Parallel Information Research Articles

Activity-dependent synaptic competition and dendrite pruning in developing mitral cells

During the early postnatal period, neurons in sensory circuits dynamically remodel their connectivity to acquire discrete receptive fields. Neuronal activity is thought to play a central role in circuit remodeling during this period: Neuronal activity stabilizes some synaptic connections while eliminating others. Synaptic competition plays a central role in the binary choice between stabilization and elimination. While activity-dependent “punishment signals” propagating from winner to loser synapses have been hypothesized to drive synapse elimination, their exact nature has remained elusive. In this review, I summarize recent studies in mouse mitral cells that explain how only one dendrite is stabilized while others are eliminated, based on early postnatal spontaneous activity in the olfactory bulb. I discuss how the hypothetical punishment signals act on loser but not winner dendrites to establish only one primary dendrite per mitral cell, the anatomical basis for the odorant receptor-specific parallel information processing in the olfactory bulb.

Multifunctional Artificial Electric Synapse of MoSe2-Based Memristor toward Neuromorphic Application.

Research on memristive devices to seamlessly integrate and replicate the dynamic behaviors of biological synapses will illuminate the mechanisms underlying parallel processing and information storage in the human brain, thereby affording novel insights for the advancement of artificial intelligence. Here, an artificial electric synapse is demonstrated on a one-step Mo-selenized MoSe2 memristor, having not only long-term stable resistive switching characteristics (reset 0.51 ± 0.01 V, on/off ratio > 30, retention > 103 s) but also diverse electrically adjustable synaptic behaviors, including multilevel conductance (synaptic weight), excitatory postsynaptic current (EPSC), paired-pulse facilitation (PPF), long-term potentiation/depression (LTP/D), spike-timing-dependent plasticity (STDP), and especially activity-dependent synaptic plasticity (ADSP). More significantly, neuromorphic functions of both image edge extraction and biological perception imitation have been successfully achieved. These results present a promising design toward synaptic devices for advancing neuromorphic systems with integrated brain-like neural sensing, memory, and recognition.

Incremental accumulation of linguistic context in artificial and biological neural networks

Large Language Models (LLMs) have shown success in predicting neural signals associated with narrative processing, but their approach to integrating context over large timescales differs fundamentally from that of the human brain. In this study, we show how the brain, unlike LLMs that process large text windows in parallel, integrates short-term and long-term contextual information through an incremental mechanism. Using fMRI data from 219 participants listening to spoken narratives, we first demonstrate that LLMs predict brain activity effectively only when using short contextual windows of up to a few dozen words. Next, we introduce an alternative LLM-based incremental-context model that combines incoming short-term context with an aggregated, dynamically updated summary of prior context. This model significantly enhances the prediction of neural activity in higher-order regions involved in long-timescale processing. Our findings reveal how the brain’s hierarchical temporal processing mechanisms enable the flexible integration of information over time, providing valuable insights for both cognitive neuroscience and AI development.

Multi-functional broadband diffractive neural network with a single spatial light modulator

Diffractive neural networks (DNNs) are emerging as a novel optical computing architecture that combines wave optics with deep-learning methods for high-speed parallel information processing. Herein, we report a reflection type, multi-functional, broadband DNN design. It consists of two phase-modulation layers based on a single spatial light modulator and a mirror facing it. The power efficiency of this design is more than 16 times higher than that of the cascaded structure utilizing beam splitters. It can function either as a two-layer DNN or a one-layer DNN with the other serving as an information input layer. Single- and dual-wavelength filtering and focusing, as well as spatial wavelength demultiplexing of supercontinuum, are experimentally demonstrated using the two-layer DNN, whereas the one-layer DNN is experimentally demonstrated by the classification of hand-written digits, which are input by the first layer via holographic imaging. The designed DNN could operate independently or be readily integrated with other optical systems and may find applications in spectroscopy, microscopy, and information technology.

Distinct retinal ganglion cell types in strictly subterranean, naturally microphthalmic mammals.

