Papers for

brain computer interface developers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Correlation aware model improves eeg brain signal decoding accuracy

CABrain: A Correlation-Aware Brain Foundation Model for EEG Decoding

Abstract: Electroencephalography (EEG) foundation models have emerged as promising paradigms for decoding brain activity through large-scale self-supervised pretraining. Notably, most existing approaches employ masked patch reconstruction, in which randomly selected EEG patches are masked and recovered by the network. However, purely random masking may limit representation learning as informative patches from highly correlated channels often remain unmasked, causing the network to underutilize weakly correlated channels or serial dependencies. Furthermore, because these models are typically based on transformer architectures, their performance is influenced by the choice of positional encoding. Although positional encoding has been extensively explored in natural language processing and computer vision, few studies have adapted it to the unique characteristics of EEG signals. To address these limitations, this study introduces a conditional positional encoding method based on graph convolutional networks (GCNs) to explicitly model inter-channel correlations, along with an adaptive masking strategy that enhances representation learning by modulating the difficulty of masked patch reconstruction. To the best of our knowledge, this is the first study to integrate graph-based representations into conditional positional encoding and introduce adaptive masking within EEG foundation model pretraining. Experimental results on ten benchmark datasets demonstrate that the proposed GCN-based positional encoding and adaptive masking strategies, both individually and in combination, achieve state-of-the-art or competitive performance, highlighting their effectiveness in EEG foundation modeling.

Mon 9 NovArtificial IntelligenceMachine Learning
The gist
Understanding brain signals from EEG recordings is tricky because the signals come from many channels that relate to each other in complex ways. The authors developed a new method that helps the computer better understand how different channels connect by treating EEG channels like points in a network. They also designed a way to train the model that makes it better at learning from harder-to-predict parts of the data. This approach helps improve the accuracy of decoding brain activity, as shown on multiple test datasets.
Open → 2611.11782v1

Model improves brain signal to image matching with detailed visual features

Look Closer, Align Better: Explicit Visual Attributes for Robust EEG-to-Image Decoding

Abstract: EEG-to-image retrieval commonly aligns neural responses with representations from pretrained vision encoders. However, these representations may underrepresent fine-grained visual details needed to distinguish semantically similar images. To address this limitation, we first investigate the EEG decodability and complementary predictive value of explicit visual attributes, identifying visual cues that capture EEG-relevant information beyond pretrained representations. Building on these findings, we propose AttrEcho, a framework that enriches pretrained visual representations with complementary, spatially aligned visual attributes. Specifically, Region-Aligned Visual Attribute Fusion (RAVF) integrates EEG-relevant local visual attributes with spatially corresponding pretrained features through a learnable signed residual gate. We further employ Feature Token Aggregation (FTA), which learns dimension-specific weights across regions to aggregate refined visual features for contrastive alignment with EEG representations. Experiments on THINGS-EEG demonstrate that AttrEcho achieves 94.8% Top-1 and 99.7% Top-5 accuracy in intra-subject 200-way retrieval, while maintaining 82.1% Top-1 accuracy as the gallery expands to 1,600 images with same-concept distractors. AttrEcho also achieves 25.3% Top-1 and 53.0% Top-5 accuracy in inter-subject retrieval. These results highlight the importance of EEG-relevant visual attributes and spatially aligned fusion for fine-grained EEG-to-image retrieval.

Fri 6 NovComputer Vision and Pattern Recognition
The gist
Matching brain signals to images often misses small but important picture details. The authors found that specific visual clues help connect brain patterns with the right images better than general image features alone. They created a new method, AttrEcho, that combines these detailed visual clues with traditional image features in a way that matches regions in the image and brain signals. Their method works much better at identifying the exact image someone saw just from their brain activity, even when many similar images are included. This shows that paying attention to specific visual traits and their locations matters for reading brain signals related to vision.
Open → 2611.09123v1

EEG signals converted into images using wavelet cross-attention and flow matching

E2I-Flow: EEG-to-Image Reconstruction via Wavelet-Temporal Cross-Attention and Conditional Flow Matching

