The contemporary Western diet is ch

The contemporary Western diet is characterised by consuming too much fat and sugar, and not enough fruit and vegetables. These unhealthy eating behaviours increase the risk of chronic health problems such as obesity (NHMRC, 2003). The global prevalence of obesity has doubled over the last few decades, with 35% of adults classified as overweight and 11% as obese (WHO, 2013). This increasing rate of overweight and obesity has been linked to many factors, including attitudes, beliefs, information, habits, cultural bias, as well as the environment (Polivy, Herman, & Coelho, 2008). One aspect of the environment that likely contributes to unhealthy eating is the ready availability of high-caloric foods, in particular the continual exposure to visual food cues through advertising in magazines and on billboards (Havermans, 2013). The present study focused on one potential link between food cue exposure and unhealthy eating, namely biased attentional processing (Polivy et al., 2008). An attentional bias is a ‘form of cognitive bias involving preferential attention to one particular type of information’ (MacLeod & Matthews, 2012, p.191). Unhealthy food cues are more likely to automatically capture attention as they are seen as attractive, rewarding and palatable (Polivy et al., 2008).

Recent studies have demonstrated attentional biases for unhealthy food cues in a range of eating-related populations. These have shown that restrained eaters (individuals who restrict their food intake) (Hollitt, Kemps, Tiggemann, Smeets, & Mills, 2010) and external eaters (individuals who eat in response to external food cues e.g., the sight of food) (Brignell et al., 2009, Hou et al., 2011 and Nijs et al., 2009), as well as overweight and obese individuals (Castellanos et al., 2009 and Nijs et al., 2010), are faster to respond to a range of high-caloric food cues relative to neutral (non-food) cues. A few studies have further examined the relationship between such biased attentional processing of food cues and unhealthy eating. In particular, Nijs et al. (2010) and Werthmann et al. (2011) reported a positive correlation between attentional bias for high-caloric snack foods and the subsequent consumption of these foods. In addition, Calitri, Pothos, Tapper, Brunstrom, and Rogers (2010) found that attentional bias for unhealthy food words predicted an increase in weight (BMI) over a 12 month period.

One theory that explains the relationship between biased attentional processing of food cues and consumption is Berridge's (2009) model of food reward. According to this model, motivational value is attributed to food cues through classical conditioning. Food cues (e.g., the sight or smell of food) in the environment become salient through continual association with a rewarding experience (i.e., eating). As a result, they capture attention, which then drives the consumption of that food. These processes can occur implicitly, without necessary conscious awareness.

Given the potential negative health consequences of unhealthy eating, it is important to modify such behaviour. According to Berridge's (2009) model of food reward, one way to counter unhealthy eating may be to focus on changing the underlying cognitive process, in particular, the attentional bias. Thus, it is proposed that decreases in the attentional bias for unhealthy food will lead to decreases in the consumption of unhealthy food. Supporting evidence comes from alcohol research (e.g., Field et al., 2005 and Field et al., 2007) which has used a visual dot probe task, originally developed to modify attentional biases in the anxiety literature (MacLeod, Rutherford, Campbell, Ebsworthy, & Holker, 2002). For example, Field and Eastwood (2005) trained heavy social drinkers to direct their attention towards, or away from, alcohol-related cues. Specifically, a dot probe consistently replaced alcohol-related pictures (‘attend alcohol’ group) or neutral pictures (‘avoid alcohol’ group). As predicted, participants in the ‘attend alcohol’ group showed an increased attentional bias for alcohol cues, whereas those in the ‘avoid alcohol’ group showed a decreased bias for such cues. Importantly, participants in the ‘avoid alcohol’ group drank less beer in a subsequent taste test than those in the ‘attend alcohol’ group. Therefore, Field and Eastwood were able to show that an experimental manipulation of attentional bias for alcohol reduced beer consumption.

One study has extended this finding into the food domain. Kemps, Tiggemann, Orr, and Grear (in press) showed that attentional bias modification for chocolate cues affected chocolate consumption in young women. Specifically, participants in the ‘attend chocolate’ group showed an increased attentional bias for chocolate cues, whereas those in the ‘avoid chocolate’ group showed a reduced bias for such cues. In addition, participants in the ‘avoid chocolate’ group subsequently ate less of a chocolate product (muffin) than those in the ‘attend chocolate’ group. Therefore, Kemps et al. were able to show that a similar experimental manipulation of attentional bias reduced chocolate consumption.

The present study aimed to determine whether attentional bias modification can discourage the consumption of unhealthy food in general (defined as food containing a high amount of sugar or fat, e.g., cake, chips), beyond the specific food of chocolate. In addition, the study sought to use the established attentional re-training paradigm in a novel way, namely to induce an attentional bias for healthy food cues and thereby, increase the consumption of healthy food (e.g., fruit, vegetables, fish). Thus, in contrast to previous attentional re-training protocols for alcohol (Field & Eastwood, 2005) and chocolate (Kemps et al., in press), people were trained towards a desirable outcome (healthy food), as well as being trained to avoid an undesirable one (unhealthy food).

A number of hypotheses were derived from these aims. Specifically, it was predicted that training would produce changes in attentional bias between groups, such that participants trained to attend to healthy food cues would show an increase in attentional bias to healthy foods (and a reduced attentional bias to unhealthy foods), whereas participants trained to attend to unhealthy food cues would show an increase in attentional bias to unhealthy foods (and a reduced attentional bias to healthy foods). It was also predicted that changes in attentional bias would be related to subsequent consumption, such that participants trained to attend to healthy food cues would eat more healthy snacks (relative to unhealthy snacks) than those trained to attend to unhealthy food cues.