African mole-rats (Bathyergidae, Rodentia) are subterranean rodents that live in extensive dark underground tunnel systems and rarely emerge aboveground. They can discriminate between light and dark but show no overt visually driven behaviours except for light-avoidance responses. Their eyes and central visual system are strongly reduced but not degenerated. Here, we focus on retinal ganglion cells (RGCs). Sighted mammals have numerous RGC types with distinct morphological and functional properties that encode different aspects of a visual scene. We analysed the morphological diversity of 216 intracellularly dye-injected RGCs in the giant mole-rat (Fukomys mechowii) and 48 RGCs in Ansell's mole-rat (Fukomys anselli). Using a hierarchical cluster analysis on 11 morphological parameters, we show that both species possess at least five RGC types with distinct dendritic field sizes and branching patterns. These resemble some RGC types of the mouse and rat, but mole-rat RGCs feature overall sparser and more asymmetric branching patterns. The dendritic trees of most RGCs in all clusters are monostratified in the inner plexiform layer, but bistratified and multistratified/diffuse cells also exist. Thus, although RGC morphologies have become disorganized, the basic retinal organization principle of parallel information processing by distinct RGC types is retained.

The human auditory cortex concurrently tracks syllabic and phonemic timescales via acoustic spectral flux.

Dynamical theories of speech processing propose that the auditory cortex parses acoustic information in parallel at the syllabic and phonemic timescales. We developed a paradigm to independently manipulate both linguistic timescales, and acquired intracranial recordings from 11 patients who are epileptic listening to French sentences. Our results indicate that (i) syllabic and phonemic timescales are both reflected in the acoustic spectral flux; (ii) during comprehension, the auditory cortex tracks the syllabic timescale in the theta range, while neural activity in the alpha-beta range phase locks to the phonemic timescale; (iii) these neural dynamics occur simultaneously and share a joint spatial location; (iv) the spectral flux embeds two timescales-in the theta and low-beta ranges-across 17 natural languages. These findings help us understand how the human brain extracts acoustic information from the continuous speech signal at multiple timescales simultaneously, a prerequisite for subsequent linguistic processing.

Autonomous and ubiquitous in-node learning algorithms of active directed graphs and its storage behavior

The brain’s memory system is extraordinarily complex, evidenced by the multitude of neurons involved and the intricate electrochemical activities within them, as well as the complex interactions among neurons. Memory research spans various levels, from cellular and molecular to cognitive behavioral studies, each with its own focus, making it challenging to fully describe the memory mechanism. Many details of how biological neuronal networks encode, store, and retrieve information remain unknown. In this study, we model biological neuronal networks as active directed graphs, where each node is self-adaptive and relies on local information for decision-making. To explore how these networks implement memory mechanisms, we propose a parallel distributed information access algorithm based on the node scale of the active directed graph. Here, subgraphs are seen as the physical realization of the information stored in the active directed graph. Unlike traditional algorithms with global perspectives, our algorithm emphasizes global node collaboration in resource utilization through local perspectives. While it may not achieve the global optimum like a global-view algorithm, it offers superior robustness, concurrency, decentralization, and biological feasibility. We also tested network capacity, fault tolerance, and robustness, finding that the algorithm performs better in sparser network structures.

FDT‐Net: Frequency‐Aware Dual‐Branch Transformer‐Based Optic Cup and Optic Disk Segmentation With Parallel Contour Information Mining and Uncertainty‐Guided Refinement

ABSTRACTAccurate segmentation of the optic cup and disc in fundus images is crucial for the prevention and diagnosis of glaucoma. However, challenges arise due to factors such as blood vessels, and mainstream networks often demonstrate limited capacity in extracting contour information. In this paper, we propose a segmentation framework named FDT‐Net, which is based on a frequency‐aware dual‐branch Transformer (FDBT) architecture with parallel contour information mining and uncertainty‐guided refinement. Specifically, we design a FDBT that operates in the frequency domain. This module leverages the inherent contextual awareness of Transformers and utilizes Discrete Cosine Transform (DCT) transformation to mitigate the impact of certain interference factors on segmentation. The FDBT comprises global and local branches that independently extract global and local information, thereby enhancing segmentation results. Moreover, to further mine additional contour information, this study develops the parallel contour information mining (PCIM) module to operate in parallel. These modules effectively capture more details from the edges of the optic cup and disc while avoiding mutual interference, thus optimizing segmentation performance in contour regions. Furthermore, we propose an uncertainty‐guided refinement (UGR) module, which generates and quantifies uncertainty mass and leverages it to enhance model performance based on subjective logic theory. The experimental results on two publicly available datasets demonstrate the superior performance and competitive advantages of our proposed FDT‐Net. The code for this project is available at https://github.com/Rookie49144/FDT‐Net.

