Arnaud Delorme

Entropy- and complexity-based measures are promising tools to characterize neural dynamics beyond traditional linear and oscillatory analyses. However, their adoption in electroencephalography (EEG) and magnetoencephalography (MEG) research has been limited by methodological fragmentation, computational cost, and a lack of standardized implementations and visualization tools. Hence, we developed ASCENT, an open-source EEGLAB plugin for computing and visualizing a broad range of entropy, fractal, and aperiodic measures in neurophysiological time series. ASCENT integrates established single-scale, multiscale, and multivariate entropy algorithms, fractal dimension estimators, and aperiodic spectral parameterization within an optimized framework, accessible through both a graphical interface and a fully parameter-tunable command-line interface. To our knowledge, it is the first toolbox to combine these measures on independent component (IC) time series and on scalp channels, and to offer time-resolved estimation of both modified multiscale entropy and aperiodic parameters. Outputs are stored in a structured format for statistical analysis with external tools, as illustrated by an accompanying group-analysis script. As a usage example, we applied ASCENT to resting-state EEG from 40 participants recorded with eyes closed and eyes open. Condition differences were found for multiple measures and temporal scales. Their direction, spatial distribution, and scale extent depended on the measure, the coarse-graining operator and, for the power spectrum, aperiodic correction. Correlation analyses helped to indicate which measures were redundant. ASCENT aims to lower the threshold for researchers to apply these measures consistently and interpret them appropriately.

This pilot study investigated whether gastrointestinal activity, measured via electrogastrography (EGG), can be influenced nonlocally by a distant, emotionally bonded partner. Using a double-blind, sham-controlled design, thirty adult pairs were randomized to a Real or Sham condition. In the Real condition, the Sender experienced sad and food-related stimuli (selecting and eating food) while intending to transmit their experience to the distant Receiver; in the Sham condition, no stimuli were presented. Although the preregistered hypothesis predicting amplitude changes was not supported, a Linear Mixed-Effects analysis identified a significant Condition × Epoch interaction for EGG frequency. Follow-up tests showed unexpectedly large decreases in Receiver gastric frequency during the Sender’s “sad” epoch and more robustly during the “eating” epoch. This specific slowing of gastric rhythm aligns with parasympathetic (vagal) modulation rather than a generalized sympathetic response. These preliminary results suggest that EGG frequency may be a sensitive physiological marker of nonlocal interpersonal influence, particularly for enteroceptive states involving eating, although the underlying autonomic pathway cannot be determined from EGG alone.

Electromyography (EMG) is fundamental to clinical assessment, rehabilitation, neuromuscular research, and human-machine interfaces. Despite decades of use, no widely adopted standard exists for organizing and sharing EMG data, limiting reusability and large-scale data aggregation. We present EMG-BIDS, an extension to the Brain Imaging Data Structure (BIDS) that standardizes the organization of EMG recordings. EMG-BIDS addresses challenges unique to EMG, including diverse electrode types (surface or intramuscular, single channel to high-density arrays), heterogeneous electrode placements across anatomical locations, montages (e.g., monopolar or bipolar sensor designs), and the critical need for transparent documentation of sensor positioning. The specification introduces hierarchical coordinate systems that link local electrode grids to anatomical landmarks, enabling precise and reproducible placement documentation. EMG-BIDS is now part of BIDS as of version 1.11.0, supported by existing tools, including MNE-BIDS and EEGLAB. We demonstrate the specification through public datasets, including high-density surface EMG recordings. EMG-BIDS provides the foundation for FAIR (Findable, Accessible, Interoperable, Reusable) EMG data sharing, enabling meta-analyses, multi-site studies, and machine learning applications that require standardized, well-documented datasets.

This preregistered study analyzed more than 25 million trials from a web-based forced-choice remote viewing task to examine patterns of clairvoyant performance across all participants and among a subset of top performers. At the aggregate level, runs ending at the four planned lengths (5, 10, 25, or 100 trials) conformed to chance expectations. In contrast, optionally stopped runs displayed systematic fluctuations: short runs (1–3 trials) were above chance before declining, runs of 11–19 fell below chance, and runs beginning at 20 showed recurring above-chance spikes at every fifth run length (e. g., 30, 35, 40, 45, 50), which diminished beyond 80 trials. A Monte Carlo simulation matched to the empirical stopping distribution clarified the extent to which these patterns could be reproduced by optional-stopping behavior alone, with much of the run-length pattern, including the 11–19 trough and round-number variability, falling within the simulated null envelope. Exploratory analyses of top performers, defined post-hoc as the 1,235 users (2.64%) who exceeded chance uncorrected, after no users survived the preregistered FDR-corrected criterion, examined belief in psi, prior precognitive experience, meditation, total trials, and optional stopping as predictors. Optional stopping was the predictor most consistently associated with hits, both on first trials and across all trials, where it also interacted with belief, prior precognitive experience, and meditation jointly with cumulative task experience. Effect sizes were small (Δp and Cohen’s d near zero for most predictors), and findings are interpreted as exploratory. The results suggest that group-level outcomes primarily reflect optional-stopping and related behavioral dynamics, whereas top-performer analyses surface more nuanced, but small, context-dependent associations between belief, experience, and behavior. These findings highlight the methodological challenges of large-scale, open online testing and the value of preregistered, participant-level approaches, combined with benchmark simulations, for distinguishing behavioral artifacts from potential psi signals.