Abstract: Electroencephalography (EEG)-based visual decoding and reconstruction aims to recover visual information from non-invasive brain signals. But there are challenges due to the complex temporal-frequency characteristics of EEG signals and the difficulty of discrete denoising framework in mapping neural activity to meaningful visual representations. Here, we propose E2I-Flow, an EEG-to-image decoding and reconstruction framework that combines a novel EEG encoder for time-frequency representation learning from EEG features, and conditional flow matching for visual decoding. A stationary-wavelet frequency branch and a temporal-mixing branch are integrated through bidirectional cross-attention and adaptive fusion, with the resulting EEG representations aligned to ground-truth image embeddings using contrastive and mean-squared-error loss supervisions. The aligned EEG representation is then used as a condition for a Transformer-based flow-matching model to predict the corresponding visual embedding for image synthesis. Experimental results show strong zero-shot EEG-to-image retrieval and improved visual reconstruction performance, outperforming several state-of-the-art baselines.

Tue 3 NovComputer Vision and Pattern RecognitionMachine Learning
The gist
Decoding pictures from brainwaves recorded via EEG is hard because EEG signals have complex patterns over time and frequency. The researchers created E2I-Flow, a method that learns important features of these brain signals using special wavelet and time-based processing combined through an attention mechanism. Then, it uses a flow-based model conditioned on these features to generate image representations that match the original visuals. This approach improved the accuracy of retrieving and reconstructing images from brain signals compared to prior methods.
Open → 2611.04524v1

Graph based eeg decoding predicts passenger danger in self driving cars

GraTED: Graph--Tensor EEG Decoding for Passenger Hazard Perception in Autonomous Vehicles

Abstract: Passenger electroencephalography (EEG) can reveal road risk through anticipatory responses before hazardous events and neural reactions around their onset. However, existing EEG decoding methods often map physiologically meaningful arrangement of electrodes onto artificial 2D grids and thus flatten the higher-order time-feature representations into matrices, obscuring meaningful structure in the data. To preserve and exploit the spatial and time-feature structure of EEG recordings, we propose the Graph-Tensor EEG Decoding Model (GraTED) for passenger risk prediction (RP) and danger identification (DI) in autonomous-vehicle scenarios. GraTED models the electrode arrangement as an undirected graph based on their scalp locations, using graph convolution to capture spatial dependencies and low-rank tensor classification to preserve the multidimensional time-feature structure, while controlling model complexity. Evaluated on the PassengerEEG dataset, comprising 11 participants and 45 hours of recordings, GraTED achieves the highest balanced accuracy, compared to the state-of-the-art, across all three evaluation settings, while requiring the fewest learnable parameters among seven baseline methods.

Mon 2 NovMachine LearningHuman-Computer Interaction
The gist
Passengers' brain signals can show when they sense danger on the road, even before anything happens. The authors created a new way to study these signals that keeps the natural layout of the brain sensors and the timing of the signals. Their method, called GraTED, uses a graph to represent sensors and a special technique to keep track of time and features. They tested this on real data and found it better at spotting risks than other methods while using fewer settings to learn.
Open → 2611.02827v1

Dataset enables decoding images from brain activity induced by sound

AVID: A 5T fMRI Dataset for Benchmarking Auditory-induced Visual Mental Imagery Decoding