Brain-Inspired Architecture for Spiking Neural Networks.

Spiking neural networks (SNNs), using action potentials (spikes) to represent and transmit information, are more biologically plausible than traditional artificial neural networks. However, most of the existing SNNs require a separate preprocessing step to convert the real-valued input into spikes that are then input to the network for processing. The dissected spike-coding process may result in information loss, leading to degenerated performance. However, the biological neuron system does not perform a separate preprocessing step. Moreover, the nervous system may not have a single pathway with which to respond and process external stimuli but allows multiple circuits to perceive the same stimulus. Inspired by these advantageous aspects of the biological neural system, we propose a self-adaptive encoding spike neural network with parallel architecture. The proposed network integrates the input-encoding process into the spiking neural network architecture via convolutional operations such that the network can accept the real-valued input and automatically transform it into spikes for further processing. Meanwhile, the proposed network contains two identical parallel branches, inspired by the biological nervous system that processes information in both serial and parallel. The experimental results on multiple image classification tasks reveal that the proposed network can obtain competitive performance, suggesting the effectiveness of the proposed architecture.

IPCT-Net: Parallel information bottleneck modality fusion network for obstructive sleep apnea diagnosis

Obstructive sleep apnea (OSA) is a common sleep breathing disorder and timely diagnosis helps to avoid the serious medical expenses caused by related complications. Existing deep learning (DL)-based methods primarily focus on single-modal models, which cannot fully mine task-related representations. This paper develops a modality fusion representation enhancement (MFRE) framework adaptable to flexible modality fusion types with the objective of improving OSA diagnostic performance, and providing quantitative evidence for clinical diagnostic modality selection. The proposed parallel information bottleneck modality fusion network (IPCT-Net) can extract local-global multi-view representations and eliminate redundant information in modality fusion representations through branch sharing mechanisms. We utilize large-scale real-world home sleep apnea test (HSAT) multimodal data to comprehensively evaluate relevant modality fusion types. Extensive experiments demonstrate that the proposed method significantly outperforms existing methods in terms of participant numbers and OSA diagnostic performance. The proposed MFRE framework delves into modality fusion in OSA diagnosis and contributes to enhancing the screening performance of artificial intelligence (AI)-assisted diagnosis for OSA.

MicroLEDs for optical neuromorphic computing—application potential and present challenges

Abstract The slow non-radiative surface recombination velocity of gallium nitride (GaN) in combination with its highly efficient radiative recombination makes this material ideally suited for microLEDs with dimensions as small as 1 µm and even below, serving as the fundamental building block of micro-displays. However, due to their superior miniaturization potential and energy efficiency, GaN-based microLEDs have applications that extend well beyond display technology. Their capability to produce optical patterns with high resolution, which can be modulated at extremely high frequencies, makes them suitable for numerous other applications. We suggest exploiting these exciting properties for a new and potentially equally significant application: utilizing microLEDs in optical processing units for artificial intelligence workloads. In neuromorphic computing, relevant aspects of biological neural networks are emulated directly with either electronic circuits or photonic devices, avoiding the shortcomings of conventional digital computer technology for AI workloads, which generally require massively parallel information processing. GaN microLEDs are discussed here as a promising enabling technology for optical neuromorphic processing units. We see great potential to substantially decrease power consumption through massively parallel in-memory processing combined with efficient photon production and detection. A theoretical analysis of scalability and energy efficiency is provided. A macroscopic bench-top optical microLED demonstrator is presented, which experimentally proves the feasibility of our approach. Future potential and challenges associated with miniaturizing and scaling microLED-based optical processing units are discussed. Finally, we summarize the open research questions that require attention before fully functional and miniaturized optical neuromorphic processing units based on GaN microLEDs can be realized.

