Aug 2026· International Conference on Systems· pp. 210-217· 0 citations· 19 references
Computer Science
TL;DR
The research rigorously investigates how neuromorphic architectures encode and integrate temporal information by conducting a comprehensive ablation study using a hybrid network, and demonstrates that shallow neuromorphic integration effectively maximizes the gains from temporal integration while mitigating the information loss inherent in binary spike quantization.
Abstract
While neuromorphic systems offer a promising path for processing dynamic, event-based data, current benchmarks often fail to isolate the specific impact of temporal integration on model performance. To address this, our research rigorously investigates how neuromorphic architectures encode and integrate temporal information by conducting a comprehensive ablation study using a hybrid network. We systematically transition from a fully spatial Convolutional Neural Network (CNN) to a fully Spiking Neural Network (SNN) by progressively replacing ReLU activations with spiking neurons across nine distinct model configurations. To challenge these architectures, we introduce the Temporal-ASL Dataset, a neuromorphic benchmark specifically curated with signs that exhibit high spatial isomorphism but distinct temporal signatures. This approach allows us to decouple spatial features from motion dynamics and quantify the marginal contribution of spiking membrane dynamics in resolving ambiguities that remain invisible to frame-isolated models. Our analysis reveals a performance hierarchy that peaks at 72.5% with a hybrid SNN configuration, representing a 6.25% improvement over the spatial CNN baseline. Logit trajectory analysis confirms this boost stems from the spiking layers’ ability to disambiguate spatially similar signs. However, accuracy declines steadily in deeper hierarchies, falling to 23.75% for the fully spiking SNN. Ultimately, these findings demonstrate that shallow neuromorphic integration effectively maximizes the gains from temporal integration while mitigating the information loss inherent in binary spike quantization.
Invasive neural recordings provide high-fidelity measurements of brain activity, with signals such as intracranial EEG (iEEG) and intracortical spiking activity capturing neural dynamics at different spatial and temporal scales. Yet existing neural foundation models have largely been developed independently for different invasive recording paradigms, leaving joint pretraining across heterogeneous invasive signals underexplored. In this work, we introduce iBrain, a unified foundation model that jointly learns from iEEG and spiking activity. iBrain employs signal-specific encoders to accommodate their distinct signal characteristics and a shared spatiotemporal Transformer backbone to model dependencies across recording channels and time. We pretrain iBrain on over 7,000 hours of heterogeneous neural recordings using masked signal reconstruction and channel-view alignment, promoting contextual modeling of neural dynamics and robustness across different channels. iBrain consistently outperforms single-signal pretraining baselines and achieves state-of-the-art performance on multiple benchmarks. Further experiments demonstrate that iBrain exhibits transferability and data efficiency across diverse recording settings. These results highlight the potential of joint pretraining on heterogeneous invasive neural recordings to support scalable neural modeling and transferable representations across recording settings and downstream tasks.
The brain uses discrete spikes for dynamic computation, yet, how neural microcircuits (NMCs) solve temporal credit assignment using local spike timing remains a fundamental open question. Dominant spiking neural network (SNN) approaches circumvent this by approximating backpropagation through surrogate gradients, decoupling learning from biological spike timing. Here, we reformulate temporal credit assignment as a state separation problem: extracting task-required components induced by historical perturbations directly from the current neural state. This enables an online feedback learning framework for NMCs through a gradient tunneling (GT) algorithm and the lead-lag expansion technique that derives credit assignment from local synaptic spike timing, while remaining compatible with ANN-SNN hybrid architectures. Experimentally, GT-trained NMCs excel at long-timescale evidence integration and noise-robust memory retention, and perform comparably to leading SNN online learning methods on real-world benchmarks with far fewer parameters. The proposed framework addresses the two-decade-old NMC feedback learning problem and suggests a computationally plausible explanation for the brain's learning mechanisms.