Abstract: Visual mental imagery (VMI) decoding uses brain activity to reconstruct internally generated visual scenes, providing a potential pathway for externalizing internal experiences and developing neuroprosthetic communication interfaces. However, progress is limited by the lack of dedicated, training-scale VMI datasets. Current VMI paradigms often rely on visual prompts or short-term memory images, introducing visual and mnemonic confounds that make it difficult to disentangle neural activity associated with endogenous image construction from signals driven by prior or concurrent visual input. Furthermore, existing VMI datasets offer limited descriptive control over individual visual attributes, making it difficult to systematically assess fine grained visual details. To address these gaps, we introduce AVID, the first dedicated training-scale 5.0 Tesla functional magnetic resonance imaging (fMRI) resource for auditory-induced VMI decoding. AVID comprises approximately 67 hours of dense recordings from nine participants. The dataset is structured into two complementary splits: a scene-level complex split using naturalistic Mandarin descriptions and a controlled simple split with explicitly specified visual attributes. We further formalize a standardized evaluation protocol that decouples source-image consistency from semantic alignment, allowing for a more nuanced assessment of imagery reconstruction. Baseline qualitative and quantitative results show that current reconstruction pipelines remain limited on AVID, although the most successful reconstructions preserve recognizable scene level semantics from spoken descriptions. By combining dedicated training-scale neural recordings and standardized evaluation resources, AVID provides both a data foundation for imagery-specific model development and a benchmark for evaluating VMI decoding with substantially reduced sensory contamination.

Mon 2 NovComputer Vision and Pattern RecognitionArtificial Intelligence
The gist
Scientists created a new brain imaging dataset that records brain activity while people imagine pictures based on sounds, without showing any visuals. This helps separate brain signals from actual images versus those created inside the mind. They also set up ways to test how well computers can guess what people imagine from the brain scans. Early tests show current methods struggle, but some meaningful parts of the imagined scenes can be detected. This dataset will help improve brain-reading technology using sounds as triggers.
Open → 2611.02593v1

Dual path method improves cross subject emotion recognition from EEG

Beyond Modeling Order: Dual-Path Spatiotemporal Representation Learning for Cross-Subject EEG Emotion Recognition

Abstract: Cross-subject generalization remains a major challenge for EEG-based emotion recognition due to pronounced inter-subject variability in neural activity and emotion-related spatiotemporal patterns. Existing methods mainly alleviate inter subject differences through distribution alignment or contrastive learning, while paying limited attention to representation bias associated with fixed spatiotemporal modeling orders. We hypothesize that high-level emotion representations of the same EEG signal should remain consistent across different spatiotemporal modeling paths. Based on this hypothesis, we propose a Dual-Path Spatiotemporal Contrastive Learning (DP-STCL) framework for cross-subject EEG emotion recognition. DP-STCL employs two complementary encoders with independent parameters, namely a Temporal-to-Spatial (TS) path and a Spatial to-Temporal (ST) path, to model EEG spatiotemporal information under different integration orders. During representation learning, the same emotional stimulus is used to establish cross subject associations, while three complementary relations jointly constrain representation variations across subjects and modeling paths. This promotes stable and discriminative stimulus-related emotion representations. To further handle limited labeled target subject data, we introduce a LoRA-based few-shot adaptation strategy with low parameter overhead. Extensive cross-subject experiments on SEED, SEED-IV, SEED-V, and FACED show that DP-STCL consistently outperforms the compared methods, demonstrating its effectiveness in improving cross-subject EEG emotion recognition.

Mon 26 OctMachine Learning
The gist
Recognizing emotions from brain signals is hard because people's brains work differently. The authors suggest using two different ways to look at brain data—first focusing on how signals change over time, then across space, and vice versa. By making sure these two views agree when identifying emotions, their method learns better, more stable emotion patterns that work across different people. They also created a way to adapt the system quickly with only a little new data. Tests showed their approach works better than older ones.
Open → 2610.35540v1

Hyperbolic aggregation improves eeg brain region representation learning

HyBRA-EEG: Hyperbolic Global-Local Brain-Region Aggregation for Self-Supervised EEG Representation Learning

