Year: 2026
Type: preprint
Abstract: Detecting behaviorally relevant sensory changes requires the coordination of modality-specific sensory processing with shared attentional and control systems, yet how these processes are organized and evolve across large-scale brain networks remains unclear. Here, we investigated the connectivity architecture and temporal dynamics of auditory and visual target detection using fMRI acquired during matched oddball tasks. Cross-modal partial least-squares analysis revealed a low dimensional architecture comprising two dominant, modality-general dimensions and additional modality-biased components. The shared dimensions distinguished a sensory attention action configuration from a higher order associative and contextual configuration, whereas modality biased dimensions showed distinct organization involving frontoparietal and attentional networks. Projecting these connectivity patterns onto time-resolved BOLD activity revealed a common dominant latent space with graded regions of preferential occupancy across modalities. Overlap-adjusted event-related analyses further showed that auditory and visual targets were associated with a similar initial displacement along the first temporal latent dimension, followed by modality-dependent trajectories, particularly along the second dimension. This later divergence reflected different balances of regional contributions within a largely shared cortical architecture. By integrating dimensionality reduction of high dimensional functional connectivity with network and temporal analyses, our findings link the spatial organization of distributed cortical systems to their evolving responses, revealing how shared brain networks can support both common and modality-dependent aspects of sensory target detection.
Institutions: McGill University, Stanford Medicine, Stanford University
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7236769058
