New papers on Computational neuroscience
42 new papers on computational neuroscience in the last 7 days, within Brain & mind. These are the 42 Pipette rates most worth reading, with the main result in the authors' own words.
The best of the week
Electronic hair: Transforming hair into a scalable bioelectronic interface
Here, we introduce electronic hair (eHAIR), a strategy that converts hair into a bioelectronic interface using mixed ionic-electronic conductive polymer dyes.
Peer-reviewed journalBold claims, read criticallyReal-world useCounterfactual aging in individual digital twin brains uncovers a latent dimension of resilience to aging
Crucially, when every individual underwent the same counterfactual aging, those with a lower trait value declined more, yet recovered up to 1.6 times as much with a reduction in body mass index, a modifiable dementia risk factor.
PreprintBold claims, read criticallyA digital twin of freely moving mouse V1 learns a general visual code bridging classical physiology and active vision
By learning this representation from the freely behaving animal and probing it under controlled conditions, freely moving digital twins bridge classical visual physiology and active vision.
PreprintNeural dynamics track the approach to physical exhaustion
Participant-specific models built on neural dynamics tracked proximity to exhaustion with accuracy comparable to cardiorespiratory measures.
PreprintPreparatory Bias Currents Alone Write Letters and Words in a Fixed Recurrent Circuit
We show that preparatory input currents, or bias currents, alone can make a recurrent circuit with fixed weights write all 52 cursive letters, and words built from them.
Preprint"The self" as shifting coalitions of brain systems in a continuous configuration space
Self-related processing therefore has no core and arises from shifting coalitions of distributed brain systems whose graded contributions define a continuous configuration space.
PreprintProbabilistic mapping of brain connections without random sampling
We show that it is unnecessary: modeling tractography as an absorbing Markov chain gives the expected result of infinitely many streamlines.
PreprintBold claims, read criticallyReal-world useCode availableSpectral and complexity-based signatures of advanced concentration absorption meditation: A high-density EEG study of jhana
ACAM-J thus emerges as a non-pharmacological model state at the high-complexity end of the spectrum of altered consciousness, and reveals a previously unquantified dissociation between neural complexity and experiential richness for specific ACAM-J states.
PreprintTwo co-occurring dynamics explain the organization of a natural movement repertoire
Analysis of thousands of movements spanning classically defined bout types revealed that repertoire diversity arises from continuous combinations of two simultaneously expressed dynamics: a slow, non-oscillatory bending component and a fast, oscillatory undulatory component.
PreprintMultistage gradient-based whole-cortex modeling for joint inference of regional circuit heterogeneity and interregional structural connectivity
Here we introduce StaGIC (Staged Gradient-based Inference of Circuitry), a method that jointly infers regional circuit heterogeneity and model-constrained structural connectivity (SC) from resting-state functional connectivity (FC) and functional connectivity dynamics (FCD).
PreprintStructure learning and context inference emerge from next-state prediction in a cerebellar-medial temporal lobe circuit
While we only trained the circuit to predict next state, it learns to infer context and represents contexts in a structured latent space.
PreprintBold claims, read criticallyMyoGen: unified biophysical modeling of human neuromotor activity and resulting signals
We present MyoGen, an open-source framework that unifies spinal circuitry, proprioceptive feedback, musculotendon dynamics, and multimodal electromyography generation.
Peer-reviewed journalBold claims, read criticallyThe global geometry of optimal neural representations
We address the general problem by analytically deriving the neural representation that minimizes the all-pairs average classification error of a one-dimensional circular variable under various forms of neural response noise.
PreprintNeurons in a dish are the newest computing machines
Brain organoids offer highly efficient processing—if researchers can learn how to harness their power
Peer-reviewed journalBold claims, read criticallyBroad interestEasy to readDisentangling tonic and phasic dopamine in fiber photometry from single recordings to cohorts
Here we present PhotoSVI, a multi-scale statistical model that splits a recording into a slow tonic signal and discrete release events, estimated together.
PreprintIdentifying aligned tensor decompositions across heterogeneous neural datasets
We introduce Neuralign, a tensor- based model that jointly aligns and decomposes recordings into components, each pairing a shared temporal pattern with its expression in individual trials and neurons, without requiring matched neurons, matched trials or behavioural labels.
PreprintLearning predictive latent structure from unclustered neural spike trains
Here we introduce Contrastive Prediction of Point Process Observations (C3PO), an unsupervised method that learns a predictive, interpretable latent variable representation directly from continuous time neural spiking without requiring spike-sorting or binning.
PreprintBold claims, read criticallySubtractive and Divisive Methods for Isolating Neural Oscillations From Broadband Activity Yield Different, Even Conflicting, Power Estimates
Using simulated power spectra with additive oscillatory and aperiodic components, we illustrate that methods that assume a multiplicative model, such as the popular specparam (formerly fooof) toolbox, systematically mis-estimate oscillatory activity in ways that can obscure real oscillatory differences between conditions, or even reverse their sign, creating an illusory effect.
PreprintRepeated self-administered EEG recordings enable longitudinal measurement of brain activity at home
By demonstrating the feasibility of repeated self-administered EEG measurements, this study provides a step toward investigating how brain activity varies both between individuals and within individuals over time in everyday environments.
PreprintGeometric analysis of postnatal maturational transformations in brain's functional connectome
We show that greater postmenstrual age at birth was associated with emergence of the frontoparietal and salience networks at the tip of the unimodal-associative hierarchical axis; in contrast, greater postnatal age was associated with increasing polarisation of primary perceptual and sensorimotor systems across functional hierarchies.
PreprintPartially shared representational space for vision and language across objects, scenes, and visual events
In summary, we find a partially shared representational space for vision and language in high-level visual cortex, most clearly in regions representing social information related to people and their actions.
