How we choose one over another: predicting trial-by-trial preference decision

Abstract

Preference formation is a complex problem as it is subjective, involves emotion, is led by implicit processes, and changes depending on the context even within the same individual. Thus, scientific attempts to predict preference are challenging, yet quite important for basic understanding of human decision making mechanisms, but prediction in a group-average sense has only a limited significance. In this study, we predicted preferential decisions on a trial by trial basis based on brain responses occurring before the individuals made their decisions explicit. Participants made a binary preference decision of approachability based on faces while their electrophysiological responses were recorded. An artificial neural network based pattern-classifier was used with time-frequency resolved patterns of a functional connectivity measure as features for the classifier. We were able to predict preference decisions with a mean accuracy of 74.3 ± 2.79% at participant-independent level and of 91.4 ± 3.8% at participant-dependent level. Further, we revealed a causal role of the first impression on final decision and demonstrated the temporal trajectory of preference decision formation.

Figure 1. Classification accuracy at individual and combined level.

(a) Prediction performance of the classifier at individual level (PAM; Pi indicates i-th Participant) and at combined level (CGM). Blue and green bars represent classifier accuracy based on brain responses related to the first face (F1X) and second face (F2X), respectively. Prediction accuracy was higher at individual level than at combined level, and also higher for F2X than F1X. Chance level is at 50% (black horizontal line). (b) Classifier performance for PAM and CGM analysis for both F1X and F2X with respect to time. The classifier performance gradually increased till 700 ms especially for CGM.

Figure 2. Time dependency of average F-ratio distribution at individual and combined level.

(a) Temporal profiles of average F-ratio for data available (from the face onset) till that time for CGM for F1X (in red) and for F2X (in black). (b) Temporal profiles of average F-ratio for successive 50 ms time window for CGM for F1X (in red) and for F2X (in black). (c) Temporal profiles of average F-ratio for data available (from the face onset) till that time for PAM for F1X (in green) and for F2X (in blue). (d) Temporal profiles of average F-ratio for successive 50 ms time window for PAM for F1X (in green) and for F2X (in blue).

Figure 3. Scalp maps of average F-ratio distribution for CGM analysis at different time periods.

(a), (b) Analysis for F1X and F2X, respectively. Note that the frontal electrode region has higher F-ratio in case of F1X while right temporal electrode region has higher F-ratio for F2X.