Food Choices after Cognitive Load: An Affective Computing Approach

Abstract

Psychology and nutritional science research has highlighted the impact of negative emotions and cognitive load on calorie consumption behaviour using subjective questionnaires. Isolated studies in other domains objectively assess cognitive load without considering its effects on eating behaviour. This study aims to explore the potential for developing an integrated eating behaviour assistant system that incorporates cognitive load factors. Two experimental sessions were conducted using custom-developed experimentation software to induce different stimuli. During these sessions, we collected 30 h of physiological, food consumption, and affective states questionnaires data to automatically detect cognitive load and analyse its effect on food choice. Utilising grid search optimisation and leave-one-subject-out cross-validation, a support vector machine model achieved a mean classification accuracy of 85.12% for the two cognitive load tasks using eight relevant features. Statistical analysis was performed on calorie consumption and questionnaire data. Furthermore, 75% of the subjects with higher negative affect significantly increased consumption of specific foods after high-cognitive-load tasks. These findings offer insights into the intricate relationship between cognitive load, affective states, and food choice, paving the way for an eating behaviour assistant system to manage food choices during cognitive load. Future research should enhance system capabilities and explore real-world applications.

Keywords: cognitive load; eating behaviour; electrodermal activity; machine learning; photoplethysmography; physiological signals; sensors.

Figure 1

Overview of the experimental framework, with the three types of data collected (physiological signals, food calories, and questionnaires) and the data processing steps to find and analyse trends in eating behaviour after cognitive loads. Abbreviations: GSR—Galvanic Skin Response, PPG—Photoplethysmography.

Figure 2

Demonstration of the Stroop task: In the first block, the font colours are identical to the colour name; in the second and third blocks, the Stroop task is demonstrated with each colour written in a different font colour.

Figure 3

High and low cognitive load recording session sequence (left to right).

Figure 4

High-cognitive-load 2-back task demonstration. The blue square which appeared in the first block repeats its position in the third block.

Figure 5

Arrangement of snacks: (left to right) 3 chocolate bars, potato chips, salted pretzels, nuts, black grapes, carrot slices, cucumber slices, orange juice, and water.

Figure 6

Recursive feature elimination with 12-fold cross validation to select features.

Figure 7

Mean leave-one-subject-out performance of machine learning algorithms for classification of low and high cognitive load. Abbreviation: RF—Random Forest, GNB—Gaussian Naive Bayes, SVM—Support Vector Machine.

Figure 8

Clustering of aggregate data based on food choices. The aggregated data on the ‘x-axis’ contains food choices made during high- and low-cognitive-load sessions, cognitive load classification accuracy, and the subject’s negative affect during the high-cognitive-load session. The ‘y-axis’ contains the subject number. Abbreviation: HL—High cognitive Load, LL—Low cognitive Load.