed  21 Doerksen S, Shimamura AP. Source memory enhancement for emotional words. Emotion  2001; 1: 5– 11. Google Scholar CrossRef Search ADS PubMed  22 Sachs O, Weis S, Zellagui N, et al.   How different types of conceptual relations modulate brain activation during semantic priming. J Cogn Neurosci  2011; 23: 1263– 73. Google Scholar CrossRef Search ADS PubMed  23 Brown MG, Gold RS. Cognitive responses to positively and negatively framed health messages: a thought-listing study. Psychol Health Med  2014; 19: 724– 9. Google Scholar CrossRef Search ADS PubMed  24 Glock S, Unz D, Kovacs C. Beyond fear appeals: contradicting positive smoking outcome expectancies to influence smokers' implicit attitudes, perception, and behavior. Addictive behaviors  2012; 37: 548– 51. Google Scholar CrossRef Search ADS PubMed  25 LaBar KS, Cabeza R. Cognitive neuroscience of emotional memory. Nature reviews Neuroscience  2006; 7: 54– 64. Google Scholar CrossRef Search ADS PubMed  26 Adolphs R, Cahill L, Schul R, Babinsky R. Impaired declarative memory for emotional material following bilateral amygdala damage in humans. Learn Mem  1997; 4: 291– 300. Google Scholar CrossRef Search ADS PubMed  27 Vardavas CI, Connolly G, Karamanolis K, Kafatos A. Adolescents perceived effectiveness of the proposed European graphic tobacco warning labels. Eur J Public Health  2009; 19: 212– 7. Google Scholar CrossRef Search ADS PubMed  28 Bansal-Travers M, Hammond D, Smith P, Cummings KM. The impact of cigarette pack design, descriptors, and warning labels on risk perception in the U.S. Am J Prev Med  2011; 40: 674– 82. Google Scholar CrossRef Search ADS PubMed  29 Garde A. Freedom of commercial expression and public health protection in Europe. Cambridge Yearb Eur Legal Stud  2010; 12: 225– 56. Google Scholar CrossRef Search ADS   30 R.J. Reynolds Tobacco Co. v. United States FDA. 845 Fd Supp 2d 266 U.S. Dist. LEXIS 26257 (D.D.C., 2012). 31 Brewer NT, Hall MG, Noar SM, et al.   Effect of pictorial cigarette pack warnings on changes in smoking behavior: a randomized clinical trial. JAMA Intern Med  2016; 176: 905– 12. Google Scholar CrossRef Search ADS PubMed  32 Strasser AA, Tang KZ, Romer D, et al.   Graphic warning labels in cigarette advertisements: recall and viewing patterns. Am J Prev Med  2012; 43: 41– 7. Google Scholar CrossRef Search ADS PubMed  33 Wang AL, Romer D, Elman I, et al.   Emotional graphic cigarette warning labels reduce the electrophysiological brain response to smoking cues. Addict Biol  2015; 20: 368– 76. Google Scholar CrossRef Search ADS PubMed  34 Wang AL, Ruparel K, Loughead JW, et al.   Content matters: neuroimaging investigation of brain and behavioral impact of televised anti-tobacco public service announcements. J Neurosci  2013; 33: 7420– 7. Google Scholar CrossRef Search ADS PubMed  35 Knapp ML, Stohl C, Reardon K. Memorable messages. J Commun  1981; 31: 27. Google Scholar CrossRef Search ADS   36 Smith SW, Ellis JB. Memorable messages as guides to self-assessment of behavior: an initial investigation. Commun Monogr  2001; 68: 14. 37 Leas EC, Myers MG, Strong DR, et al.   Recall of anti-tobacco advertisements and effects on quitting behavior: results from the California smokers cohort. Am J Public Health  2015; 105: e90– 7. Google Scholar CrossRef Search ADS PubMed  38 Benowitz NL, Hukkanen J, Jacob P3rd. Nicotine chemistry, metabolism, kinetics and biomarkers. Handb Exp Pharmacol  2009; 192: 29– 60. Google Scholar CrossRef Search ADS   39 Schneider S, Gadinger M, Fischer A. Does the effect go up in smoke? A randomized controlled trial of pictorial warnings on cigarette packaging. Patient Educ Couns  2012; 86: 77– 83. Google Scholar CrossRef Search ADS PubMed  40 WTO Regional Office for Europe. Alcohol Labeling: A Discussion Document on Policy Options. 2017 (http://www.euro.who.int/__data/assets/pdf_file/0006/343806/WH07_Alcohol_Labelling_full_v3.pdf?ua=1). © The Author(s) 2018. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The European Journal of Public HealthOxford University Press

Published: Apr 26, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off