Are we Lovin' It? A single blind randomised controlled trial to investigate the influence of high in fats, salt and sugar food brand advertising on cognitive, affective and behavioural intention outco

Human faces carry crucial social signals that play a central role in interpersonal communication and observers exhibit remarkable face processing proficiency in rapidly extracting a wide range of facial features to support various cognitive and social functions (, , , , , , . This expertise in face processing is, however, strongly influenced by an observer’s experience with specific categories of faces. Behaviorally, this is reflected in the differential processing of faces from the observer’s own race (i.e. same-race; SR) compared to those of other-races (OR). Two well-documented phenomena illustrate this effect: the Same-Race Recognition Advantage (SRRA), characterized by higher recognition accuracy for SR faces, and the Other-Race Categorization Advantage (ORCA), where race categorization is faster for OR faces. One of the main explanations for these phenomena is that perceptual expertise tunes the face processing system towards the most frequently encountered and socially relevant category (, ). While this enhanced specialized neural and cognitive processing increases SR face recognition, it has been hypothesized that it also slows down their categorization by race (,, ). In contrast, OR faces, which are processed more coarsely, might allow faster access to racial information, thereby allowing for a faster categorization. Race does not only influence face processing under typical viewing conditions, but it also modulates the occurrence of other well-established face-specific effects. One such phenomenon is the face inversion effect (FIE). The FIE refers to the well-documented decline in performance when faces are presented inverted rather than upright, an effect that is significantly stronger for faces compared to any other non-face object (, ). While the majority of behavioral studies using face categorization by race tasks reported no race-related differences in the FIE (, , , ), those using face recognition or discrimination paradigms reported larger FIE for SR than OR faces, with the decline in face recognition accuracy being more pronounced for SR faces (; , , , ; but see ). A prevailing explanation for this difference is rooted in how visual expertise shapes the face processing mechanisms. Holistic processing, which refers to the perception of a stimulus as a unified whole by relying on the configural integration of its features (i.e. the integration of spatial relationships between facial features), is more prominent for faces with which we possess extensive experience, such as SR faces (, , , ). In contrast, the limited familiarity with OR faces leads to a greater reliance on piecemeal processing, wherein individual facial features are analyzed independently. As inversion disrupts the holistic perception of a stimulus (, ), or more broadly impairs perceptual mechanisms optimized for frequently encountered stimuli (, ), it is more detrimental for SR faces (e.g. ). Although behavioral differences in the recognition and categorization by race of SR and OR faces are well established, there is limited consensus in the literature on whether, and how, these effects occur at the neural level. In the electrophysiological domain, several studies using event-related potentials (ERPs) have investigated the impact of face race on the N170 component–a face-sensitive negative deflection, peaking around 170 ms after stimulus onset and typically observed over posterior temporal sites (). Despite extensive investigation, findings remain inconclusive: while some studies have reported larger N170 amplitudes for upright SR than OR faces (e.g. , , ), others have found no race-related modulation of the N170 (e.g. , , , , , , ), or even a reversed pattern, with larger amplitudes for OR faces (e.g. , , , , ). Similar inconsistencies have been observed in how the neural FIE manifests across face races. For same-race faces, the neural FIE is typically characterized by increased amplitude and delayed latency of the N170 component for inverted faces (e.g. , , , , , , , , ). As for SR vs. OR FIE differences, some studies revealed greater FIE for SR than OR faces (, ), whereas others reported similar FIEs on the N170 component for both races (, , , , ). Finally, one study reported the reverse pattern, with a larger FIE for OR than SR faces (). Such inconsistency in the literature may be attributed to a lack of precision, sensitivity or objectivity in the ERP measure. Specifically, different levels of attention or memory load may modulate the N170 (, ). This component is also susceptible to noise, including artifacts from muscle activity (, ). In addition, variability in latency across trials can reduce the precision of ERP measures (, ). When responses are averaged, such variability may blur the neural signal, leading to less well-defined peaks. The low-level properties of the stimuli when not controlled could also modulate the ERPs responses. Furthermore, the objectivity of ERP analysis is often compromised by the subjective