Neural reference groups: a synchrony-based classification approach for predicting attitudes using fNIRS

Abstract

Social neuroscience research has demonstrated that those who are like-minded are also 'like-brained.' Studies have shown that people who share similar viewpoints have greater neural synchrony with one another, and less synchrony with people who 'see things differently.' Although these effects have been demonstrated at the 'group level,' little work has been done to predict the viewpoints of specific 'individuals' using neural synchrony measures. Furthermore, the studies that have made predictions using synchrony-based classification at the individual level used expensive and immobile neuroimaging equipment (e.g. functional magnetic resonance imaging) in highly controlled laboratory settings, which may not generalize to real-world contexts. Thus, this study uses a simple synchrony-based classification method, which we refer to as the 'neural reference groups' approach, to predict individuals' dispositional attitudes from data collected in a mobile 'pop-up neuroscience' lab. Using functional near-infrared spectroscopy data, we predicted individuals' partisan stances on a sociopolitical issue by comparing their neural timecourses to data from two partisan neural reference groups. We found that partisan stance could be identified at above-chance levels using data from dorsomedial prefrontal cortex. These results indicate that the neural reference groups approach can be used to investigate naturally occurring, dispositional differences anywhere in the world.

Keywords: dmPFC; fNIRS; intersubject correlation; neural reference groups; neural synchrony.

Fig. 1.

(a) Locations of 20 NIRS channels, which are formed between adjacent sources and detectors. (b) Experimental setup, showing participant fitted with fNIRS cap in the mobile laboratory, which was established in a market research company’s office space.

Fig. 2.

Depiction of the neural reference group classification approach. (a) Neural timecourses from channel 9 for participants holding a pro-life stance are averaged together to form a pro-life neural reference group timecourse. (b) Timecourses for participants holding a pro-choice stance are averaged together to form a pro-choice neural reference group timecourse. (c) A participant’s timecourse, whose data were not included in the reference group timecourses, is compared to the timecourses of the two neural reference groups. The participant is then categorized as belonging to one group or the other by demonstrating greater similarity with one group over the other, as measured by a distance metric (Euclidean distance in this case, though Pearson correlation might also be used). Areas that are shaded in purple demonstrate overlap where the participant‘s timecourse differed from both reference groups. Areas shaded blue or red correspond to where the participant’s timecourse diverged more from one of the reference groups (blue=diverging further from pro-choice, red=diverging further from pro-life). These red and blue areas are key to determining which reference group the participant differs from most in order to match the participant as being likely to belong to one group or the other. Blue and red bars shown above the graph indicate sections of the timecourse where the participant differed more than (i.e. had a greater Euclidean distance from) one group or the other. For the participant shown here, a larger blue area than red area across all timepoints indicates that the participant differed more from the pro-choice group, and thus this participant was classified as being pro-life. In future studies, it may be valuable to examine regions of the timecourse when most participants tend to show similarity to one group over the other and identify moments in the video to which those timepoints correspond.

Fig. 3.

For each video within each channel, group-level differences between pro-life and pro-choice participants were computed as the Euclidean distance between the mean timecourse for each group. These Euclidean distance values are shown projected onto a 3D cortical surface for each video: pro-life (left) and pro-choice (right). These maps were used to identify ROIs for conducting the classification-based analyses.

Fig. 4.

Classification accuracy for channels (in dmPFC) that could distinguish between partisan groups at above-chance levels for the pro-life (left) and pro-choice (right) videos. For each video, the observed classification accuracy is shown relative to a null distribution of accuracy scores generated for shuffled data.