How critical is brain criticality?

Trends in Neurosciences, Journal Year: 2022, Volume and Issue: 45(11), P. 820 - 837

Published: Sept. 9, 2022

Language: Английский

---

Null models in network neuroscience

Nature reviews. Neuroscience, Journal Year: 2022, Volume and Issue: 23(8), P. 493 - 504

Published: May 31, 2022

Language: Английский

---

Connectome-based modelling of neurodegenerative diseases: towards precision medicine and mechanistic insight

Nature reviews. Neuroscience, Journal Year: 2023, Volume and Issue: 24(10), P. 620 - 639

Published: Aug. 24, 2023

Language: Английский

---

Neuromorphic learning, working memory, and metaplasticity in nanowire networks

Science Advances, Journal Year: 2023, Volume and Issue: 9(16)

Published: April 21, 2023

Nanowire networks (NWNs) mimic the brain's neurosynaptic connectivity and emergent dynamics. Consequently, NWNs may also emulate synaptic processes that enable higher-order cognitive functions such as learning memory. A quintessential task used to measure human working memory is n-back task. In this study, variations inspired by are implemented in a NWN device, external feedback applied brain-like supervised reinforcement learning. found retain information at least n = 7 steps back, remarkably similar originally proposed "seven plus or minus two" rule for subjects. Simulations elucidate how synapse-like junction plasticity depends on previous modifications, analogous "synaptic metaplasticity" brain, consolidated via strengthening pruning of conductance pathways.

Language: Английский

---

Physical reservoir computing with emerging electronics

Nature Electronics, Journal Year: 2024, Volume and Issue: 7(3), P. 193 - 206

Published: March 12, 2024

Language: Английский

---

Prefrontal connectomics: from anatomy to human imaging

Neuropsychopharmacology, Journal Year: 2021, Volume and Issue: 47(1), P. 20 - 40

Published: Sept. 28, 2021

Abstract The fundamental importance of prefrontal cortical connectivity to information processing and, therefore, disorders cognition, emotion, and behavior has been recognized for decades. Anatomic tracing studies in animals have formed the basis delineating direct monosynaptic connectivity, from cells origin, through axon trajectories, synaptic terminals. Advances neuroimaging combined with network science taken lead developing complex wiring diagrams or connectomes human brain. A key question is how well these magnetic resonance imaging (MRI)-derived networks hubs reflect anatomic “hard wiring” first proposed underlie distribution large-scale interactions. In this review, we address challenge by focusing on what known about connections non-human primates compares MRI-derived measurements organization humans. First, outline pathways each cortex (PFC) region. We then review available MRI-based techniques indirectly measuring structural functional introduce graph theoretical methods analysis hubs, modules, topologically integrative features connectome. Finally, bring two approaches together, using specific examples, demonstrate connections, demonstrated tract-tracing studies, can directly inform understanding composition PFC nodes edges that connect subcortical areas.

Language: Английский

---

Brain connectivity meets reservoir computing

PLoS Computational Biology, Journal Year: 2022, Volume and Issue: 18(11), P. e1010639 - e1010639

Published: Nov. 16, 2022

The connectivity of Artificial Neural Networks (ANNs) is different from the one observed in Biological (BNNs). Can wiring actual brains help improve ANNs architectures? we learn about what network features support computation brain when solving a task? At meso/macro-scale level connectivity, ANNs’ architectures are carefully engineered and such those design decisions have crucial importance many recent performance improvements. On other hand, BNNs exhibit complex emergent patterns at all scales. individual level, results development plasticity processes, while species adaptive reconfigurations during evolution also play major role shaping connectivity. Ubiquitous been identified years, but their brain’s ability to perform concrete computations remains poorly understood. Computational neuroscience studies reveal influence specific only on abstract dynamical properties, although implications real networks topologies machine learning or cognitive tasks barely explored. Here present cross-species study with hybrid approach integrating connectomes Bio-Echo State Networks, which use solve memory tasks, allowing us probe potential computational task solving. We find consistent across showing that biologically inspired as well classical echo state networks, provided minimum randomness diversity connections allowed. framework, bio2art , map scale up can be integrated into recurrent ANNs. This allows show interareal patterns, stressing stochastic processes determining neural general.

