Bain on Neural Networks

In his bookMind and body(1873), Bain set out an account in which he related the processes of associative memory to the distribution of activity in neural groupings—or neural networks as they are now termed. In the course of this account, Bain anticipated certain aspects of connectionist ideas that are normally attributed to 20th-century authors—most notably Hebb (1949). In this paper we reproduce Bain's arguments relating neural activity to the workings of associative memory which include an early version of the principles enshrined in Hebb's neurophysiological postulate. Nonetheless, despite their prescience, these specific contributions to the connectionist case have been almost entirely ignored. Eventually, Bain came to doubt the practicality of his own arguments and, in so doing, he seems to have ensured that his ideas concerning neural groupings exerted little or no influence on the subsequent course of theorizing in this area.     

Neural Darwinism and consciousness

Neural Darwinism (ND) is a large scale selectionist theory of brain development and function that has been hypothesized to relate to consciousness. According to ND, consciousness is entailed by reentrant interactions among neuronal populations in the thalamocortical system (the 'dynamic core'). These interactions, which permit high-order discriminations among possible core states, confer selective advantages on organisms possessing them by linking current perceptual events to a past history of value-dependent learning. Here, we assess the consistency of ND with 16 widely recognized properties of consciousness, both physiological (for example, consciousness is associated with widespread, relatively fast, low amplitude interactions in the thalamocortical system), and phenomenal (for example, consciousness involves the existence of a private flow of events available only to the experiencing subject). While no theory accounts fully for all of these properties at present, we find that ND and its recent extensions fare well.     

审美的神经机制

审美能力是人类独有的高级功能。本文介绍了审美的认知神经模型, 以及相关研究的实验范式, 对各种刺激材料和任务进行了评价。回顾了近年来脑成像研究的主要进展, ERP和MEG研究证明审美包含发生在不同时段的不同加工阶段, 与审美判断相关的ERP成分为300~400 ms 额中部负成分和440~880 ms 头皮后部晚期正成分。fMRI研究发现与审美相关的主要脑区包括尾状核、框额部皮质和扣带前回。未来研究应进一步区分审美的不同方面的神经机制之间的差别, 扩大刺激材料范围, 更深入地探讨复杂因素, 如文化因素对审美神经机制的影响。     

Neural mechanisms of timing

A crucial step in timing research is to isolate clock components from other sources of temporal variability. Significant progress has been made both behaviorally and neurologically, using elaborate experimental designs that separate timing mechanisms from motoric sensory and mnemonic processes. Marked similarities between the temporal characteristics of the clock in perception and production tasks implicate a common timing system. Similar conclusions can be reached from clinical studies, indeed individuals with neocerebellar damage are impaired at discriminating and reproducing short intervals. However, other patient populations, especially those with disorders affecting the basal ganglia, also exhibit deficits in timing tasks. It therefore appears that temporal computation may be distributed throughout the brain with specific roles for different neural structures.     

Estrogen and Neural Plasticity

Converging clinical evidence suggests that postmenopausal estrogen therapy in women is associated with improved cognition and a reduced incidence of Alzheimer's disease. In experimental work, investigators have found estrogen to promote changes in synaptic plasticity within the nervous system. In this article, we review both the clinical and the experimental literature, and consider mechanisms of action of estrogen on neurons and synaptic plasticity, and how they might protect against the cognitive impairments of old age.     

Neural Policy Style Transfer

Style Transfer has been proposed in a number of fields: fine arts, natural language processing, and fixed trajectories. We scale this concept up to control policies within a Deep Reinforcement Learning infrastructure. Each network is trained to maximize the expected reward, which typically encodes the goal of an action, and can be described as the content. The expressive power of deep neural networks enables encoding a secondary task, which can be described as the style. The Neural Policy Style Transfer (NPST)¹ algorithm is proposed to transfer the style of one policy to another, while maintaining the content of the latter. Different policies are defined via Deep Q-Network architectures. These models are trained using demonstrations through Inverse Reinforcement Learning. Two different sets of user demonstrations are performed, one for content and other for style. Different styles are encoded as defined by user demonstrations. The generated policy is the result of feeding a content policy and a style policy to the NPST algorithm. Experiments are performed in a catch-ball game inspired by the Deep Reinforcement Learning classical Atari games; and a real-world painting scenario with a full-sized humanoid robot, based on previous works of the authors. The implementation of three different Q-Network architectures (Shallow, Deep and Deep Recurrent Q-Network) to encode the policies within the NPST framework is proposed and the results obtained in the experiments with each of these architectures compared.     

Neural Mechanisms in Dyslexia

ABSTRACT— Within the last two decades, evidence from many laboratories has converged to indicate the cognitive basis for dyslexia: Dyslexia is a disorder within the language system and, more specifically, within a particular subcomponent of that system, phonological processing. Converging evidence from a number of laboratories using functional brain imaging indicates that there is a disruption of left-hemisphere posterior neural systems in child and adult dyslexic readers when they perform reading tasks. The discovery of a disruption in the neural systems serving reading has significant implications for the acceptance of dyslexia as a valid disorder—a necessary condition for its identification and treatment. Brain-imaging findings provide, for the first time, convincing, irrefutable evidence that what has been considered a hidden disability is "real," and these findings have practical implications for the provision of accommodations, a critical component of management for older children and young adults attending postsecondary and graduate programs. The utilization of advances in neuroscience to inform educational policy and practices provides an exciting example of translational science being used for the public good.     