Abstract: Electroencephalography (EEG) representation learning is challenged by noisy signals, inter-subject variability, and heterogeneous electrode montages in different datasets and recording settings. Recent self-supervised EEG pretraining methods improve transferability by converting heterogeneous recordings into unified token or sequence representations, but channel-aware designs may still encode montage- or dataset-specific regularities. We propose HyBRA, a self-supervised framework based on hyperbolic global-local brain-region aggregation. HyBRA maps heterogeneous channel observations into a fixed regional token space by combining electrode coordinates, region priors, and feature-dependent hyperbolic distances. A channel-evidence injection branch preserves fine-grained discriminative cues during aggregation. The resulting tokens are processed by a Region-Time Transformer and pretrained with masked reconstruction of latent features and region-level signals. HyBRA outperforms the compared methods in average balanced accuracy on six benchmark datasets. Under a controlled six-fold evaluation on OpenBMI-MI, HyBRA matches CSBrain in balanced accuracy, improves AUROC by 1.84 percentage points and AUPRC by 5.23 percentage points, and uses 96.8\% fewer parameters. Matched comparisons on TUAB and TUEV show higher balanced accuracy for hyperbolic matching than for Euclidean distance or cosine similarity.

Mon 12 OctMachine Learning
The gist
Electroencephalography (EEG) recordings vary a lot between people and setups, making it hard for computers to learn from them reliably. The authors developed HyBRA, a new technique that groups EEG signals from different brain regions using hyperbolic geometry and extra information about where electrodes are placed. This approach creates stable signal summaries that work well across multiple EEG datasets. Tests show HyBRA is more accurate and efficient than previous methods.
Open → 2610.16589v1

Siamese deep learning detects ALS early from brain and eye data

Siam-ALS: An Explainable Siamese Deep Learning Framework for Early ALS Detection Using EEG and Eye-Tracking Signals

Abstract: Amyotrophic lateral sclerosis (ALS) diagnosis remains challenging due to heterogeneous symptoms and reliance on clinical assessment, underscoring the need for objective, data-driven approaches that leverage patient-generated physiological signals. We propose Siam-ALS, an explainable Siamese deep learning framework that jointly models synchronized electroencephalography (EEG) and eye-tracking (ET) signals. EEG captures ALS-related changes in brain activity, while ET provides information about oculomotor control, allowing the model to learn complementary patterns associated with ALS. The framework employs two modality-specific encoders: a convolutional EEG encoder that captures spatiotemporal neural patterns and an MLP-based ET encoder that models oculomotor dynamics. These modality-specific features are processed through weight-sharing Siamese branches and fused into a compact 64-dimensional joint embedding. A lightweight classifier is used to predict ALS, while pairwise metric learning helps the model learn more effective patterns from the small, diverse ALS datasets. Performance was evaluated on the public EEGET-ALS dataset using strict subject-level cross-validation, where Siam-ALS achieved an accuracy of 0.988 and an AUROC of 0.998, demonstrating strong performance compared with classical machine learning, DL, and non-Siamese fusion approaches. Cross-dataset evaluation on the independent BigP3BCI dataset achieved an AUROC of approximately 0.98, supporting the generalization capability of Siam-ALS across different BCI paradigms. To enhance interpretability, six XAI methods were evaluated in the learned embedding space, with SHAP achieving the highest, though statistically comparable, fidelity in identifying important EEG and ET patterns that contribute to ALS predictions. Overall, Siam-ALS provides an accurate, generalizable, and explainable framework for multimodal ALS detection.

Sun 11 OctMachine LearningArtificial Intelligence
The gist
ALS is a hard disease to diagnose early because it shows different symptoms in different people. The authors created a computer model called Siam-ALS that looks at brain signals (EEG) and eye movement patterns together to find signs of ALS. Their method uses two connected deep learning parts that learn from both types of data and combines them to spot ALS with high accuracy. The model also includes ways to explain what features it used to make its decisions. This approach was tested on multiple datasets and worked well, showing it can generalize across different settings.
Open → 2610.15351v1

Decoding dynamic video perception and recall from brain signals

Replaying the Movie in Your Mind Decoding Dynamic Visual Perception and Recall from EEG

Abstract: Decoding dynamic visual content from EEG has advanced rapidly for visual perception, yet it remains unclear which aspects of neural representations are preserved when people internally replay dynamic scenes. We introduce PAIR, an EEG benchmark collected from 20 participants using an interleaved, clip-wise paradigm that pairs naturalistic video viewing with immediate visual recall. This design enables a comparison between externally driven perception and internally generated recall for the same stimulus content within a fixed trial sequence. We benchmark multiple decoding models on hierarchical semantic classification at both fine-grained and coarse-grained levels, meta-feature decoding tasks covering motion, color, object number, face, and human presence, and a state-decoding task that discriminates perception from recall. Within-subject validation results show numerically similar performance for several model--task combinations, alongside higher perception accuracy in two coarse-grained comparisons after multiple-comparison correction. In contrast, perception and recall are highly separable at the state level under the present validation protocol. State accuracy remains high after alpha-band ablation, although this does not exclude ocular or other task-related contributions.