Exploring multi-qubit analogue operations through acoustic wave dynamics

Quantum computing harnesses quantum phenomena like superposition and entanglement to surpass classical computers, showing promise across various fields. Current techniques utilize qubits in quantum circuits for parallel information processing, though managing and measuring these systems poses significant challenges. We propose a novel approach to simplify executing complex quantum algorithms without relying on multiple quantum gate operations. Our method introduces logical phi-bits—classical counterparts to qubits, using nonlinear acoustic waves in an externally driven acoustic metastructure. We demonstrate that complex multi-phi-bit unitary operations, akin to those in quantum circuits, can be conducted through a single action on this metastructure. This method starkly contrasts traditional quantum computing, which requires decomposed sequences of qubit gates for equivalent operations. The phi-bit system simplifies processes that are typically complex in quantum mechanics, potentially enhancing robustness and ease of implementation. Our results indicate that phi-bits could expand computational models by merging classical wave dynamics with quantum computational principles, thereby widening the potential of computational technologies. This research advances quantum-analogue computation and introduces new prospects for utilizing wave physics in information processing, thereby challenging and expanding existing paradigms in both classical and quantum computing. [Funding: NSF grant 2204382, 2204400, and 2242925.]

Multilayer spatial-spectral segmentation network for detecting AFB1

Aflatoxin, a toxin produced by Aspergillus flavus, is commonly found in peanuts and poses a significant threat to human health. In this paper, we propose an adjustable Mul-U-Net model, capable of learning both spectral and spatial information in parallel, facilitating rapid detection of aflatoxins by adjusting the number of central spatial U units. First, AFB1 was semantically segmented by inputting hyperspectral image cubes into six semantic segmentation models: Mul-U-Net, U-Net, FCN, SegNet, CNN, and RefineNet. Experimental results demonstrate that Mul-U-Net outperformed the other models, achieving an overall accuracy of 95.64 % and an RMSE of 0.2079. We further examined the influence of different spatial U units on model performance and found that the optimal spatial U count matched the number of labels. In the aflatoxin dataset, the Mul-2 U-Net model achieved the highest accuracy of 96.12 %. Further validation on the Indian Pines and Pavia University datasets showed that Mul-16 U-Net and Mul-9 U-Net achieved the highest accuracies of 98.12 % and 98.01 %, respectively. This study presents a robust, high-performance, and tunable Mul-U-Net semantic segmentation network, offering a valuable new approach for the identification of aflatoxins in peanuts.

Parallel multi-scale dynamic graph neural network for multivariate time series forecasting

Accurately modeling and predicting multivariate time series (MTS) are crucial in real-world scenarios, where precise predictions aid decision-making. MTS exhibit complex variable relationships and dynamic evolution patterns across various time scales, posing significant modeling challenges. To address these, a parallel multi-scale dynamic graph neural network (PMEDGN) is proposed for parallel modeling of dynamic information across multiple time scales. Spatial-temporal embedding module based on spatial-temporal attention mechanisms and multi-scale dynamic graph generation module are designed. These modules capture implicit spatial-temporal dependencies and self-learn to generate multi-scale dynamic graph structures from MTS without predefined graphs, automatically obtaining the importance of different time scale patterns. Additionally, a global graph convolution module is developed to integrate parallel multi-scale information, enhancing collaboration across time scales for final predictions. Comprehensive experiments on real-world datasets demonstrate the effectiveness and superiority of PMEDGN over state-of-the-art methods, underscoring its potential for practical applications.

Language networks of normal-hearing infants exhibit topological differences between resting and steady states: An fNIRS functional connectivity study.