Language: Английский

---

Scale-free behavioral dynamics directly linked with scale-free cortical dynamics

eLife, Journal Year: 2023, Volume and Issue: 12

Published: Jan. 27, 2023

Naturally occurring body movements and collective neural activity both exhibit complex dynamics, often with scale-free, fractal spatiotemporal structure. Scale-free dynamics of brain behavior are important because each is associated functional benefits to the organism. Despite their similarities, scale-free have been studied separately, without a unified explanation. Here, we show that mouse neurons in visual cortex strongly related. Surprisingly, limited specific subsets neurons, these stochastic winner-take-all competition other subsets. This observation inconsistent prevailing theories systems, which stem from criticality hypothesis. We develop computational model incorporates known cell-type-specific circuit structure, explaining our findings new type critical dynamics. Our results establish underpinnings clear behavioral relevance activity.As go about days, how do fidget, compared frequently make larger movements, like walking down hall? And rare trek across town same walk Animals tend follow mathematical law relates size them. posits small-to-medium large-to-huge related way, is, ‘scale-free’, it holds for different scales movement. measurements also this law: level activation group they activated way levels activation. Although behave mathematically similar two facts had not previously linked. Jones et al. mice, found were linked activity, but only certain neurons. hidden compete When turn on, competing groups off. averaged together, fluctuations cancel out. The provide understanding orchestrated healthy organisms. In particular, suggest complex, multi-scale nature may emerge operating at tipping point between order disorder, edge chaos.

Language: Английский

---

Biological neurons act as generalization filters in reservoir computing

Proceedings of the National Academy of Sciences, Journal Year: 2023, Volume and Issue: 120(25)

Published: June 12, 2023

Reservoir computing is a machine learning paradigm that transforms the transient dynamics of high-dimensional nonlinear systems for processing time-series data. Although was initially proposed to model information in mammalian cortex, it remains unclear how nonrandom network architecture, such as modular cortex integrates with biophysics living neurons characterize function biological neuronal networks (BNNs). Here, we used optogenetics and calcium imaging record multicellular responses cultured BNNs employed reservoir framework decode their computational capabilities. Micropatterned substrates were embed architecture BNNs. We first show response static inputs can be classified linear decoder modularity positively correlates classification accuracy. then timer task verify possess short-term memory several 100 ms finally this property exploited spoken digit classification. Interestingly, BNN-based reservoirs allow categorical learning, wherein trained on one dataset classify separate datasets same category. Such not possible when directly decoded by decoder, suggesting act generalization filter improve performance. Our findings pave way toward mechanistic understanding representation within build future expectations realization physical based

Language: Английский

---

Unbiased construction of constitutive relations for soft materials from experiments via rheology-informed neural networks

Proceedings of the National Academy of Sciences, Journal Year: 2024, Volume and Issue: 121(2)

Published: Jan. 3, 2024

The ability to concisely describe the dynamical behavior of soft materials through closed-form constitutive relations holds key accelerated and informed design processes. conventional approach is construct simplifying assumptions approximating time- rate-dependent stress response a complex fluid an imposed deformation. While traditional frameworks have been foundational our current understanding materials, they often face twofold existential limitation: i) Constructed on ideal generalized assumptions, precise recovery material-specific details usually serendipitous, if possible, ii) inherent biases that are involved by making those commonly come at cost new physical insight. This work introduces leveraging recent advances in scientific machine learning methodologies discover governing equation from experimental data for fluids. Our rheology-informed neural network framework found capable hidden rheology limited number experiments. followed construction unbiased relation accurately describes wide range bulk material. extremely efficient model discovery real-world system, also provides insight into underpinning physics

Language: Английский

---