The hemodynamic response function (HRF) characterizes temporal variations of blood oxygenation level-dependent (BOLD) signals. Although a variety of HRF models have been proposed for gray matter responses to functional demands, few studies have investigated HRF profiles in white matter particularly under resting conditions. In the present work we quantified the nature of the HRFs that are embedded in resting state BOLD signals in white matter, and which modulate the temporal fluctuations of baseline signals. We demonstrate that resting state HRFs in white matter could be derived by referencing to intrinsic avalanches in gray matter activities, and the derived white matter HRFs had reduced peak amplitudes and delayed peak times as compared with those in gray matter. Distributions of the time delays and correlation profiles in white matter depend on gray matter activities as well as white matter tract distributions, indicating that resting state BOLD signals in white matter encode neural activities associated with those of gray matter. This is the first investigation of derivations and characterizations of resting state HRFs in white matter and their relations to gray matter activities. Findings from this work have important implications for analysis of BOLD signals in the brain.
1 August 2024 in Journal of Intelligent Information Systems
In recent years, an increasing body of research supports the "disconnection" hypothesis of schizophrenia, which posits that many of the clinical symptoms of the disorder are strongly associated with abnormal functional connectivity between brain regions, particularly those involving white matter (WM) dysfunction. However, most current schizophrenia detection methods mainly focus on signals or functional connectivity features in gray matter (GM) regions, overlooking functional associations involving WM regions and pathways. To address this gap, we propose a novel feature construction method, called Three-Way Cross-Correlation Coefficient (TW3C), incorporating WM functional connectivity for schizophrenia detection based on resting-state functional magnetic resonance imaging data. This method quantifies joint GM-WM-GM correlation patterns by calculating the cross-correlation of blood oxygen level-dependent (BOLD) signals between paired GM nodes and WM bundles. Experiments on two publicly available datasets, COBRE and UCLA, demonstrate that this method achieves mean cross-validated classification accuracies of 68.85% and 74.36%, respectively. Further analysis identified the bilateral superior longitudinal fasciculus, left cingulum, left posterior limb of the internal capsule, and right anterior limb of the internal capsule as key WM pathways associated with schizophrenia. Twenty-five highly discriminative GM-WM-GM tuples were identified across the two datasets, with recurrent involvement of the thalamus, cerebellum, sensorimotor, and limbic regions. Most tuples showed reduced TW3C values in schizophrenia, indicating altered joint correlation patterns involving GM and WM regions. Our findings indicate that incorporating WM functional signals provides a complementary perspective for schizophrenia classification and may facilitate further investigation of functional abnormalities involving GM and WM regions.
Multi-modal multi-label emotion recognition (MMER) aims to identify multiple co-occurring emotions from heterogeneous modalities. Existing methods tend to emphasize either modality complementarity or modality-specific information, while rarely achieving a balanced integration of both. Moreover, dependencies between modalities and emotion labels remain insufficiently explored, and the challenge of label imbalance in MMER has received little attention. To address these challenges, we propose COLIN (COmplementary and competitive baLanced learnIng Network), a unified framework designed to simultaneously balance modality complementarity and specificity while alleviating learning imbalance. Specifically, COLIN incorporates a Mixture-of-Experts (MoE) mechanism with attention to capture label dependencies and optimize multi-modal feature representations. It further adopts a dual-branch architecture that integrates modality-specific feature reconstruction with multi-level fusion, ensuring balanced utilization of information across different modalities. In addition, an improved loss function is designed to mitigate label imbalance and enhance overall model robustness. Extensive experiments on two benchmark datasets, CMU-MOSEI and \(\text {M}^3\text {ED}\), demonstrate that COLIN consistently outperforms state-of-the-art methods, validating its effectiveness in both modality balancing and robust multi-label learning.