Fri 9 OctComputer Vision and Pattern Recognition
The gist
It is difficult to know what parts of what you see are remembered in your brain. The authors created a new brainwave (EEG) data set from people watching short video clips and then immediately remembering them. They compared brain activity when people saw videos and when they mentally replayed them, using computer models to try to tell the difference. The results show it’s possible to decode details about what’s seen and recalled, and that the brain signals for seeing and remembering are distinguishable. This work helps us understand how memories of visual scenes form and differ from actual perception.
Open → 2610.13120v1

Neural embeddings reveal personal traits beyond conversational roles

Contrastive Neural Embeddings Reveal Individual Traits Beyond Conversational Role

Abstract: Contrastive representation learning is increasingly used to recover low-dimensional structure from neural recordings, but its output is typically validated by decoding accuracy rather than by the geometry of the manifold it produces. We apply CEBRA to EEG recorded from dyads in conversation, and analyze the resulting embedding, which training constrains to the 2D sphere. Labels describing the dyads, including the absolute difference between partners' autism-quotient scores, decode well above chance (0.77 against a 0.55 majority baseline for binary AQ magnitude; 0.44 against 0.25 for the six-class $|Δ$AQ$|$ partition). However, the two permutation controls have notable differences in results: permuting labels over a frozen embedding yields p = 0.001, whereas retraining the encoder under each permutation yields p = 0.50. Only the latter tests the label rather than the geometry. Consistent with this, spherical mixture structure and per-class dispersion track identity rather than autism trait differences in dyads; frequency-band and non-oscillatory activity ablation controls do not change the results. However, participant-level model does separate from its identity-aware null (p = 0.0099) while speaker-versus-listener role analysis performs at chance in the same embedding, indicating a manifold organized by individual -- and, in contrast with current neurolinguistics models, almost invariant to speaking vs. listening. Based on these results, we suggest that retraining-based nulls should be the default for grouped-data contrastive embeddings.

Fri 2 OctMachine Learning
The gist
This paper examines how brain signals recorded during conversations can reveal more about individual traits than just the role someone plays in the talk, such as speaker or listener. The authors use a technique called contrastive learning to create two-dimensional representations of these brain signals. They find that these representations mainly reflect identity-related features rather than differences in traits like autism scores or speaking roles. They also caution that certain testing methods might overestimate what these embeddings reveal about specific labels.
Open → 2610.03410v1

Structured visual targets improve eeg image retrieval accuracy across people

Structured Visual Target Learning For Cross-Subject eeg-to-image retrieval

Abstract: Cross-subject EEG-to-image retrieval requires a neural represen- tation trained on source subjects to remain aligned with a visual embedding space for an unseen subject. Whereas existing methods primarily focus on the EEG side, we address this problem from the perspective of the visual target. Our approach preserves the spatial information of the Perception Encoder, converts its patch grid into a compact set of learned visual views, and aggregates them for each image with a block-structured, content-dependent router. The target is learned jointly with the EEG encoder through contrastive learning with MMD regularization across source subjects. For deployment, we propose a training-free representation refinement that aligns frozen embeddings without updating either encoder. Under leave- one-subject-out evaluation on THINGS-EEG2, the structured target achieves 35.3%/65.6% Top-1/Top-5 accuracy, the best among com- pared methods. Refinement raises this to 48.1%/77.1%, an 18.5% Top-1 gain over the strongest compared method, improving all ten held-out subjects.

