Behavioral traits that define social dominance are the same that reduce social influence in a consensus task

Mariana Rodriguez-Santiago, Paul Nührenberg, James Derry, Oliver Deußen, Fritz A. Francisco, Linda Karen Garrison, Sylvia Fernanda Garza, Hans A. Hofmann, Alex Jordan

TL;DR

In the animal kingdom, "alpha" status doesn't always translate to effective leadership. This study shows that while dominant cichlid fish lead routine group movements, their aggressive and aversive behaviors actually prevent them from influencing groups during complex consensus tasks—allowing "passive" subordinates to emerge as the true agents of social change.

What Problem Does This Paper Address?

We often conflate social dominance with social influence. In many vertebrate societies, individuals reach the top of the hierarchy through aggression, intimidation, and coercion. While these traits are effective for securing resources or mating opportunities, they are also socially aversive.

The researchers sought to answer a fundamental question: Do the same traits that make an individual dominant also make them a better leader? Specifically, they looked at the "signal-to-noise" problem of social information. If a leader is constantly chasing or attacking others, can the group distinguish between "I am attacking you" and "I am moving toward a food source"? By dissociating dominance from influence, the authors challenge the assumption that the most powerful individuals are the most effective at guiding a group toward a consensus.

Methods

The researchers used the cichlid fish Astatotilapia burtoni, a species where males exist in two distinct, reversible phenotypes: highly aggressive, brightly colored "dominants" and passive, cryptically colored "subordinates."

  1. The Association Task: Fish were trained to associate a specific colored light (orange vs. cyan) with a food reward.
  2. The Consensus Test: Once an "informant" (either a dominant or subordinate male) learned the task, they were placed into a new group of eight naive fish. The researchers measured how quickly the naive group reached a consensus—defined as seven or more fish moving toward the correct stimulus within one second.
  3. Deep Learning Tracking: To capture granular behavioral data, the team used Mask R-CNN (a convolutional neural network) to segment and track individual fish. This allowed them to reconstruct 2D trajectories, estimate spine poses, and calculate "visual connectivity" (what each fish could actually see from its perspective).
  4. Social Network Analysis: They built directed interaction networks based on "initiator-responder" events (e.g., one fish starts a rapid movement and another follows) to determine network centrality.

Datasets

The study involved a total of 168 fish. While the paper does not utilize a standard external benchmark dataset, it introduces a robust pipeline for behavioral analysis:

  • Video Data: High-resolution recordings of group interactions in oval PVC tubs.
  • Training Labels: A subset of manually labeled images used to train the Mask R-CNN for automated fish detection and segmentation.
  • Behavioral Metrics: A custom dataset of trajectories, speed events, and social network positions.

Key Insights and Ideas

  • The "Noise" of Dominance: Dominant males are "noisy" informants. Because they frequently chase and attack others, their rapid movements toward a food source are often indistinguishable from an aggressive charge. This low signal-to-noise ratio masks informative social cues.
  • Subordinate Success: Subordinate males, despite being socially peripheral and having little influence over routine movement, were the most effective at driving group consensus. Their lack of aggression makes their directed movements "cleaner" and easier for the group to interpret.
  • Spatial Separation: Dominants are physically isolated. Because they are aversive, other fish maintain a greater distance from them. Subordinates stay closer to the group, providing more opportunities for "physical leadership" and observational learning.
  • Context-Dependency: Influence is not a fixed trait. A dominant male might be highly influential in a "routine" scenario (like a simple flight response) but completely ineffective in a "complex" scenario requiring group coordination and learning.

Key Results

  • Consensus Speed: Groups with subordinate informants reached consensus significantly faster than those with dominant informants ($p = .015$).
  • Dominant Failure: Groups with dominant informants were no faster at reaching consensus than groups with no informant at all (completely naive groups).
  • Network Centrality: Dominant males had significantly higher betweenness centrality in routine social networks ($p < .01$), yet this "centrality" was driven by aggression rather than helpful guidance.
  • Signal-to-Noise Ratio: Subordinate males had a significantly higher ratio of informative "signal" (moving toward the goal) versus uninformative "noise" (aggressive chasing) compared to dominants ($p < .001$).

Limitations and Future Work

  • Species Specificity: The authors note that in other species, like rhesus macaques, subordinates may "play dumb" or underperform in the presence of dominants. The dynamics of influence may vary significantly across the phylogenetic tree.
  • Sensory Modalities: The study focused heavily on visual cues. However, cichlids also communicate via water-borne metabolites (androgens) and the lateral line system (mechanosensation). Future work could explore how these non-visual signals contribute to a leader's "aversiveness."
  • Gender Bias: Only males were used as informants because they possess clear phenotypic markers of status. The role of female dominance hierarchies in social influence remains an open question.
  • Human Parallels: The authors suggest these findings might mirror "toxic leadership" in human organizations, where aggressive individuals rise to power but ultimately degrade group performance. Testing this hypothesis in human social networks is a logical next step.

Summary updated

Abstract

Dominant individuals are often most influential in their social groups, affecting movement, opinion, and performance across species and contexts. Yet, behavioral traits like aggression, intimidation, and coercion, which are associated with and in many cases define dominance, can be socially aversive. The traits that make dominant individuals influential in one context may therefore reduce their influence in other contexts. Here, we examine this association between dominance and influence using the cichlid fish <i>Astatotilapia burtoni</i>, comparing the influence of dominant and subordinate males during normal social interactions and in a more complex group consensus association task. We find that phenotypically dominant males are aggressive, socially central, and that these males have a strong influence over normal group movement, whereas subordinate males are passive, socially peripheral, and have little influence over normal movement. However, subordinate males have the greatest influence in generating group consensus during the association task. Dominant males are spatially distant and have lower signal-to-noise ratios of informative behavior in the association task, potentially interfering with their ability to generate group consensus. In contrast, subordinate males are physically close to other group members, have a high signal-to-noise ratio of informative behavior, and equivalent visual connectedness to their group as dominant males. The behavioral traits that define effective social influence are thus highly context specific and can be dissociated with social dominance. Thus, processes of hierarchical ascension in which the most aggressive, competitive, or coercive individuals rise to positions of dominance may be counterproductive in contexts where group performance is prioritized.
BibTeX
@inproceedings{rodriguezsantiago2020behavioral,
  author = {Mariana Rodriguez-Santiago and Paul Nührenberg and James Derry and Oliver Deußen and Fritz A. Francisco and Linda Karen Garrison and Sylvia Fernanda Garza and Hans A. Hofmann and Alex Jordan},
  title = {{Behavioral traits that define social dominance are the same that reduce social influence in a consensus task}},
  booktitle = {Proceedings of the National Academy of Sciences},
  year = {2020},
  doi = {10.1073/pnas.2000158117},
}
Figure