In an online sequential card-search task, the target card position was generated server-side and stored in a location inaccessible to the participant. Back-of-card images presented to the participant were identical across all five positions, and no target information appeared in any client-accessible form. A line-by-line code audit confirmed no pathway through which target information could reach the participant. Under these conditions, 106,996 participants contributed 35,945,094 trials from 2002–2018. Overall, they found the target in slightly fewer steps than expected by chance (mean = 2.9975 vs. 3.0; z = −10.73, Cohen's d = 0.0018). Although the deviation was extremely small, its direction and magnitude were estimated with high precision.Permutation tests preserved both the observed distribution of user click patterns and the exact distribution of generated targets while disrupting only the trial-level pairings between them. The observed mean fell more than 10 standard deviations below the permutation null mean, ruling out explanations based on interactions between user behavior and target distributions. Additional analyses found no evidence that target autocorrelation, duplicate records, pseudorandom algorithm artifacts, or behavioral strategies accounted for the effect.A preregistered analysis of data collected from 2021 – 2026 on an independently developed web platform found a consistent below-chance effect across approximately one million trials. However, data irregularities found in this study’s dataset precluded treating it as a satisfactory replication; it is best interpreted as convergent evidence, with primary evidential weight resting on the first study.Across both datasets, a small but precisely estimated deviation from chance was observed under conditions designed to preclude access to target information. The deviation survives permutation testing, code audit, and replication on an independent platform. Its origin remains unexplained.

Open data infrastructures are rapidly reshaping the landscape of neuroscience by enabling greater reproducibility through transparency, reusability, and collaboration, thus providing a foundation for sustainable and scalable scientific progress. This paper presents a systematic audit of the Neuroelectromagnetic Data Archive and Repository (NEMAR), a platform supporting the sharing and reuse of EEG, MEG, and iEEG datasets standardized through the Brain Imaging Data Structure (BIDS) format. Drawing on a detailed analysis of over 300 publicly available datasets, we assess the composition, metadata quality, and population coverage of data in NEMAR. Our findings reveal significant momentum in data sharing practices, particularly in EEG, and highlight key challenges, including inconsistent metadata, underrepresentation of clinical populations, and variable adherence to data standards. These insights highlight the importance of targeted support for policy and cultural development to ensure open repositories like NEMAR are not only accessible but also equitable, reusable, and scientifically robust.

Detecting single-trial P300 from EEG is difficult when only a few labeled trials are available. When attempting to boost a small target set with a large source dataset through transfer learning, cross-dataset shift arises. To address this challenge, we study transfer between two public visual-oddball ERP datasets using five shared electrodes (Fz, Pz, P3, P4, Oz) under a strict small-sample regime (target: 10 trials/subject; source: 80 trials/subject). We introduce Adaptive Split Maximum Mean Discrepancy Training (AS-MMD), which combines (i) a target-weighted loss with warm-up tied to the square root of the source/target size ratio, (ii) Split Batch Normalization (Split-BN) with shared affine parameters and per-domain running statistics, and (iii) a parameter-free logit-level Radial Basis Function kernel Maximum Mean Discrepancy (RBF-MMD) term using the median-bandwidth heuristic. Implemented on an EEG Conformer, AS-MMD is backbone-agnostic and leaves the inference-time model unchanged. Across both transfer directions, it outperforms target-only and pooled training (Active Visual Oddball: accuracy/AUC 0.66/0.74; ERP CORE P3: 0.61/0.65), with gains over pooling significant under corrected paired t-tests. Ablations attribute improvements to all three components.

Independent Component Analysis (ICA) is an important step in EEG processing for a wide-ranging set of applications. However, ICA requires well-designed studies and data collection practices to yield optimal results. Past studies have focused on quantitative evaluation of the differences in quality produced by different ICA algorithms as well as different configurations of parameters for AMICA, a multimodal ICA algorithm that is considered the benchmark against which other algorithms are measured. Here, the effect of the data quantity versus the number of channels on decomposition quality is explored. AMICA decompositions were run on a 71 channel dataset with 13 subjects while randomly subsampling data to correspond to specific ratios of the number of frames in a dataset to the channel count. Decomposition quality was evaluated for the varying quantities of data using measures of mutual information reduction (MIR) and the near dipolarity of components. We also note that an asymptotic trend can be seen in the increase of MIR and a general increasing trend in near dipolarity with increasing data, but no definitive plateau in these metrics was observed, suggesting that the benefits of collecting additional EEG data may extend beyond common heuristic thresholds and continue to enhance decomposition quality.