Tue 29 SeptComputer Vision and Pattern Recognition
The gist
Matching brain signals (EEG) to images is hard because brain signals vary a lot between different people. The authors focus on improving how images are represented so that a system trained on some people can better understand brain signals from a new person. They do this by organizing visual information into meaningful parts and learning to match these with brain signals during training. Their method works better than previous ones, especially after a refinement step that improves alignment without retraining.
Open → 2609.36971v1

Frozen natural image autoencoders match fMRI models in brain trait prediction

Natural Image Autoencoder-Based fMRI Representations for Trait and State Prediction

Abstract: Foundation models pre-trained on large-scale fMRI datasets have shown strong downstream performance, but at substantial data and computation cost. To investigate how much fMRI-specific pre-training is actually needed for such performance, we introduce FReD, which derives fMRI representations from a frozen Deep Compression AutoEncoder (DCAE) pre-trained exclusively on natural images and pairs them with a task specific readout. For trait prediction, FReD summarizes frame-wise representations by their temporal mean and log-standard deviation and applies linear probing, with late fusion across two normalization schemes. For state prediction, it represents each frame as a single token and models temporal dependencies with a shallow Transformer. Across four resting-state datasets spanning six trait-prediction targets, linear probes on frozen DCAE features generally outperform those on fMRI foundation model representations and remain competitive with fully fine-tuned fMRI foundation models. On three task-fMRI state-prediction tasks, a temporal readout on DCAE features performs comparably to the strongest foundation models evaluated. A Gaussian injection analysis further shows that localized signal changes are recovered more accurately from the frozen DCAE features than from the evaluated foundation-model representations. Together, these results show that strong performance on current fMRI benchmarks is possible without fMRI-specific representation pre-training, making frozen natural-image features as a useful baseline for assessing its added value.

Mon 28 SeptComputer Vision and Pattern Recognition
The gist
Understanding brain activity using fMRI usually requires training special AI models on lots of brain data, which is expensive and time-consuming. The authors show that using an AI model trained only on natural images, without any specific brain training, can predict personal traits and brain states from fMRI data just as well as specialized models. This suggests simpler, cheaper ways to analyze brain scans without the need for heavy fMRI-focused training. Their approach might help researchers and developers compare new brain analysis methods to a strong baseline.
Open → 2609.34167v1

Neuron level growth improves some EEG convolutional decoders accuracy

Neuron-Level Architecture Growth: A Controlled Evaluation for EEG Time-Series Decoding

Abstract: Convolutional EEG decoders are trained at a fixed width, usually set by their authors on other data. Growing methods add neurons during training where the loss could decrease the most, but whether they improve compared to a reference width is untested on EEG. Here, we grow three convolutional backbones on 12 motor-imagery datasets under three protocols and compare each with its reference model per subject. The growing ShallowFBCSPNet scores 2.9 points above its reference model with only half the parameters (0.57x), SCCNet changes by at most 1.2 points. Deep4Net growing models show decreased accuracy, but they require adaptation that prevent to compare faithfully the results. These differences follow the selection step, which keeps a candidate neuron relying on a dynamic threshold from singular values decomposition. Overall, these results suggest that growth helps when its criterion can rank the candidate neurons, and that the rate of skipped neuron addition tells where a decoder can be grown small from scratch.

Sun 27 SeptHuman-Computer InteractionMachine LearningNeural and Evolutionary Computing
The gist
Many EEG decoding models are built with a fixed number of neurons, which might not be the best size for every dataset. The authors tested methods that add neurons during training only where it helps the model improve the most. They found that for some simpler brain activity decoders, growing the network helped improve accuracy and reduced the number of parameters needed. However, for deeper models, growing neurons did not improve performance as expected. This suggests that carefully deciding where to add neurons matters for improving EEG decoding.
Open → 2609.33880v1