Task-related studies have consistently reported that listening to speech sounds activate the temporal and prefrontal regions of the brain. However, it is not well understood how functional organization of auditory and language networks differ when processing speech sounds from its resting state form. The knowledge of language network organization in typically developing infants could serve as an important biomarker to understand network-level disruptions expected in infants with hearing impairment. We hypothesized that topological differences of language networks can be characterized using functional connectivity measures in two experimental conditions (1) complete silence (resting) and (2) in response to repetitive continuous speech sounds (steady). Thirty normal-hearing infants (14 males and 16 females, age: 7.8 ± 4.8 months) were recruited in this study. Brain activity was recorded from bilateral temporal and prefrontal regions associated with speech and language processing for two experimental conditions: resting and steady states. Topological differences of functional language networks were characterized using graph theoretical analysis. The normalized global efficiency and clustering coefficient were used as measures of functional integration and segregation, respectively. We found that overall, language networks of infants demonstrate the economic small-world organization in both resting and steady states. Moreover, language networks exhibited significantly higher functional integration and significantly lower functional segregation in resting state compared to steady state. A secondary analysis that investigated developmental effects of infants aged 6-months or below and above 6-months revealed that such topological differences in functional integration and segregation across resting and steady states can be reliably detected after the first 6-months of life. The higher functional integration observed in resting state suggests that language networks of infants facilitate more efficient parallel information processing across distributed language regions in the absence of speech stimuli. Moreover, higher functional segregation in steady state indicates that the speech information processing occurs within densely interconnected specialized regions in the language network.

Open-Source Data Formalization through Model-Based Systems Engineering for Concurrent Preliminary Design of CubeSats

Market trends in the space sector suggest a notable increase in satellite operations and market value for the coming decade. In parallel, there has been a shift in the industrial and academic sectors from traditional Document-Based System Engineering to Model-based systems engineering (MBSE) combined with Concurrent engineering (CE) practices. Due to growing demands, the drivers behind this change have been the need for quicker and more cost-effective design processes. A key challenge in this transition remains to determine how to effectively formalize and exchange data during all design stages and across all discipline-specific tools; as representing systems through models can be a complex endeavor. For instance, during the Preliminary design (PD) phase, the integration of system models with external mathematical models for simulations, analyses, and system budgeting is crucial. The introduction of CubeSats and their standard has partly addressed the question of standardization and has aided in reducing overall development time and costs in the space sector. Nevertheless, questions about how to successfully exchange data endure. This paper focuses on formalizing a CubeSat model for use across various stages of the PD phase. The entire process is conducted with the exclusive use of open-source tools, to facilitate the transparency of data integration across the PD phases, and the overall life cycle of a CubeSat. The paper has two primary outcomes: (i) developing a generic CubeSat model using Systems modeling language (SysML) that includes data storage and visualization through the application of Unified modeling language (UML) stereotypes, streamlining in parallel information exchange for integration with various simulation and analysis tools; (ii) creating an end-to-end use case scenario within the Nanostar software suite (NSS), an open-source framework designed to streamline data exchange across different software during CE sessions. A case study from a theoretical academic space mission concept is presented as the illustration of how to utilize the proposed formalization, and it serves as well as a preliminary validation of the proposed formalization. The proposed formalization positions the CubeSat SysML model as the central data source throughout the design process. It also supports automated trade-off analyses by combining the benefits of SysML with effective data instantiating across all PD study phases.

Visual sequence encoding is modulated by music schematic structure and familiarity.

Music is omnipresent in daily life and may interact with critical cognitive processes including memory. Despite music's presence during diverse daily activities including studying, commuting, or working, existing literature has yielded mixed results as to whether music improves or impairs memory for information experienced in parallel. To elucidate how music memory and its predictive structure modulate the encoding of novel information, we developed a cross-modal sequence learning task during which participants acquired sequences of abstract shapes accompanied with paired music. Our goal was to investigate whether familiar and structurally regular music could provide a "temporal schema" (rooted in the organized and hierarchical structure of music) to enhance the acquisition of parallel temporally-ordered visual information. Results revealed a complex interplay between music familiarity and music structural regularity in learning paired visual sequences. Notably, compared to a control condition, listening to well-learned, regularly-structured music (music with high predictability) significantly facilitated visual sequence encoding, yielding quicker learning and retrieval speed. Conversely, learned but irregular music (where music memory violated musical syntax) significantly impaired sequence encoding. While those findings supported our mechanistic framework, intriguingly, unlearned irregular music-characterized by the lowest predictability-also demonstrated memory enhancement. In conclusion, this study demonstrates that concurrent music can modulate visual sequence learning, and the effect varies depending on the interaction between both music familiarity and regularity, offering insights into potential applications for enhancing human memory.

