Brain connectivity meets reservoir computing.
Brain connectivity meets reservoir computing.
复制标题
DOI：
10.1371/journal.pcbi.1010639
复制
发表时间：
2022-11
影响因子：
4.3
通讯作者：
中科院分区：
文献类型：
--
作者：
人工神经网络 (ANN) 的连接性与生物神经网络 (BNN) 中观察到的连接性不同。实际大脑的连接可以帮助改善人工神经网络的架构吗？我们能否从人工神经网络中了解到在解决任务时哪些网络特征支持大脑中的计算？在连接的中观/宏观层面上，人工神经网络的架构经过精心设计，这些设计决策对于最近的许多性能改进至关重要。另一方面，BNN 在所有尺度上都表现出复杂的突发连接模式。在个体层面，BNN 的连通性是大脑发育和可塑性过程的结果，而在物种层面，进化过程中的适应性重新配置也在塑造连通性方面发挥着重要作用。近年来，人们已经发现了大脑连接的普遍特征，但它们在大脑执行具体计算的能力中所扮演的角色仍然知之甚少。计算神经科学研究揭示了特定大脑连接特征仅对抽象动力学特性的影响，尽管真正的大脑网络拓扑对机器学习或认知任务的影响几乎没有被探索过。在这里，我们提出了一项跨物种研究，采用集成真实大脑连接组和生物回声状态网络的混合方法，我们用它来解决具体的记忆任务，使我们能够探索真实大脑连接模式对任务解决的潜在计算影响。我们发现跨物种和任务的结果是一致的，这表明，只要允许最低水平的随机性和连接多样性，受生物学启发的网络的性能与经典的回声状态网络一样。我们还提出了一个框架，bio2art，来绘制和扩展可以集成到循环人工神经网络中的真实连接体。这种方法还使我们能够展示区域间连接模式多样性的至关重要性，强调随机过程决定神经网络连接性的重要性。人工神经网络 (ANN) 和生物神经网络 (BNN) 表现出不同的连接模式。人工神经网络通常拥有精心设计的架构，这些架构在其性能中发挥着重要作用。另一方面，BNN 的布线显示出由发育和神经元可塑性等过程产生的自组织涌现模式。尽管大脑连接的普遍特性已经被识别出来，并且与大脑的抽象动态特性相关联，但真正的大脑网络拓扑对具体机器学习任务的影响却几乎没有被探索过。这项混合、跨物种研究的目标是通过在具体的机器学习任务上探测真实的大脑连接组，朝这个方向迈出一步。我们的方法集成了真实的大脑连接体和生物回声状态网络，我们用它来解决具体的记忆任务。为了实现这一目标，我们还在这里提出了一个框架，bio2art，来映射和扩展可以集成到循环人工神经网络中的真实连接组。我们发现跨物种和任务的结果是一致的，这表明，只要允许最低水平的随机性和连接多样性，受生物学启发的网络的性能与经典的回声状态网络一样。我们的研究结果强调了神经网络连接中随机性的重要性，特别是关于区域间连接的异质性。
The connectivity of Artificial Neural Networks (ANNs) is different from the one observed in Biological Neural Networks (BNNs). Can the wiring of actual brains help improve ANNs architectures? Can we learn from ANNs about what network features support computation in the brain when solving a task? At a meso/macro-scale level of the connectivity, ANNs’ architectures are carefully engineered and such those design decisions have crucial importance in many recent performance improvements. On the other hand, BNNs exhibit complex emergent connectivity patterns at all scales. At the individual level, BNNs connectivity results from brain development and plasticity processes, while at the species level, adaptive reconfigurations during evolution also play a major role shaping connectivity. Ubiquitous features of brain connectivity have been identified in recent years, but their role in the brain’s ability to perform concrete computations remains poorly understood. Computational neuroscience studies reveal the influence of specific brain connectivity features only on abstract dynamical properties, although the implications of real brain networks topologies on machine learning or cognitive tasks have been barely explored. Here we present a cross-species study with a hybrid approach integrating real brain connectomes and Bio-Echo State Networks, which we use to solve concrete memory tasks, allowing us to probe the potential computational implications of real brain connectivity patterns on task solving. We find results consistent across species and tasks, showing that biologically inspired networks perform as well as classical echo state networks, provided a minimum level of randomness and diversity of connections is allowed. We also present a framework, bio2art, to map and scale up real connectomes that can be integrated into recurrent ANNs. This approach also allows us to show the crucial importance of the diversity of interareal connectivity patterns, stressing the importance of stochastic processes determining neural networks connectivity in general. Artificial Neural Networks (ANNs) and Biological Neural Networks (BNNs) exhibit different connectivity patterns. ANNs’ have tyically carefully hand-crafted architectures that play an important role in their performance. On the other hand, BNNs’ wiring shows self-organized emergent patterns resulting from processes such as development and neuronal plasticity. Although ubiquitous properties of brain connectivity have beed identified and associated with abstract dynamical properties of the brain, the implications of real brain networks topologies on concrete machine learning tasks have been barely explored. The goal of this hybrid, cross-species study was to give a step in that direction by probing real brain connectomes on concrete machine learning tasks. Our approach integrates real brain connectomes and Bio-Echo State Networks, which we use to solve concrete memory tasks. To achieve that, we also present here a framework, bio2art, to map and scale up real connectomes that can be integrated into recurrent ANNs. We find results consistent across species and tasks, showing that biologically inspired networks perform as well as classical echo state networks, provided a minimum level of randomness and diversity of connections is allowed. Out findings stress the importance of stochasticity in neural networks connectivity, especially regarding the heterogeneity of interareal connectivity.
登录
查看更多内容
影响因子：
5.7
作者：
Mišić B;Sporns O
通讯作者：
Sporns O
影响因子：
4.3
作者：
Ellefsen KO;Mouret JB;Clune J
通讯作者：
Clune J
DOI：
10.1093/cercor/bhs270
发表时间：
2014-01
期刊：
Cerebral cortex (New York, N.Y. : 1991)
影响因子：
--
作者：
Markov NT;Ercsey-Ravasz MM;Ribeiro Gomes AR;Lamy C;Magrou L;Vezoli J;Misery P;Falchier A;Quilodran R;Gariel MA;Sallet J;Gamanut R;Huissoud C;Clavagnier S;Giroud P;Sappey-Marinier D;Barone P;Dehay C;Toroczkai Z;Knoblauch K;Van Essen DC;Kennedy H
通讯作者：
Kennedy H
影响因子：
12.9
作者：
Lukosevicius, Mantas;Jaeger, Herbert
通讯作者：
Jaeger, Herbert
影响因子：
13.6
作者：
Goulas, A.;Betzel, R. F.;Hilgetag, C. C.
通讯作者：
Hilgetag, C. C.