Erp foundation model improves brain response analysis across datasets

ERP-FM: A Foundation Model for Universal ERP Analysis

Abstract: Foundation models have recently shown strong potential for learning generalizable EEG representations, yet their effectiveness for event-related potential (ERP) analysis remains unclear. In this work, we investigate two fundamental questions: 1) can foundation-model learning benefit ERP analysis, and what limits the transfer of existing EEG foundation models to ERP tasks? 2) can the complementary advantages of single-trial and averaged-trial ERP be integrated into a unified training pipeline? To study these questions, we curate a large-scale ERP corpus comprising 1,517,157 single-trial ERPs from 3,696 subjects across 38 datasets and 18 paradigms. Leveraging this corpus, we introduce ERP-FM, to the best of our knowledge, the first foundation model specifically developed for ERP representation learning. ERP-FM uses single-channel tokenization, temporal and spatial positional embeddings, and mixed masked autoencoding for large-scale single-trial pretraining. We compare our model against 17 existing methods on 12 downstream datasets covering ERP event/condition classification and neurological disease classification. Our model achieves the best overall average rank across all evaluated methods. Furthermore, our analyses reveal that both ERP and non-ERP EEG pretraining can provide transferable representations for ERP tasks, while fine-grained temporal tokenization is critical for effectively modeling transient ERP dynamics. We further find that single-trial and averaged-trial ERP play complementary rather than competing roles. Combining single-trial pretraining with averaged-trial downstream adaptation substantially improves neurological disease analysis. Overall, these findings establish an effective foundation-model training pipeline for ERP analysis and represent significant progress toward generalizable ERP representation learning. Source code: https://github.com/DL4mHealth/ERP-FM

Sat 26 SeptComputational Engineering, Finance, and Science
The gist
Analyzing brain signals called event-related potentials (ERPs) can be tricky because they change quickly and vary among people. The authors created a special AI model named ERP-FM that learns from lots of ERP data across many people and experiments. This model helps computers better understand and classify different brain responses and can even improve detecting neurological diseases. They found combining data from individual brain signal events and averaged events makes the model work better. Overall, ERP-FM offers a new way to study brain responses more accurately across different tasks and people.
Open → 2609.32796v1

Dual-invariance learning improves eeg workload recognition across people

CDBG: Causally Motivated Dual-Invariance Learning against Topological and Predictive Shifts in EEG Workload Recognition

Abstract: Generalizing Electroencephalography (EEG)-based mental workload recognition to unseen subjects remains a formidable challenge due to severe inter-subject variability. While functional brain graphs effectively model distributed cognitive dynamics, their inherent subject-specificity induces two coupled distribution shifts: a class-conditional topological shift in the underlying functional connectivity, and a predictive mechanism shift in the learned representation-to-label mapping. Motivated by the subject-induced distribution shifts, we propose CDBG, a Causally motivated Dual-invariance learning framework for Brain Graphs. CDBG disentangles and mitigates these shifts via a two-stage rationale learning pipeline. First, it employs stochastic edge masking to extract sparse, workload-predictive graph rationales, regularized by workload-conditional Laplacian spectral alignment to enforce topological invariance across subjects. Second, it applies subject-wise Invariant Risk Minimization (IRM) to the graph representations, ensuring environment-wise risk stationarity. Extensive experiments on a self-built air traffic controller EEG cognitive workload dataset and multiple public datasets under a strict leave-one-subject-out protocol demonstrate that CDBG significantly outperforms state-of-the-art cross-subject and graph-based baselines, improving the Macro-F1 score by up to 4.23%, while simultaneously providing neurophysiologically interpretable functional rationales.

Fri 25 SeptHuman-Computer Interaction
The gist
It is hard to use brain signals from EEG to recognize mental workload on new people because everyone's brain connections are a little different. The paper's authors developed a method called CDBG that learns to ignore these differences by focusing on stable patterns in brain connectivity and prediction rules. Their method works in two steps: first, it finds important brain connection patterns that stay consistent across people; second, it trains a model to make predictions that remain reliable no matter who the data comes from. Their tests show this helps recognize workload more accurately on new subjects, and it also helps understand which brain connections are involved.
Open → 2609.30831v1

