Self-association enhances early attentional selection through automatic prioritization of socially salient signals
Hypotheses The paper tests whether arbitrary self-association acts as an automatic attentional salience signal at early perceptual stages, whether self-prioritization mobilizes absolute TVA processing capacity rather than merely redistributing fixed attentional weights, and whether social and perceptual salience capture attention through largely independent mechanisms that would produce additive effects when both are applied to the same stimulus.
Claims In a temporal-order-judgment task with hierarchical Bayesian TVA modeling across two experiments (N = 69, N = 71), self-associated stimuli are processed about 1.5 Hz faster than other-associated under perceptual decisions but the advantage vanishes or reverses under explicit social decoding, and a condition-specific capacity model decisively outperforms a weights-only model (Delta-loo = 14.2, weight 0.86). For other-associated stimuli the social-by-perceptual interaction is near zero (-0.54 Hz) consistent with additivity, while for self-associated stimuli the perceptual-salience benefit is markedly reduced (2.5 Hz vs 5.2 Hz for other, 6 Hz neutral).
Inferences Self-prioritization operates automatically at early, pre-attentive stages and reduces rather than amplifies under active social decoding, and it does so by recruiting additional processing capacity rather than merely reweighting attention. The sub-additive interaction in the self condition partially refutes the strong form of the independence hypothesis: social and perceptual salience stream independently except when self-relevance obligatorily dominates, forcing a qualified rather than unrestricted independence account.
▸ Summary
▸Hypotheses tested
Self-association acts via absolute processing capacity (TVA C parameter), not merely through redistribution of fixed attentional weights — predicting that condition-specific C should fit better than a single-C model.
A condition-specific TVA capacity model (indiv-C) should outperform a fixed-capacity, weights-only model on LOO cross-validation if the mechanism is capacity change rather than pure weight redistribution.
Tested by
Model comparison via leave-one-out cross-validation shows that a model with condition-specific processing capacity parameters outperforms a single-capacity model (Δloo=14.2, Δse=6.41, weight=0.86 pooled across N=140 participants), indicating that social salience changes absolute processing rates rather than merely redistributing attentional weights.
Arbitrary self-association acts as an automatic attentional salience signal at the perceptual feature level — detectable in TOJ tasks that demand only a perceptual report, and modulated (or reversed) by the social-decision context.
In the perceptual-decision TOJ condition, the self-associated stimulus should show a faster TVA processing rate than the other-associated stimulus — even though social identity is irrelevant to the task.
Tested by
In the perceptual decision dimension (report which shape flickered first), self-associated stimuli were processed 1.5 Hz faster than other-associated stimuli [HDI95: −0.16 to 3.2 Hz], driven by a capacity increase of 2.6 Hz and a self-associated rate increase of 2.1 Hz [HDI95: 0.13 to 4.1 Hz]; other-associated stimulus rate showed no consistent increase (0.53 Hz [HDI95: −1.4 to 2.5 Hz]).
Requiring an explicit social-identity decision should attenuate, eliminate, or reverse the self-advantage seen in the perceptual decision condition — counter to the naive prediction that explicit attention to identity should amplify it.
Tested by
In Experiment 2, other-associated stimuli show a processing rate advantage over self-associated stimuli of -1.6 Hz [HDI95: -3 to -0.26 Hz] relative to neutral baseline in the social decision condition, replicating the cross-experimental finding that social decoding context reverses or eliminates the self-advantage.
In the social decision dimension (report whose shape flickered first), there is no processing advantage for self-associated stimuli: the relative advantage is 0.87 Hz [HDI95: −0.96 to 2.7 Hz], with 63.9% of the HDI favoring the other-associated stimulus (1.2 Hz [HDI95: −0.78 to 3.1 Hz]).
Social and perceptual salience capture attention via largely independent mechanisms, predicting additive (non-interacting) effects when both occur on the same stimulus.
For other-associated stimuli, social and perceptual salience effects should be additive in Experiment 2 — the interaction term should not reliably differ from zero if independence holds.
Tested by
The combination of social association and perceptual salience for other-associated stimuli produces effects consistent with additivity: the interaction term for other-associated salient stimuli is -0.54 Hz [HDI95: -2.4 to 1.4 Hz], indistinguishable from zero, indicating that social and perceptual salience operate via independent mechanisms for other-associated stimuli.
