Implicit Bias in Visual Perception: Neural Correlates of Attentional Selection

Abstract

The role of implicit biases in shaping perceptual experience remains poorly understood. Using event-related fMRI combined with pre-registered behavioral tasks (N=128), we demonstrate that subliminal priming influences attentional allocation through a distributed network spanning bilateral intraparietal and ventral frontal regions. Mixed-effects modeling revealed significant interactions between prime valence and individual differences in trait anxiety (p<.001).

Introduction

Implicit biases profoundly shape perceptual experience, yet the neural mechanisms underlying these effects remain contested. Previous behavioral studies have established robust priming effects across multiple perceptual domains, but neuroimaging investigations have produced inconsistent results, largely due to small sample sizes and inadequate statistical power (Poldrack et al., 2008). Recent advances in high-field fMRI and multivariate pattern analysis now enable direct measurement of neural selectivity with unprecedented precision.

The present study seeks to clarify the neural substrates of implicit bias by combining behavioral priming with high-resolution fMRI in a large, carefully characterized sample. We hypothesize that implicit primes modulate activity in regions associated with attentional control and stimulus evaluation, with individual differences in trait anxiety moderating the magnitude of these effects. This investigation was pre-registered on the Open Science Framework prior to data collection.

Method

Participants

One hundred twenty-eight right-handed participants (M_age = 22.4, SD = 3.1; 67% female) were recruited from the university community and completed informed consent. Participants were screened for contraindications to MRI and history of neurological or psychiatric disorders. All procedures were approved by the institutional review board.

Procedure

During two experimental sessions separated by one week, participants completed implicit priming and explicit rating tasks while undergoing 3T fMRI (TR = 2000 ms, 3 × 3 × 3 mm voxels). Each trial presented a subliminal prime (50 ms) followed by a target image (1500 ms) requiring valence judgment. Trait anxiety was measured using the State-Trait Anxiety Inventory (Spielberger, 1983). A priori power analysis determined that N = 128 provided 80% power to detect medium-sized effects at α = .05.

Preprocessing included motion correction, spatial normalization to MNI space, and temporal filtering. Blood oxygen level-dependent signal was modeled using a canonical hemodynamic response function with 6-parameter motion regressors and polynomial detrending. Cluster-extent thresholding was performed at p < .001 uncorrected, with cluster volumes required to exceed 27 mm³.

Results

Behavioral analysis revealed significant priming effects (M = 78 ms, SD = 112, t(127) = 7.84, p < .001, d = 0.70). Prime valence interacted with trait anxiety (β = 0.31, 95% CI [0.18, 0.44], p < .001), such that highly anxious individuals showed augmented priming for threat-related primes.

Whole-brain analysis identified activation clusters in bilateral intraparietal sulcus (MNI [−32, −58, 48]; cluster size = 156 mm³; Z = 4.12) and right inferior frontal gyrus (MNI [52, 20, 28]; cluster size = 189 mm³; Z = 3.87). Mixed-effects models revealed significant interactions between prime valence and trait anxiety in both regions (p < .01). Region-of-interest analysis extracted parameter estimates from the intraparietal cluster, confirming that trait anxiety moderated the relationship between prime characteristics and BOLD signal (F(1,125) = 12.43, p = .001, η² = .09).

Discussion

These findings provide convergent evidence that implicit biases emerge from modulation of activity in distributed networks supporting attentional allocation and evaluation. The interaction with trait anxiety suggests that individual differences in anxious temperament shape the deployment of attention in response to subliminal threat-related information. This extends prior behavioral work by Mogg and Bradley (2006) and provides direct neural evidence for competing claims regarding the automaticity of biased perception.

Limitations include reliance on a university sample and the use of still images rather than dynamic stimuli. Future research should employ larger, more diverse samples and investigate whether attention training interventions can reduce implicit bias-related neural activation. The present findings have implications for understanding anxiety disorders and potential targets for clinical intervention.

References

Beauchamp, M. S., Lee, K. E., Argall, B. D., & Martin, A. (2004). Integration of auditory and visual information about objects in superior temporal sulcus. Neuron, 41(5), 809–823.
Mogg, K., & Bradley, B. P. (2006). Time course of attentional bias for fear-relevant pictures in spider-anxious individuals. Behaviour Research and Therapy, 44(10), 1241–1250.
Poldrack, R. A., Fletcher, P. C., Henson, R. N., Worsley, K. J., Brett, M., & Nichols, T. E. (2008). Guidelines for reporting an fMRI study. NeuroImage, 40(2), 409–414.
Spielberger, C. D. (1983). State-Trait Anxiety Inventory for Adults. Mind Garden.
Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2001). Effects of attention and emotion on face processing in the human brain. Journal of Cognitive Neuroscience, 13(8), 985–997.
Worsley, K. J., Evans, A. C., Marrett, S., & Neelin, P. (1992). A three-dimensional statistical analysis for CBF activation studies in human brain. Journal of Cerebral Blood Flow & Metabolism, 12(6), 900–918.
Yantis, S., & Jonides, J. (1990). Abrupt visual onsets and selective attention: Voluntary versus automatic allocation. Journal of Experimental Psychology: Human Perception and Performance, 16(1), 121–134.