Xiangnan Zhang, Jingxin Liu, Ranqi Lu et al.· 0 citations
This work proposes a noisy group neuron (NGN) model, which incorporates population-level synchronous resetting and neural stochasticity as fundamental computational mechanisms, and develops the NGN method as a framework that combines the NGN model with backpropagation learning based on mean-field dynamics.
Yajie Zhai, Yanmei Kang, Meng Li et al.· 0 citations
Extensive experiments on static and neuromorphic benchmarks show that lower-bit BASC models match or outperform higher-bit baselines and retain this accuracy advantage after structured pruning, while further reducing model storage and synaptic operations.
Linliang Chen, Yan Zhong, Xin Liu et al.· 0 citations
Spiking neural networks (SNNs) have garnered significant attention in reinforcement learning tasks for their low power consumption. However, traditional spiking reinforcement learning (SRL) methods, which rely on local-connected encoding and fixed-threshold learning, struggle to capture the inter-dimensional correlations of input information within short timesteps, limiting the network’s expressive capacity at low timesteps. While increasing timesteps can significantly enhance performance, excessive timesteps result in substantial delays. To address this contradiction and enhance the expressive and decision-making capabilities of SNNs within short timesteps, we propose Mask-Adaptive Global Connection (MAGC), a novel encoding method that efficiently captures long-range dependencies via sparse, adaptively masked connections—enabling global feature interaction in a single timestep. Additionally, dynamic-threshold spiking neurons are introduced to effectively capture and distinguish subtle changes in input signals at each timestep, thereby enhancing the spatial-temporal state representation during spike information transmission. Extensive experimental results demonstrate that the proposed method achieves performance comparable to state-of-the-art algorithms using only a single timestep, while significantly reducing inference latency and energy consumption. When extended to multiple timesteps, our approach consistently outperforms existing methods, showing substantial improvements across eight continuous control tasks from OpenAI Gym.
Rong Xiao, Zhiyuan Hu, Ping He et al.· IEEE Transactions on Image P...· 0 citations
Whole-brain transcriptomic atlases are now widely available, yet computational neural models are almost exclusively parameterized from rodent data and used to infer human brain function, an extrapolation whose cost remains unquantified. To address this, we constructed a biophysically detailed, conductance-based Hodgkin–Huxley spiking microcircuit of a five-population prefrontal network, where every ion-channel, receptor, and gap-junction conductance was scaled by cell-type-specific gene expression. We parameterized the identical circuit using single-nucleus RNA-seq from mouse mPFC and human DLPFC, alongside a literature-derived baseline, and compared their high-frequency-oscillation (HFO) outputs across seven physiological and pathological states. While population firing rates differed only modestly between the two refinements (∼20% for pyramidal and PV cells), the oscillatory dynamics diverged dramatically. The human-refined circuit generated strongly synchronized PV activity and robust ripple- and fast-ripple-band power (e.g., healthy-wake ripple power, in arbitrary units: 322 vs. 24 and 22), whereas the mouse-refined and literature arms remained asynchronous (interneuron synchrony: 0.21 vs. 0.02). This human ≫mouse ≈ original hierarchy was statistically consistent across all seven states (significant arm differences in 75/77 comparisons). Mechanistically, the human transcriptome drove markedly stronger PV–PV electrical coupling (gap-junction scale: 1.78 vs. 0.96) paired with stronger recurrent pyramidal excitation, which collectively synchronized the fast-spiking PV population into a coherent rhythm that perisomatic inhibition then imposed on the local field potential. Critically, these results are model-dependent; the gene-to-conductance mapping is phenomenological, and mRNA expression does not linearly translate to functional conductance. Nonetheless, under this mapping the divergence localizes PV-mediated coupling and excitation–inhibition balance as the parameters most in need of human-specific recalibration. More broadly, this work establishes transcriptome-informed spiking simulation as a powerful strategy for uncovering species-specific computational principles and for building mechanistically grounded, human-relevant models of prefrontal circuit dysfunction, an approach that moves beyond generic rodent defaults to enable targeted, species-appropriate modeling of neurological and psychiatric disorders.
