Brain responses to vehicles predict individual interest and forecast changes in aggregate demand
Poster No:
1341
Submission Type:
Abstract Submission
Authors:
Dylan Christiano1,2, Cynthia Wu1, Tara Srirangarajan1, Shabnam Hakimi3, Matthew Klenk3, Charlene Wu3, Brian Knutson1
Institutions:
1Stanford University, Stanford, CA, 2University of Michigan, Ann Arbor, MI, 3Toyota Research Institute, Palo Alto, CA
First Author:
Co-Author(s):
Introduction:
Demand for vehicles using alternative fuel sources is increasing, leading to the introduction of new vehicle models and types. Little is known, however, about market demand for these new vehicles. We sought to examine whether brain activity could predict individuals' desire to purchase and learn more about vehicles (Erk et al., 2002), as well as forecast changes in demand for vehicles out-of-sample in the US market (Knutson & Genevsky, 2018). Drawing from the AIM (Affect, Integration, Motivation) framework, we hypothesized that activity in the Nucleus Accumbens (NAcc), associated with dopaminergic signaling and positive arousal to stimuli, should predict choice (Samanez-Larkin & Knutson, 2015) and might forecast changes in demand for new vehicles.
Methods:
13 subjects who reported being interested in purchasing a vehicle in the next two years participated in a vehicle rating task as their brain activity was monitored using Functional Magnetic Resonance Imaging (FMRI). Participants viewed 48 trials per task, each lasting an average of 16 seconds. A mixed set of 12 Electric Vehicles (EV) and 12 Internal Combustion Engine (ICE) vehicles from various brands. Each vehicle appeared twice, in black or white colors, at a ¾ angle, and in a pseudorandom order. During each vehicle rating task trial, subjects initially saw a centrally presented image of the vehicle and its' model name (2 sec), followed by its fuel source (2 sec), followed by rating prompts querying interest in learning more and desire to purchase the vehicle (4 sec each). To assess changes in aggregate United States market demand, we collected publicly available data indicating units of each model sold for each quarter of 2022 (from goodcarbadcar.net). We then calculated the average and slope of units sold over quarters of 2022 to estimate average demand and change in demand.
For individual prediction analyses, activity was averaged prior to choice and extracted from predicted Volumes Of Interest (VOIs) in the NAcc, Medial PreFrontal Cortex (MPFC), and Anterior Insula (AIns) and regressed against rated desire to know more and to purchse on each trial (Samanez-Larkin and Knutson, 2015). For aggregate market forecasts, activity in these VOIs was averaged by model and regressed against the slope of sales for 2022.
For individual prediction analyses, activity was averaged prior to choice and extracted from predicted Volumes Of Interest (VOIs) in the NAcc, Medial PreFrontal Cortex (MPFC), and Anterior Insula (AIns) and regressed against rated desire to know more and to purchse on each trial (Samanez-Larkin and Knutson, 2015). For aggregate market forecasts, activity in these VOIs was averaged by model and regressed against the slope of sales for 2022.
Results:
Within individuals, both VOI and whole brain analyses revealed that subjects' initial NAcc response to vehicles predicted rated desire to know more about and purchase those vehicles. A main effect of fuel type indicated that subjects preferred electric to gas vehicles (beta=15.21±0.07, p < 0.001) and NAcc activity also predicted vehicle preference (beta=2.63±0.11, p < 0.01). At the aggregate level, average NAcc response to vehicles also forecast the slope in market demand for units sold (p<.006), but not the average of units sold during 2022. In combination with NAcc activity, averaged behavior (i.e., desire to know and purchase ratings), however, did not forecast the slope or average units sold.
·Whole brain maps confirm that NAcc activity predicts individual desire to know more and to purchase different vehicle models.
·Pairwise plot indicating that NAcc response to vehicles forecasts aggregate change in U.S. demand for different models over 2022.
Conclusions:
Brain responses to vehicles, specifically early responses in the NAcc, predict individuals' desire learn more about and to purchase them. Further, average group NAcc activity forecast changes in demand for cars, beyond behavioral forecasts. These findings suggest that brain activity might add value to conventional measures for forecasting market demand for new vehicles.
Emotion, Motivation and Social Neuroscience:
Emotion and Motivation Other 2
Higher Cognitive Functions:
Decision Making
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Subcortical Structures
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
Data analysis
Emotions
MRI
Sub-Cortical
Univariate
1|2Indicates the priority used for review
Provide references using author date format
Knutson, B., & Genevsky, A. (2018). Neuroforecasting aggregate choice. Current Directions in Psychological Science, 27(2), 110-115.
Samanez-Larkin, G. R., & Knutson, B. (2015). Decision making in the ageing brain: changes in affective and motivational circuits. Nature Reviews Neuroscience, 16(5), 278-289.