❌

Normal view

Received β€” 27 May 2026 ⏭

MindAlign: Bridging EEG, Vision, and Language for Zero-Shot Visual Decoding

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Zexuan Chen Β· Sichao Liu Β· Runhao Lu Β· Huichao Qi Β· Alexandra Woolgar Β· Xi Vincent Wang Β· Lihui Wang
26 May 2026 at 12:00
arXiv:2605.24523v1 Announce Type: cross Abstract: Visual decoding from brain signals is a key challenge at the intersection of computer vision and neuroscience, requiring methods that bridge neural representations and computational models of vision. We introduce a tri-modal contrastive framework for EEG-based visual decoding that aligns EEG, visual, and textual representations within a unified latent space. Our approach follows a two-stage design. First, we pre-train an EEG encoder via masked reconstruction on unlabeled trials, learning spatio-temporal regularities that transfer robustly to downstream tasks. Second, we jointly align EEG, image, and LLM-generated textual descriptions through contrastive learning, where text supervision acts as a semantic regularizer that injects linguistic structure into the shared space without overwhelming the primary EEG-image signal. The encoder integrates subject-specific adaptation, graph-attention over channels, and temporal-spatial convolutional embeddings. On the Things-EEG2 200-way zero-shot benchmark, our framework achieves 54.1% Top-1 and 83.4% Top-5 accuracy, substantially exceeding the strongest prior baseline (32.4% / 64.0%), with paired Wilcoxon tests confirming significance (p

What Are We Actually Decoding? Source Attribution for Non-Invasive Brain-to-Language Retrieval

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Xinyu Zhang Β· Sichao Liu Β· Runhao Lu Β· Alexandra Woolgar Β· Lihui Wang
26 May 2026 at 12:00
arXiv:2605.24524v1 Announce Type: cross Abstract: In non-invasive neural language decoding, results can be inflated by sources that are not stimulus-evoked neural evidence: decoder priors, embedding-based metrics, and non-neural structural nuisances such as signal duration. The methodological challenge is therefore attribution: a reported gain is more informative when it can be traced to a specific source. We recast stimulus-locked MEG-to-audio retrieval as an auditing framework that separates apparent performance into three sources - structural shortcuts, window-level stimulus-locked evidence, and cross-window contextual aggregation - and provides a diagnostic for each. Signal-blind Gaussian noise reaches 66.3% Rank@1 (R@1) under variable-length decoding but collapses to near chance once fixed-duration windows and stimulus-identity splits are enforced, isolating structural leakage. Under these controls, fixed-window retrieval recovers measurable MEG-audio discriminability, while an oracle sentence-bucket diagnostic shows that 95.7% of Top-1 errors select the wrong sentence, localising the residual bottleneck to sentence-level competition. We audit this contextual source with Group Context Bias (GCB), an inference-time additive logit bias that pools sentence-consistent evidence across windows while leaving the base retrieval scores and candidate pool fixed. Used as a score-space intervention, GCB makes the contextual source measurable: R@1 shifts from 44% to 52% on Gwilliams and from 22% to 29% on MOUS under the same fixed setting. GCB is auditable under this design: its effect collapses under random-grouping perturbations and vanishes when local evidence is attenuated in MEG or is near chance in EEG, supporting its use as a controlled source-attribution intervention. These results suggest that brain-to-language performance should be source-attributed, not merely reported.

Data-driven precision: artificial intelligence redefining immunoradiotherapy in advanced pancreatic cancer

βœ‡(Multiomics OR Omics) AND (Pancreatic)
By: Yao-Wen Liu Β· Xiao-Ding Men Β· Bin-Ru Di Β· Yu-Lin Lei Β· Li Xiang Β· Yu-Hao Luo
25 May 2026 at 18:00

Front Pharmacol. 2026 May 8;17:1804673. doi: 10.3389/fphar.2026.1804673. eCollection 2026.

ABSTRACT

Advanced pancreatic ductal adenocarcinoma (PDAC) remains among the most formidable challenges in oncology, driven by a profoundly immunosuppressive tumor microenvironment (TME) and pervasive resistance to systemic and local therapies. Although immune checkpoint inhibitors (ICIs) can synergize with radiotherapy (RT) in several malignancies, the clinical benefit of immunoradiotherapy (iRT) in PDAC has been modest, highlighting the limitations of population-averaged paradigms that fail to capture extensive inter- and intratumoral heterogeneity. Here, we synthesize an artificial intelligence (AI)-enabled framework to refine both the biological rationale and clinical implementation of iRT for advanced PDAC through integrative analysis of multimodal data (clinical variables, imaging, RT dose distributions, and multi-omics). We highlight advances in three domains. First, AI-based deconvolution of TME heterogeneity can delineate clinically relevant molecular subtypes and spatial immune architectures that may be therapeutically tractable. Second, AI-driven modeling can optimize spatiotemporal RT-immunotherapy interactions, informing individualized dose, fractionation, and biologically guided target definition. Third, AI-supported predictive modeling and adaptive feedback can enable response-guided treatment adjustment beyond static planning. We also discuss unresolved clinical questions and key translational barriers, including data scarcity, lack of standardization, and limited interpretability. Finally, we outline priorities for translation-prospective digital biobanks, hybrid mechanistic-data-driven modeling, and adaptive trial designs-to enable rigorous validation and clinical deployment. Collectively, these developments position AI as a catalyst to move iRT for PDAC from empiricism toward real-time, individualized precision medicine.

PMID:42181887 | PMC:PMC13194001 | DOI:10.3389/fphar.2026.1804673

There are no more articles
❌