共情的神经机制

共情是社会认知中不可或缺的一部分,是维系积极社会关系的社会动机因素。近年来,脑成像技术的发展为共情的神经机制研究开拓了新的视野,共情的神经机制研究取得了很大的进展。共情神经机制的研究主要包括共情的"表征共享"基础、共情的镜像神经元特征、共情调节的早期、晚期评价模型、共情的调节因素以及特殊个体的共情研究。在此基础上指出现有研究中存在的问题和未来研究的方向。     

自我面孔识别的脑机制

自我面孔识别是自我参照加工的一种研究范式, 反映了人们通过自我与他人的区分识别出自我面孔的过程。自我面孔识别的脑区定位涉及前额叶、脑岛、扣带回、颞叶和顶叶等脑区的协同作用, 其认知加工有三个阶段：低水平的感觉处理阶段, 对自我参照的面孔信息的处理阶段和身份辨别阶段。在今后的研究中, 应该在时程上区分自我面孔识别的各个加工过程, 并且结合脑区定位结果, 明确自我面孔识别的各加工阶段及认知成分。     

Neural correlates of consciousness reconsidered

It is widely accepted among philosophers that neuroscientists are conducting a search for the neural correlates of consciousness, or NCC. Chalmers (2000) conceptualized this research program as the attempt to correlate the contents of conscious experience with the contents of representations in specific neural populations. A notable claim on behalf of this interpretation is that the neutral language of "correlates" frees us from philosophical disputes over the mind/body relation, allowing the science to move independently. But the experimental paradigms and explanatory canons of neuroscience are not neutral about the mechanical relation between consciousness and the brain. I argue that NCC research is best characterized as an attempt to locate a causally relevant neural mechanism and not as an effort to identify a discrete neural representation, the content of which correlates with some actual experience. It might be said that the first C in "NCC" should stand for "causes" rather than "correlates."     

Neural correlates of the first-person-perspective

Human self-consciousness depends on the metarepresentation of mental and bodily states as one's own mental and bodily states. First-person-perspective taking is not sufficient, but necessary for human self-consciousness. To assign a first-person-perspective is to center one's own multimodal experiential space upon one's own body, thus operating in an egocentric reference frame. The brain regions involved in assigning first-person-perspective comprise medial prefrontal, medial parietal and lateral temporoparietal cortex. These empirical findings complement recent neurobiologically oriented theories of self-consciousness which focus on the relation between the subject and his/her environment by supplying a neural basis for its key components.     

创造性思维的脑机制

创造性思维是创造性的核心。近年来,脑电和脑成像技术的发展为研究创造性思维的神经基础提供了有力的技术支持,对创造性思维的脑机制研究取得了较大进展。创造性思维的脑机制研究主要包括顿悟的脑机制、发散性思维的脑机制、远距离联想的脑机制、言语创造性和图画创造性的对比的脑机制研究。研究结果显示创造性思维需要多个脑区的参与,因不同的认知任务其关键脑区而有所不同。对创造性思维的脑机制研究进行了总结,对这些研究可能存在创造性思维究竟应该定义为“领域一般的”还是“领域特殊的”的解释进行了讨论,并指出了研究中可能存在的基线任务设置问题和未来研究中需要在研究手段、研究设计、研究领域等方面加以改进的建议     

抑制引起的遗忘及其神经机制

当面对不愉快的提醒时, 人们通常会试图阻止那些不想要的记忆出现在脑海中。先前的研究表明, 抑制不想要的记忆的检索会降低这些记忆的可及性, 引起遗忘, 这一效应称为抑制引起的遗忘(suppression- induced forgetting, SIF)。SIF的神经机制涉及右背外侧前额叶皮层和额中回的激活增加, 以及海马的激活受到抑制。SIF的程度受记忆材料的情绪效价、个体的病理性情绪状态和训练的影响。未来研究应在深入了解SIF神经机制的基础上, 考虑如何提高临床病理性记忆的SIF效果, 以达到治疗目的。     

Neural codes for conscious vision

What makes cerebral activity conscious - is it the location of the activity or the type of processing taking place? A recent study of binocular fusion adds to the increasing evidence in favour of processing, not location, as the crucial factor. It also provides evidence for the co-localization of perceptual and non-perceptual activity within individual cortical areas.     

Neural computations of decision utility

How are decision alternatives represented in the primate brain? A recent study by Sugrue et al. sought to answer this question by integrating behavioral, computational and physiological methods in examining the choice patterns of monkeys placed in a dynamic foraging environment. They observed specific encoding of the relative value of alternatives by neurons in the parietal cortex, providing an important starting point for researchers interested in how value and probability are combined in the brain to arrive at decision outcomes.     

Neural correlates of person recognition

Rapidly identifying known individuals is an essential skill in human society. To elucidate the neural basis of this skill, we monitored brain activity while experimental participants demonstrated their ability to recognize people on the basis of viewing their faces. Each participant first memorized the faces of 20 individuals who were not known to the participants in advance. Each face was presented along with a voice simulating the individual speaking their name and a biographical fact. Following this learning procedure, the associated verbal information could be recalled accurately in response to each face. These learned faces were subsequently viewed together with new faces in a memory task. Subjects made a yes-no recognition decision in response to each face while also covertly retrieving the person-specific information associated with each learned face. Brain activity that accompanied this retrieval of person-specific information was contrasted to that when new faces were processed. Functional magnetic resonance imaging in 10 participants showed that several brain regions were activated during blocks of learned faces, including left hippocampus, left middle temporal gyrus, left insula, and bilateral cerebellum. Recordings of event-related brain potentials in 10 other participants tracked the time course of face processing and showed that learned faces engaged neural activity responsible for person recognition 300-600 msec after face onset. Collectively, these results suggest that the visual input of a recently learned face can rapidly trigger retrieval of associated person-specific information through reactivation of distributed cortical networks linked via hippocampal connections.