Bioinspired iontronic synapse fibers for ultralow-power multiplexing neuromorphic sensorimotor textiles

Artificial neuromorphic devices can emulate dendric integration, axonal parallel transmission, along with superior energy efficiency in facilitating efficient information processing, offering enormous potential for wearable electronics. However, integrating such circuits into textiles to achieve biomimetic information perception, processing, and control motion feedback remains a formidable challenge. Here, we engineer a quasi-solid-state iontronic synapse fiber (ISF) comprising photoresponsive TiO2, ion storage Co-MoS2, and an ion transport layer. The resulting ISF achieves inherent short-term synaptic plasticity, femtojoule-range energy consumption, and the ability to transduce chemical/optical signals. Multiple ISFs are interwoven into a synthetic neural fabric, allowing the simultaneous propagation of distinct optical signals for transmitting parallel information. Importantly, IFSs with multiple input electrodes exhibit spatiotemporal information integration. As a proof of concept, a textile-based multiplexing neuromorphic sensorimotor system is constructed to connect synaptic fibers with artificial fiber muscles, enabling preneuronal sensing information integration, parallel transmission, and postneuronal information output to control the coordinated motor of fiber muscles. The proposed fiber system holds enormous promise in wearable electronics, soft robotics, and biomedical engineering.

Guest Editorial: Quantum industry: Applications in quantum communication (Quantum.Tech Europe 2022)

AbstractQuantum technology harnesses the principles of quantum mechanics to accomplish tasks in a different way as compared to the classical technologies. This includes quantum computing, which uses qubits for parallel information processing, greatly enhancing computation speed and entanglement and empowering problem‐solving abilities. Quantum communication provides secure data transmission through quantum cryptography, while quantum sensing offers improved measurement precision, benefiting areas such as cryptography, material science, and pharmaceuticals. Additionally, the implementation and commercialisation of quantum technology involve transitioning theoretical quantum concepts into practical applications and marketable products. To achieve widespread adoption of quantum industry, significant research efforts are crucial among academia, industry, and government.

A Hybrid Deep Learning Framework Using Scaling‐Basis Chirplet Transform for Motor Imagery EEG Recognition in Brain–Computer Interface Applications

ABSTRACTThe emerging field of brain–computer interface has significantly facilitated the analysis of electroencephalogram signals required for motor imagery classification tasks. However, the accuracy of EEG classification models has been restricted by the low signal‐to‐noise ratio, nonlinear nature of brain signals, and a lack of sufficient EEG data for training. To address these challenges, this study proposes a new approach that combines time‐frequency analysis with a hybrid parallel–series attention‐based deep learning network for EEG signal classification. The proposed framework comprises three main elements: first, a scaling‐basis chirplet transform designed to effectively capture the characteristics of nonstationary EEG signals; second, a hybrid parallel–series attention‐based deep learning network to extract features. The serial information flow continuously expands the receptive fields of output neurons, whereas parallel information flow extracts features based on different regions. Finally, machine learning classifiers are utilized to predict the corresponding motor imagery state. The developed EEG‐based motor imagery classification framework is assessed by two open‐source datasets, BCI competition III, dataset IIIa and BCI competition IV, dataset IIa and has achieved the average classification accuracy of 95.55% on BCI competition III, dataset IIIa and 90.18% on BCI competition IV, dataset IIa. The experimental findings illustrate that this study has attained promising motor imagery discrimination performance, surpassing existing techniques in terms of classification accuracy and kappa coefficient.