Brain decoding using functional magnetic resonance imaging (fMRI) has high potential for evaluating unobservable mental information that varies across individuals in everyday situations. Although estimating individual differences in mental information using brain decoding could be effective for social applications, especially in neuromarketing, the physical constraints and high costs of fMRI limit the practical application in real-world settings. To overcome these limitations, we propose a framework to estimate individual differences in decodable mental information using brain responses predicted by deep neural networks (DNNs). In this framework, DNN-based models are trained to predict fMRI responses to natural movie scenes. Then, additional fMRI measurements are no longer required because mental information for new movie scenes can be decoded from the predicted responses instead of the measured ones. A large-scale analysis revealed that the proposed predicted-response decoding captures individual difference patterns consistent with those derived from conventional measured-response decoding across various cognitive domains. Of the 88 cognitive labels associated with movie scenes, individual difference patterns could be captured for more than 70% of the labels. This framework offers a promising approach for evaluating personal mental information with substantially reduced reliance on fMRI, thereby expanding the potential for social applications of brain decoding.
Background Autism spectrum disorder (ASD) is hallmarked by pronounced inter-individual heterogeneity in brain function and clinical manifestations. Conventional neuroimaging studies yield highly inconsistent findings, as extrinsic confounding factors and intrinsic disease heterogeneity mask ASD-specific functional connectivity (FC) signatures, hampering the identification of individualized neural mechanisms linked to symptom severity. Methods We developed a digital twin control analytical framework based on deep contrastive learning. By comparing each ASD participant with their matched digital twin control, we quantified individual-specific FC deviations. Principal component analysis was applied to extract the core aberrant pattern, followed by multi-dimensional analyzes and validation including clinical correlation, transcriptomic decoding, neurotransmitter and mitochondrial co-localization, as well as disease and cognitive enrichment. Results The model effectively disentangled disease‑related from non‑disease‑related variations. The identified dominant aberrant FC pattern was significantly associated with Autism Diagnostic Observation Schedule (ADOS) communication scores, suggesting it is associated with to communication impairment. This pattern showed convergent associations with synaptic pathways, excitatory/inhibitory neuronal signatures, critical developmental windows and subcortical structures, and exhibited negative spatial association with multiple neurotransmitter systems and mitochondrial respiratory complexes. Biological validation further supported its ASD specificity and communication‑related cognitive relevance. Conclusions Leveraging the twin control analytical framework, this study provides an exploratory cross-scale framework linking individualized FC deviations with clinical, transcriptomic, neurochemical, and mitochondrial maps. The findings indicate that communication difficulties in ASD may partly arise from multi-scale disturbances in neural integration, social-perceptual processing and activity-dependent cellular energy regulation.
AIM To characterise short- and long-term plastic changes in cortical and subcortical structures induced by ultra-long (five-day) monocular deprivation (MD), and to assess the extent and regional specificity of structural changes and their reversibility following the restoration of binocular vision. METHODS In this single-arm longitudinal study, 21 healthy adults underwent five structural MRI sessions: at baseline, during MD on days 1, 3 and 5, and after a seven-day recovery period (day 12). MD was induced by occluding the right eye with a translucent eye patch for five consecutive days. Grey matter volume (GMV) changes were analysed using voxel-based morphometry (VBM) for whole-brain effects and generalised estimating equations (GEE) to model temporal trajectories in predefined cortical and subcortical regions. RESULTS Ultra-long MD induced dynamic GMV changes in primary and higher visual cortices, with exploratory changes also in subcortical and cerebellar structures. VBM showed cumulative volume increases in bilateral occipital cortex, including calcarine, lateral occipital and fusiform regions, and adjacent cerebellar lobules, peaking between days 3 and 5. Region-wise modelling resolved the temporal build-up of these changes across visual, subcortical and sensorimotor regions; after correction for multiple comparisons, a time-dependent effect remained in occipital and ventral occipito-temporal cortex (occipital pole, fusiform and calcarine cortex), converging with the voxel-wise result. Following binocular restoration, an exploratory GMV decrease was observed selectively in the left middle occipital gyrus, while an exploratory increase was observed in the left hippocampus, compatible with persistent plasticity in associative memory networks. CONCLUSION Prolonged sensory imbalance via ultra-long MD induces rapid, region-specific and possibly persistent GMV plasticity in the adult brain. These changes follow temporally stratified trajectories across visual, subcortical and integrative systems, consistent with models of homeostatic and hierarchical plasticity. Corrected evidence supports cumulative occipital plasticity during deprivation; whether the exploratory extra-visual and post-recovery changes reflect persistent adaptation requires confirmation. MD-based paradigms thus offer a controlled window on structural plasticity in both healthy individuals and clinical populations.
Recent advances in the physics of complex systems have motivated the study of higher-order structure in neuroimaging data, including functional magnetic resonance imaging (fMRI), electroencephalography (EEG), and magnetoencephalography (MEG). Despite the proliferation of proposed measures, the meaning of "higher-order" is often left implicit, and the term is used inconsistently across studies. In this work, we clarify the distinction between higher-order interactions and higher-order dependencies. Higher-order interactions necessarily imply higher-order dependencies, but the converse does not hold: systems can exhibit higher-order dependencies in the absence of higher-order interactions, as in the case of Gaussian variables. We present several equivalent definitions of higher-order interactions for continuous data, illustrate them using simple model systems, and use them to assess the specificity of measures commonly applied in neuroimaging. We show that these measures are not specific to higher-order interactions. Consequently, although existing studies support the presence of higher-order dependencies in neuroimaging data, they do not provide conclusive evidence for the presence of higher-order interactions.
Presurgical motor mapping in patients with tumors near eloquent motor cortex is commonly based on task-evoked activation maps, yet local BOLD responses may be difficult to interpret in lesion-affected tissue because of anatomical distortion, altered vascular coupling, and variability in task performance. In this study, we used a four-state motor paradigm spanning Action, kinesthetic motor imagery (IMG), Watch, and recent-action recall (REC) - to test whether coarse task-state relationships are preserved across paired preoperative fMRI and intraoperative ECoG in10 patients with brain tumors. Activation summaries and preoperative task familiarization were used to support task engagement, representational similarity analysis (RSA) was used to characterize within-modality state geometry, and cross-modal RSA was used to assess shared organization. Across modalities, Action and IMG showed relatively similar response patterns, Watch was more distinct, and REC occupied an intermediate position that should be interpreted cautiously because of the fixed task order. Cross-modal RSA showed positive correspondence between fMRI and ECoG representational dissimilarity matrices, supporting a shared coarse state structure rather than fine-grained equivalence between modalities. These findings support a hypothesis-generating, relation-based complementary framework for interpreting functional organization in patients with brain tumors, while the limited state space, clinically determined ECoG coverage, fixed task order, and indirect verification of internally generated states warrant cautious interpretation.