▸Dissociations
Across individuals, processing rate changes in the social decision dimension and the perceptual decision dimension are negatively correlated (r = -0.243, BF10 = 6.58), driven by differential allocation of relative attentional weights (r = -0.268, BF10 = 16.93): individuals who show stronger automatic self-prioritization at the perceptual level show less facilitation of self-associated information at the social level.
Individual self-prioritization effects in the shape-label matching task correlate with socially induced processing rate changes in the social decision dimension in Experiment 1 (r = 0.354, BF10 = 8.23, 95% CI: 0.11 to 0.54) but not in the perceptual decision dimension (r = 0.069, BF10 = 0.181).
In the social decision dimension (report whose shape flickered first), there is no processing advantage for self-associated stimuli: the relative advantage is 0.87 Hz [HDI95: −0.96 to 2.7 Hz], with 63.9% of the HDI favoring the other-associated stimulus (1.2 Hz [HDI95: −0.78 to 3.1 Hz]).
In the perceptual decision dimension (report which shape flickered first), self-associated stimuli were processed 1.5 Hz faster than other-associated stimuli [HDI95: −0.16 to 3.2 Hz], driven by a capacity increase of 2.6 Hz and a self-associated rate increase of 2.1 Hz [HDI95: 0.13 to 4.1 Hz]; other-associated stimulus rate showed no consistent increase (0.53 Hz [HDI95: −1.4 to 2.5 Hz]).
Perceptual salience benefit is substantially reduced for self-associated stimuli (2.5 Hz [HDI95: 0.86 to 4.1 Hz]) compared to other-associated (5.2 Hz [HDI95: 3.6 to 6.9 Hz]) or non-socially-associated perceptually salient stimuli (6 Hz [HDI95: 4.6 to 7.3 Hz]), indicating a sub-additive interaction when social self-relevance and perceptual salience co-occur.
The combination of social association and perceptual salience for other-associated stimuli produces effects consistent with additivity: the interaction term for other-associated salient stimuli is -0.54 Hz [HDI95: -2.4 to 1.4 Hz], indistinguishable from zero, indicating that social and perceptual salience operate via independent mechanisms for other-associated stimuli.
▸Eliminations & validating controls
Perceptual salience (local color contrast) produces a 6 Hz processing rate advantage for the salient stimulus [HDI95: 4.6 to 7.3 Hz] in Experiment 2, driven by a 2.5 Hz increase in the salient stimulus rate [HDI95: 0.8 to 4.2 Hz] and a 3.4 Hz decrease in the non-salient stimulus rate [HDI95: -4.9 to -2.0 Hz].
Strong self-prioritization effects are present in the shape-label matching task in both experiments: Experiment 1 (N=69) d = -1.064 [CI95: -1.38 to -0.75], BF10 = 3.23×10^95; Experiment 2 (N=71) d = -0.982 [CI95: -1.20 to -0.77], BF10 = 4.47×10^109, confirming participants learned and retained the self-associations used in the TOJ task.
▸Interpretations
Self-association enhances attentional selection automatically at the perceptual feature level — it emerges without explicit social decoding (perceptual decision condition, Experiment 1) and is reduced or reversed when social identity decoding is required (social decision condition, Experiments 1 and 2), indicating that self-prioritization in attentional selection is an early, pre-conscious process distinct from later social cognition.
Social salience and perceptual salience operate via largely independent mechanisms in attentional selection: their effects are additive for other-associated stimuli (interaction term -0.54 Hz, HDI includes zero), confirming that arbitrary social association and physical salience capture attention through separate processing streams rather than a single shared resource.
▸Synthesis claims
For perceptually salient self-associated stimuli, social salience effects are the stronger predictor (BFinclusion = 2458.52) while perceptual salience also contributes; for perceptually salient other-associated stimuli, perceptual salience dominates (BFinclusion = 4638.74), with social salience having a weaker role.