PreprintGeneralized hippocampal replay exhibits memory recall-like dynamics
Applying model comparison to three neural recording datasets from adult male rats, we find that the recall transition dynamics offer a statistically superior description of spiking during population burst events compared to a commonly used heuristic model of smooth trajectories.
PreprintBold claims, read criticallyShared-Encoder Residual Task-Bridging Network for Motor Execution to Motor Imagery EEG Transfer in Brain-Computer Interfaces
The proposed method achieves average classification accuracies of 69.39%, 74.61%, and 98.94%, respectively, and outperforms existing transfer-learning algorithms in overall performance.
PreprintReal-world useCode availableCan rat cortical cultures predict what comes next, unpredictable or learned?
Pre-stimulus activity thus carried no detectable information about the stimulus, and none emerged with repetition.
PreprintPROGMF: Progressive Multimodal Fusion and Representation Learning for Incomplete Neuroimaging Data
We introduce Progressive Multimodal Fusion (ProgMF), a novel fusion framework for learning multimodal source representations from incomplete neuroimaging data without imputing missing observations.
PreprintRestoring Marcenko-Pastur noise estimation and denoising of diffusion MRI under common image reconstruction
We restore the intended denoising behavior with a noise covariance model estimated from the reconstructed images, and validate the estimated model on repeated ex vivo and in vivo scans.
PreprintReal-world useRobust Feature Engineering Techniques for Designing Efficient Motor Imagery-Based Systems Using Multidomain Electroencephalography
These findings demonstrate that reducing 416 multidomain EEG features to four discriminative features using MRMR can maintain strong classification performance.
PreprintReal-world useState-Dependent Stimulation for Seizure Termination in a Morris-Lecar Network: A Reinforcement Learning Approach
These results provide a proof of concept that combining dynamical-systems-based state reconstruction with online reinforcement learning can identify effective stimulation timing for terminating seizure-like activity.
PreprintStability and flexibility in the Wisconsin Card Sorting Task emerge from goal-directed inference and planning under the free-energy principle
This implies that the two processes and this balance emerge from the FEP.
PreprintA Mechanistic Account of Dedifferentiation in the Aging Visual System
Severing connections or adding noise, whether in inferotemporal cortex (IT) or upstream, reduced IT differentiation, indicating that dedifferentiation need not originate locally.
PreprintWithin-Subject Precision Measurement: Theory and Utility of Assessing Change Scores and Their Variance
Here we describe the statistical theory behind estimating measurement properties from change scores.
PreprintInter-protocol reproducibility of multi-shell diffusion MRI metrics across major neuroimaging study protocols
These findings demonstrate good reproducibility of quantitative higher-order diffusion MRI metrics across commonly used multi-shell acquisition protocols, which can be further improved by aligning b-values, providing a foundation for data pooling, harmonization, and normative modeling in large-scale neuroimaging studies.
PreprintDynamic functional connectivity reveals evolving neural activity states of alcohol-seeking in rats
The inability of P rats to acquire stable neural states following devaluation suggests impaired formation of latent states as a neural mechanism of ARD, offering insights into the cognitive vulnerabilities that may sustain AUD.
Peer-reviewed journalWaveform morphology links local circuit kinetics to cortical interactions
Together, these findings show that waveform shape carries information about local neuronal dynamics, provide a mechanistic framework for interpreting cortical interactions, and motivate testing waveform morphology as a guide to stimulation timing.
PreprintTheoretical frameworks for credit assignment via behavioral timescale synaptic plasticity
To study how BTSP-like learning can be integrated into network models, we propose a generalized theoretical framework, gBTSP, which we apply to unsupervised and supervised tasks in feedforward and recurrent networks.
Peer-reviewed journalCircuit decoupling with surrogate compensation for baseline tuning in biophysically detailed neural models
Together, the results demonstrate that our method can facilitate baseline tuning in a biophysically detailed model by enabling decoupled optimization and guiding subsequent refinement.
PreprintA Variational Principle for the Emergence of Neural Geometry Underlying Bayesian-Like Behavior
We propose a variational principle in which learning balances task accuracy and representational smoothness under stimulus-dependent weighting.
PreprintBold claims, read criticallyA toolkit to identify parameter degeneracy and solution manifolds of Hodgkin-Huxley type current models
After fitting all model parameters simultaneously using synthetic data originating from a representative range of experimental cases, this reveals that a manifold of solutions is possible depending on the type of voltage clamp data available and on the selected HH-type model.
PreprintBold claims, read criticallyNeuroMASS v4: Geometry-Aware Neural Population Dynamics via Learnable Hyperbolic Latent Representations and Biologically-Calibrated Diagonal State Space Models
Our central scientific claim is that the curvature c of a Poincaré-ball--learned end-to-end from extracellular spike recordings--constitutes a biologically interpretable geometric signature: significantly positive for ring-manifold motor circuits (MC_Maze, MC_RTT) and near-zero for topologically unstructured frontal circuits (DMFC_RSG).
PreprintBold claims, read criticallyRobustness of cortical distortion metrics to unmodelled OPM errors
We found that the approach was tolerant to independent sensor position errors of less than 2 mm (per axis); orientation errors of less than 5 degrees; fixed array displacements of less than 10 degrees and to random sensor gain errors of less than 10%.
PreprintRank Deficiency in Parameter Estimation for Hodgkin-Huxley Type Current Models
Descriptions of the null spaces of these Jacobians are used to prove that the rank deficiency can be reduced or even be removed by fixing certain parameters, or by adding additional experiments.
PreprintNew tools for an old problem
How does one map, measure, and model the aging brain?
Peer-reviewed journalEasy to read