Implicit evaluation of the emotional response to premium organic herbal infusions through a temporal dominance approach: Development of the temporal dominance of facial emotions (TDFE)

• Published in: Food quality and preference 2019-09, Vol.76, p.71-80
• Main Authors: Rocha, C., Lima, R.C., Moura, A.P., Costa, T., Cunha, L.M.
• Format: Article
• Language: English
• Subjects: EsSense Profile; FaceReader; Implicit method, lemon verbena; Overall liking; Temporal dominance of emotions
• ISSN: 0950-3293; 1873-6343
• DOI: 10.1016/j.foodqual.2019.04.001

•A new dynamic approach on facial behavioural evaluation.•Food-evoked Facial Emotions evaluated following a temporal dominance approach.•The TDFE as a complementary approach to the EsSense Profile®.•An Excel-based program is accessible to convert FaceReader™ data into binary data. One of the greatest challenges when designing research on emotions is the conception of an efficient tool for their measurement. The FaceReader™ emerged as an implicit tool capable of analysing facial expression patterns whether in recorded or real-time videos. This software produces large datasets on temporal variation of the intensity of basic emotions. The main goals of this research were: (i) to develop a new data analysis procedure, based on temporal dominance approaches to evaluate temporal data from Face Reader, called the Temporal Dominance of Facial Emotions (TDFE), (ii) to relate the results from TDFE with the EsSense Profile® evaluation, (iii) to evaluate different market samples and to establish its emotional profile. An untrained panel of 50 naïve panellists tasted five samples. For each infusion, the overall liking was evaluated using a 9-point scale (1-dislike extremely to 9-like extremely) followed by the evaluation of the emotional profile, using an adapted version of the EsSense Profile®. The entire test was video- recorded using web cameras placed in front of the panellists and adjusted, to obtain a correct measurement of their facial expressions. Significant differences (p ≤ 0.05) between samples were found for overall liking. The correlation between TDFE and EsSense Profile® was 0.704, above the reference value of 0.680. However, from both methods one can identify a clear emotional profile of the samples under evaluation. Both methodologies appear to be promising regarding the emotional profile characterization. However, the lower direct correlation can be explained by the difference in positive and negative emotions. FaceReader™ identifies more negative than positive emotions, while the EsSense Profile® identified the opposite. Therefore, both approaches should be regarded as complementary.