▸Standalone empirical findings
Processing capacity increases by 2.6 Hz [HDI95: -1.7 to 6.8 Hz] in the perceptual decision condition, with the self-associated stimulus showing a rate increase of 2.1 Hz [HDI95: 0.13 to 4.1 Hz]; the other-associated stimulus shows no consistent increase (0.53 Hz [HDI95: -1.4 to 2.5 Hz]), confirming the model selection favoring condition-specific capacity parameters.
▸Methodological warrants
Model comparison via leave-one-out cross-validation shows that a model with condition-specific processing capacity parameters outperforms a single-capacity model (Δloo=14.2, Δse=6.41, weight=0.86 pooled across N=140 participants), indicating that social salience changes absolute processing rates rather than merely redistributing attentional weights.
▸Scope qualifiers
All processing-rate claims derive from a TOJ + shape-label matching + hierarchical Bayesian TVA paradigm; preregistered N=69 (Exp 1) and N=71 (Exp 2); arbitrary shape-self associations only; OSF-deposited posteriors.
▸All claims (alphabetical)
- decisional-dimension-tradeoff fig9
- hypothesis-capacity-mechanism-not-weights hypothesis
- hypothesis-self-association-alters-attentional-selection hypothesis
- hypothesis-social-perceptual-independent-mechanisms hypothesis
- interprets-bundesen-tva-framework introduction, methods (TVA model variants), results (fig4, fig5–7), discussion
- other-association-advantage-social-condition fig6
- perceptual-salience-6hz-advantage fig6
- prediction-additive-effects-other-associated prediction
- prediction-capacity-model-outperforms-weights-only prediction
- prediction-self-advantage-attenuated-social-decision prediction
- prediction-self-advantage-perceptual-decision prediction
- processing-capacity-rises-perceptual-self fig5
- scope-toj-tva-paradigm scope
- self-prioritization-absent-social-decision fig5
- self-prioritization-automatic-early fig5 (synthesis)
- self-prioritization-perceptual-decision-automatic fig5
- self-salience-dominates-other-associated fig9 (Tables 1, 2)
- self-salience-reduces-perceptual-benefit fig6, fig7
- self-social-additive-perceptual fig7
- social-perceptual-salience-independent-streams fig7 (synthesis)
- spe-matching-correlates-social-decision fig9
- spe-robust-matching-both-experiments fig8
- tva-capacity-model-wins fig4
Abstract mapped to claims
The paper's abstract is shown with each sentence linked to the claim(s) it represents in the dependency graph. Hover or click a sentence to highlight the corresponding claim cards. Below: what the graph contains that the abstract leaves out, and vice versa.
1Efficiently processing self-related information is critical for cognition, yet the earliest mechanisms enabling this self-prioritization in humans remain unclear. 2By combining a temporal order judgement task with computational modeling based on the Theory of Visual Attention (TVA), we show how mere, arbitrary associations with the self can fundamentally alter attentional selection of sensory information into aware short-term memory, by enhancing the attentional weights and processing capacity devoted to encoding socially loaded information. 3This self-prioritization in attentional selection occurs automatically at early perceptual stages but reduces when active social decoding is required. 4Importantly, the processing benefits obtained from attentional selection via self-relatedness and via physical salience were additive, suggesting that social and perceptual salience captured attention via separate mechanisms. 5Furthermore, intra-individual correlations revealed an ‘obligatory’ self-prioritization effect, whereby self-relatedness overpowered the contribution of perceptual salience in guiding attentional selection. 6Together, our findings provide evidence for the influence of self-relatedness during earlier, automatic stages of attentional selection at the gateway to perception, distinct from later post-attentive processing stages.
- C2 spe-robust-matching-both-experiments fig8 Strong self-prioritization effects are present in the shape-label matching task in both experiments: Experiment 1 (N=69) d = -1.064 [CI95: -1.38 to -0.75], BF10 = 3.23×10^95; Experiment 2 (N=71) d = -0.982 [CI95: -1.20 to -0.77], BF10 = 4.47×10^109, confirming participants learned and retained the self-associations used in the TOJ task.
- H1.P1 prediction-capacity-model-outperforms-weights-only prediction A condition-specific TVA capacity model (indiv-C) should outperform a fixed-capacity, weights-only model on LOO cross-validation if the mechanism is capacity change rather than pure weight redistribution.