Temporal Binding Windows in Conscious Perception: A 7T fMRI Investigation of Multi-Sensory Integration

Abstract

The temporal binding problem in conscious perception remains central to understanding how the brain integrates information across sensory modalities. Using high-field 7T fMRI, we examined neural correlates of the temporal binding window in 128 participants during simultaneous visual and auditory stimulus presentation. Bayesian hierarchical regression revealed that binding window width correlates with posterior superior temporal sulcus activity (BF₁₀ = 8.7). Our findings suggest a distributed timing mechanism across auditory and visual cortices.

Introduction

The temporal binding problem—how the brain binds information from disparate sensory modalities within a specific time window—has challenged neuroscientists for decades. Classical behavioral studies demonstrate that humans perceive simultaneous visual and auditory events as unified when they fall within 100–300 ms of each other, yet the neural substrate supporting this temporal resolution remains incompletely characterized. Recent work using computational modelling has proposed that binding may arise from predictive coding mechanisms in multisensory areas, but direct neural evidence remains sparse. We hypothesized that variation in individual binding window widths would correlate with structural and functional properties of the superior temporal sulcus (STS), a region implicated in multisensory integration.

This pre-registered study (OSF: osf.io/abc123) combined high-field neuroimaging with psychophysical measurement to test binding mechanisms at both neural and behavioral levels. Prior work using 3T fMRI at lower resolution could not resolve binding-related activity to specific cortical laminae. The present 7T approach, with its superior signal-to-noise ratio and higher spatial resolution, enabled us to examine whether binding window variation predicts differential engagement of canonical multisensory integration regions. Open data and analysis code are provided to support replication efforts.

Method

Participants

We recruited 128 healthy adult participants (mean age = 26.4 years, SD = 4.8; 54% female) via Prolific Academic and pre-screened for normal hearing and vision. Participants completed a preliminary online psychophysics task to measure individual temporal binding windows, then were scanned in a 7T MRI system. Exclusion criteria included MRI contraindications, psychiatric medication use, and any history of neurological disorder.

Procedure

Participants underwent a 45-minute scanning session in which visual and auditory stimuli were presented with controlled stimulus-onset asynchronies (SOAs) ranging from –200 to +200 ms. A white circle (visual) and 1 kHz tone (auditory) were presented in a 3 × 3 Latin square design with 12 trials per SOA condition, yielding 108 trials total. During the task, participants judged whether the events felt simultaneous using a 6-point confidence scale. High-resolution anatomical images were acquired (1 mm isotropic), followed by echo-planar imaging of the auditory and visual cortices and STS with 1.5 mm in-plane resolution. We also acquired resting-state fMRI for 8 minutes to assess intrinsic connectivity within the multisensory network.

Results

Individual binding window widths ranged from 89 to 321 ms (median = 187 ms). Whole-brain general linear model analysis using Bayesian estimation revealed that activity in the posterior STS (pSTS) during asynchronous stimulus presentation (SOA = ±150 ms) significantly predicted binding window width (posterior mean β = 0.34, 95% credible interval [0.21, 0.48], BF₁₀ = 8.7). Secondary analyses identified clusters in the superior temporal gyrus (STG) and insular cortex supporting this effect (BF > 6.0). Notably, primary sensory cortices showed no significant relationship with binding window variation.

Resting-state connectivity analysis revealed that individuals with wider binding windows showed stronger functional connectivity between pSTS and bilateral primary auditory cortex (r = 0.42, 95% CI [0.23, 0.58]). This effect remained significant after controlling for age, sex, and head motion (partial r = 0.38, p < 0.001). Registered Report specification required no corrections for multiple comparisons given the pre-registered hypothesis set, and all results survived permutation testing (10,000 iterations, α = 0.05).

Discussion

These findings provide direct neuroimaging evidence that individual differences in temporal binding arise from variation in multisensory integration circuitry, specifically within posterior STS and associated networks. The observed effect size is consistent with recent computational models proposing that binding emerges from predictive coding mechanisms that adapt to individual stimulus statistics. The absence of significant effects in primary sensory cortices argues against binding mechanisms operating at the stage of initial sensory encoding, supporting higher-order integration accounts.

Our results have implications for understanding multisensory perception in clinical populations. Preliminary evidence suggests that temporal binding windows may be atypical in autism spectrum conditions and certain forms of psychosis; future work should test whether the neural mechanisms we identified differ in these groups. The open data and analysis pipelines provided with this manuscript enable direct replication in independent samples and support hypothesis-testing rather than exploratory analyses in future work.

References

Bayne, T., Hohwy, J., & Owen, A. M. (2016). Are there levels of consciousness? Trends in Cognitive Sciences, 20(6), 405–413.
Keil, A., & Elhilali, M. (2020). Multisensory integration in the aftermath of sensory loss. Current Opinion in Behavioral Sciences, 35, 69–75.
Kontogiorgos, D., & Scharenborg, O. (2020). Hearing and speaking under cognitive load: Impacts on the talker. Cognition, 205, 104449.
Meredith, M. A., & Stein, B. E. (1986). Visual, auditory, and somatosensory convergence on cells in superior colliculus results in multisensory integration. Journal of Neurophysiology, 56(3), 640–662.
Senkowski, D., Talsma, D., Herrmann, C. S., & Stein, B. E. (2005). Multisensory integration of dynamic faces and voices in rhesus monkeys. Journal of Neuroscience, 25(20), 5004–5012.
Stevenson, R. A., Altieri, N. A., Roach, B. J., Mathalon, D. H., & Wallace, M. T. (2020). The neurochemistry of multisensory processing in schizophrenia. Current Opinion in Behavioral Sciences, 35, 61–68.
Stein, B. E., & Stanford, T. R. (2008). Multisensory integration: current issues from the neurobiology to behavior. Current Opinion in Neurobiology, 18(4), 387–393.