Rehumanizing the homeless: Altered BOLD response following contact with an extreme outgroup

Peter Alexander Kirk, Alexandra O. Cohen, Walter Sinnott‐Armstrong, Lasana T. HarrisbioRxiv 2018 · DOI

TL;DR

This study investigates whether direct contact with homeless individuals can "rehumanize" them in the eyes of participants, as measured by brain activity. The findings suggest that while direct interaction can alter brain responses associated with social cognition, the effects are complex and not always consistent, indicating that contact is a promising but not fully understood intervention.

What Problem Does This Paper Address?

The paper tackles the phenomenon of dehumanization, where individuals fail to attribute mental states (like thoughts and feelings) to certain social groups, particularly those perceived as "extreme outgroups" such as the homeless. This lack of mental state attribution is linked to reduced activity in brain regions associated with social cognition, like the medial prefrontal cortex (mPFC). The core problem is understanding if and how this dehumanizing perception can be reversed, specifically by examining the impact of direct, real-world contact with homeless individuals on brain activity. This is crucial because dehumanization can lead to apathy and support for negative social policies.

Methods

The researchers employed functional magnetic resonance imaging (fMRI) to measure brain activity. They conducted two experiments:

  • Experiment 1 (Mixed Design): 20 participants underwent fMRI scans before and after a direct contact intervention with homeless individuals. During the scans, participants viewed images of various social groups and made emotional judgments. The intervention involved participants interacting with homeless individuals at a soup kitchen.
  • Experiment 2 (Repeated Measures): 11 participants underwent a similar fMRI procedure, also with pre- and post-intervention scans, and a contact intervention. This experiment aimed to replicate and further explore the findings of Experiment 1.

The fMRI data was preprocessed using standard techniques, including motion correction and spatial smoothing. Data analysis involved general linear models (GLMs) to compare brain activity between different conditions (e.g., humanized vs. dehumanized targets, pre- vs. post-intervention). Regions of interest (ROIs) in social cognition networks were specifically analyzed. Behavioral data, including enjoyment ratings and time spent in interaction, were also collected and analyzed.

Datasets

Not discussed in the paper. The study used images of individuals from 8 different social groups as stimuli.

Key Insights and Ideas

  • Dehumanization is linked to reduced mPFC activity: Consistent with prior research, viewing homeless individuals (a "low-low" group in the Stereotype Content Model) elicited less activity in the mPFC compared to more "humanized" groups, suggesting a lack of mental state attribution.
  • Contact can alter brain responses: The intervention of direct contact with homeless individuals showed some evidence of altering brain activity in regions associated with social cognition, such as the mPFC, insula, IPL, and MTG.
  • Mixed evidence for rehumanization: While some brain regions showed changes post-contact, the overall results were mixed, suggesting that contact is not a simple "on/off" switch for dehumanization.
  • Affective responses are complex: Participants consistently rated drug-addicted individuals with disgust, while homeless individuals were rated with pity. This suggests different affective pathways might be involved in dehumanization.
  • Time spent in interaction matters: In Experiment 1, the amount of time participants spent interacting with homeless individuals positively correlated with changes in mPFC activity, highlighting the importance of the quality and duration of contact.

Key Results

  • Experiment 1: Showed a trend towards increased mPFC activity in response to homeless individuals after the intervention, particularly correlating with time spent in interaction. However, this effect was not statistically significant across all participants.
  • Experiment 2: Found significant changes in activity in the left IPL and MTG following contact with homeless individuals, suggesting these regions are involved in rehumanization. The right insula showed decreased activity pre-intervention but no significant change post-intervention.
  • Behavioral Data: Participants spent significantly more time in the direct interaction condition than a video condition and reported higher enjoyment in the interaction.
  • Emotional Judgments: Drug-addicted targets consistently elicited disgust, while homeless targets elicited pity.

Limitations and Future Work

  • Sample Size: Both experiments had relatively small sample sizes (20 and 11 participants), which may have limited the statistical power to detect significant effects.
  • Mixed Findings: The inconsistent results across experiments and regions highlight the complexity of dehumanization and rehumanization processes.
  • Mechanisms of Change: The exact neural mechanisms by which contact influences social cognition remain unclear.
  • Future Research: The authors suggest:
  • Re-evaluating the Stereotype Content Model's categories and emotional responses.
  • Using richer measures of affect beyond forced-choice paradigms.
  • Employing advanced fMRI statistics (e.g., effective connectivity analysis) to understand schema-triggered affect.
  • Investigating the role of motivation and anticipated exhaustion as factors influencing contact's efficacy.
  • Exploring cross-group effects and the dynamic relationship between attitudes towards different social groups.

Summary updated

Abstract

Abstract Dehumanized perception strips people of their humanity by ignoring their mental states. Evidence has accumulated to suggest many individuals do not spontaneously attribute mental states towards certain social groups, such as the homeless, and drug-addicted (Fiske et al., 2002, Harris & Fiske, 2006; 2007; 2011). These groups tend to elicit differential BOLD signal within areas associated with social cognition. To investigate the versatility of this response, two experiments were conducted: a mixed design study (20 participants); and a repeated-measures design (11 participants). These investigated the malleability of social cognition following a contact intervention with the homeless. Both experiments had participants make emotional judgements toward eight different social groups whilst under fMRI. The two studies found mixed evidence, demonstrating altered response to homeless people in regions such as the mPFC, Insula, IPL, and MTG following social contact. This lends some support to the use of contact as an effective intervention.
BibTeX
@misc{kirk2018rehumanizing,
  author = {Peter Alexander Kirk and Alexandra O. Cohen and Walter Sinnott‐Armstrong and Lasana T. Harris},
  title = {{Rehumanizing the homeless: Altered BOLD response following contact with an extreme outgroup}},
  year = {2018},
  doi = {10.1101/462671},
  howpublished = {bioRxiv},
}
Figure