Brain image retrieval improves by using multiple visual hierarchy levels

The Visual Target Matters: Learning across the Visual Hierarchy for Brain-to-Image Retrieval

Abstract: Brain-to-image retrieval seeks to identify the visual stimulus that elicited a non-invasive neural response. Candidate images are typically represented by pretrained vision models, whose internal representations vary in abstraction across depth. Existing methods usually train the neural encoder to recover a fixed final-layer visual target. Under this formulation, the visual hierarchy is reduced to a single prescribed endpoint, preventing representations at other depths from directly shaping the visual target. This limitation motivates learning how information across visual depths should contribute to the retrieval target. To this end, we introduce NeuroGlyph, which learns a trial-independent visual target from multiple depths of a frozen visual backbone. NeuroGlyph decomposes the target into factor-specific subspaces. Each subspace learns an image-conditioned allocation over visual depth. The resulting subspaces are fused into a single embedding for retrieval. Across THINGS-EEG and THINGS-MEG, NeuroGlyph outperforms final-layer supervision in all controlled comparisons. It also surpasses the post hoc best fixed-layer oracle in three of four comparisons. Parameter-matched ablations support both factorized target construction and image-conditioned depth allocation. Under comparable 200-way retrieval protocols, NeuroGlyph achieves the strongest system-level performance in six of eight reported metrics. These results support learning retrieval targets across the visual hierarchy rather than prescribing one visual depth.

Mon 21 SeptComputer Vision and Pattern Recognition
The gist
Figuring out which image someone saw from their brain signals is tricky. Previous methods looked only at the most abstract visual features in a model. These authors show that using information from several levels of visual processing helps find the right image better. They developed a method called NeuroGlyph that combines different feature levels and changes how much each level contributes based on the image. Their tests show this approach outperforms older ways that focus on just the final features.
Open → 2609.24136v1

Hierarchical model improves word retrieval from non-invasive brain signals

HDND: Hierarchical Dynamic Neural Decoding for Multilingual Word/Character Retrieval from Non-Invasive Brain Recordings

Abstract: While deep learning has enabled language decoding from intracranial brain recordings, extending this capability to non-invasive recordings remains an unresolved challenge. Decoding individual words from non-invasive brain recordings is particularly difficult, as word-level neural evidence is weak, temporally distributed, and entangled with acoustic, lexical, and semantic structure. Existing retrieval pipelines often collapse these factors into a single representation, potentially discarding information available at intermediate temporal scales. Here, we introduce Hierarchical Dynamic Neural Decoding (HDND), a hierarchical dynamic decoding framework that treats word decoding as structured refinement rather than flat label retrieval. HDND combines intermediate neural representations, contextual semantic predictions, and, for selected reading conditions, an auxiliary character-form objective. We evaluate HDND across seven electroencephalography (EEG) and magnetoencephalography (MEG) datasets spanning English, Dutch, Mandarin, and Cantonese listening, reading, and reading-aloud conditions. Across the nine-condition word-retrieval benchmark, the proposed HDND yields a higher participant-averaged balanced Top-10 point estimate than the matched contextual word-decoding baseline in every condition and achieves the highest mean among all compared methods in eight of nine conditions. Across the same nine matched conditions, HDND also yields higher token-micro and pooled word-macro Top-10 point estimates in every setting. Sentence retrieval favors HDND in eight of nine conditions, while auditory speech-segment retrieval is mixed across the six listening conditions. These results show that hierarchical residual refinement can improve multilingual word retrieval from heterogeneous non-invasive brain recordings.

Mon 21 SeptComputer Vision and Pattern Recognition
The gist
Decoding individual words from non-invasive brain recordings is hard because the brain signals are weak and mixed with many other factors. The authors developed a new method called HDND that breaks down the decoding process into smaller steps, refining the word guess over time instead of guessing all at once. They tested HDND on brain signals recorded while people listened to, read, or spoke words in four different languages. This method consistently outperformed previous approaches in identifying the correct words from brain activity. The results suggest that using hierarchical, stepwise models can better capture the complex brain signals involved in language.
Open → 2609.24095v1