How Interoceptive Awareness Promotes Decision-making: An fMRI study on Neuroforecasting Mobile Games Engagement

Description

Neuroscientists in decision science have advanced an affect-integration-motivation (AIM) framework demonstrating that the neural activity associated with anticipatory affect is predictive of individual behavior and aggregate choice. While traditional psychological theories view aggregate choice as a summation of individual choices, it remains unclear why sampled neural responses can be used to forecast aggregate choices above and beyond sampled individual choices and self-reported affective ratings. We hypothesized that interoceptive awareness may bridge the gap between predicting individual choice and aggregate choice, such that individuals who are better at accessing their bodily sensations are more likely to integrate their physiological responses into decision making, as manifested in consistent self-reporting, behavior, and neural measures jointly forecasting aggregate choice. Therefore, we examined whether participants’ neural responses, choices, and affective ratings forecast download times and revenue from mobile games in app markets, and investigated the role of interoceptive awareness by comparing prediction models from different subgroups of participants. Generally, positive arousal and neural activity in posterior cingulate cortex positively predict individuals’ decisions to download mobile games, whereas individual download choices and sampled affective neural responses (nucleus accumbens, NAcc) forecast aggregate download times. For the high interoceptive sensibility (IS) group, their sampled positive arousal and brain activity associated with positive affect (NAcc) and value integration (medial prefrontal cortex, mPFC) collectively forecasted aggregate download times while, for the low IS group, it was their individual choice that forecasted aggregate download times. Finally, positive arousal robustly predicted aggregate revenue in all groups except for individuals with low interoceptive accuracy. These findings not only support, but also extend the AIM framework by shedding light on the influence of interoceptive awareness on the neurobehavioral mechanisms underlying human decision-making. Keywords: forecasting, interoceptive awareness, fMRI, accumbens, mPFC

Steps to reproduce

fMRI Data analysis Individual Choice Model In this general linear model, three event types (download, non-download, and baseline) were used to construct regressors in which event onsets were convolved with the canonical hemodynamic response function of the SPM. These events were modeled with a 4-s duration for the trials in which participants responded. The onsets of the questions, responses, and inter-trial intervals (ITI) were also modeled in three regressors across all event types, with durations of reaction time (RT), 0-s, and jittered time, respectively. An additional seventh regressor may be added to model missing trials in which participants did not respond. Six head-motion parameters modeled the residual effects of head motion as covariates of no interest. The contrast images included downloaded vs. not downloaded images, and vice versa. These contrast images were input into a second-level one-sample t-test to examine the regions that were significantly more activated for each contrast. The threshold of the statistical maps was a whole-brain voxel-wise intensity of p <.001, with false discovery rate (FDR) correction. The resulting regions of activation were reported in terms of the peak voxels in the MNI coordinate space. Game Model In this general linear model, 15 event types (14 games and one baseline) in each block were used to construct regressors in which event onsets were convolved with the SPM's canonical hemodynamic response function. These events were modeled with a 4-s duration for participant response. The onset of the questions, responses, and ITI were also modeled in three regressors across all event types, with reaction time duration (RT), 0-s, and jitter time, respectively. Six head-motion parameters were used to model the residual effects of head motion as covariates of no interest. Region of Interest Analysis We selected the bilateral NAcc, bilateral mPFC, bilateral AIns, and PCC as a priori regions of interest (ROIs), based on previous neuroimaging studies that reported these regions as individual/aggregate choice predictors (Genevsky and Knutson 2015; Tong, Acikalin et al. 2020,). Behavioral Analysis The positive arousal is calculated as (arousal/√2)+(valence/√2); and negative arousal as (arousal/√2)-(valence/√2).. The aggregate mobile game download times equals to log⁡((download times (iso & android))/(age of game (months) )), and the aggregate revenue is log⁡((revenue (iso & android))/(age of game (months) )). Linear regression analyses using a stepwise method were performed to examine the variables that could predict individual/aggregate indices. To predict individual download choices, the independent variables were positive arousal, negative arousal, and parameter estimates of the NAcc, AIns, mPFC, and PCC. To predict the aggregate download times and aggregate revenue, in addition to the above mentioned variables, the download rate was included as an independent variable.

Institutions

Categories

Funding

Licence