Cognitive neuroimaging: cognitive science out of the armchair

Cognitive scientists were not quick to embrace the functional neuroimaging technologies that emerged during the late 20th century. In this new century, cognitive scientists continue to question, not unreasonably, the relevance of functional neuroimaging investigations that fail to address questions of interest to cognitive science. However, some ultra-cognitive scientists assert that these experiments can never be of relevance to the study of cognition. Their reasoning reflects an adherence to a functionalist philosophy that arbitrarily and purposefully distinguishes mental information-processing systems from brain or brain-like operations. This article addresses whether data from properly conducted functional neuroimaging studies can inform and subsequently constrain the assumptions of theoretical cognitive models. The article commences with a focus upon the functionalist philosophy espoused by the ultra-cognitive scientists, contrasting it with the materialist philosophy that motivates both cognitive neuroimaging investigations and connectionist modelling of cognitive systems. Connectionism and cognitive neuroimaging share many features, including an emphasis on unified cognitive and neural models of systems that combine localist and distributed representations. The utility of designing cognitive neuroimaging studies to test (primarily) connectionist models of cognitive phenomena is illustrated using data from functional magnetic resonance imaging (fMRI) investigations of language production and episodic memory.     

Beta-MPT: Multinomial processing tree models for addressing individual differences

Traditionally, multinomial processing tree (MPT) models are applied to groups of homogeneous participants, where all participants within a group are assumed to have identical MPT model parameter values. This assumption is unreasonable when MPT models are used for clinical assessment, and it often may be suspect for applications to ordinary psychological experiments. One method for dealing with parameter variability is to incorporate random effects assumptions into a model. This is achieved by assuming that participants’ parameters are drawn independently from some specified multivariate hyperdistribution. In this paper we explore the assumption that the hyperdistribution consists of independent beta distributions, one for each MPT model parameter. These beta-MPT models are ‘hierarchical models’, and their statistical inference is different from the usual approaches based on data aggregated over participants. The paper provides both classical (frequentist) and hierarchical Bayesian approaches to statistical inference for beta-MPT models. In simple cases the likelihood function can be obtained analytically; however, for more complex cases, Markov Chain Monte Carlo algorithms are constructed to assist both approaches to inference. Examples based on clinical assessment studies are provided to demonstrate the advantages of hierarchical MPT models over aggregate analysis in the presence of individual differences.     

Functional magnetic resonance imaging: Overview and methods for psychological research

Functional magnetic resonance imaging (fMRI) allows noninvasive imaging of hemodynamic changes related to neural activity. This technique can be used in single-subject designs and can provide millimeter spatial resolution and temporal resolution in the range of 5–10 sec. This paper provides a brief introduction to MRI techniques and their application to functional neuroimaging, focusing on methodological issues that are of particular concern to psychologists, including methods for presenting computerized stimuli to subjects without disrupting the scanner, experimental design issues, and statistical analysis and image processing procedures. To illustrate methodological issues, recent results from a series of studies looking at the topographic organization of visual cortex are presented. General issues concerning limitations in this technique, future directions in its development, its relationship to other neuroimaging techniques, and the role of functional neuroimaging in psychological research are addressed in the Discussion.     

Functional neuroimaging in neurology and psychiatry

How can functional neuroimaging be applied to clinical neurology and psychiatry? This article reviews selected contributions of functional neuroimaging to the clinical neurosciences. We review selected technical aspects of positron emission tomography, single photon emission tomography, and functional magnetic resonance imaging with a focus on the relative strengths and weaknesses of these techniques. Consumers of functional neuroimaging research are encouraged to consider the limitations of imaging techniques and theoretical pitfalls of cognitive task design when interpreting results of functional imaging studies. Then, we selectively review the contributions of functional neuroimaging to neurology and psychiatry, including the areas of epilepsy, stroke, chronic pain, schizophrenia, depression, and obsessive-compulsive disorder. Future directions of functional neuroimaging research are offered, with the emphasis that the best conclusions are informed by a convergence of research from functional neuroimaging, neurophysiological, and lesion studies.     

Lexical organization and competition in first and second languages: computational and neural mechanisms

How does a child rapidly acquire and develop a structured mental organization for the vast number of words in the first years of life? How does a bilingual individual deal with the even more complicated task of learning and organizing two lexicons? It is only until recently have we started to examine the lexicon as a dynamical system with regard to its acquisition, representation, and organization. In this article, I outline a proposal based on our research that takes the dynamical approach to the lexicon, and I discuss how this proposal can be applied to account for lexical organization, structural representation, and competition within and between languages. In particular, I provide computational evidence based on the DevLex model, a self-organizing neural network model, and neuroimaging evidence based on functional magnetic resonance imaging (fMRI) studies, to illustrate how children and adults learn and represent the lexicon in their first and second languages. In the computational research, our goal has been to identify, through linguistically and developmentally realistic models, detailed cognitive mechanisms underlying the dynamic self-organizing processes in monolingual and bilingual lexical development; in the neuroimaging research, our goal has been to identify the neural substrates that subserve lexical organization and competition in the monolingual and the bilingual brain. In both cases, our findings lead to a better understanding of the interactive dynamics involved in the acquisition and representation of one or multiple languages.     

生物运动信息加工模型

人类可以从生物体的各种运动行为中获得丰富的社会信息,以满足社会交往的需求。视觉系统对生物运动信息的加工是一个复杂的过程,不同于对其他普通客体的加工能力。研究者们采用不同的方法,分别从各自的角度来研究这一过程,同时也建立了一系列模型。其中早期模型关注视觉系统加工生物运动信息的过程和方法;近期模型则采用脑成像手段构建生物运动信息加工的神经网络。这些模型包含了很多有价值的研究成果,但是也存在需要进一步完善的地方。     

如何利用开源工具分析fMRI数据：基于多体素模式的有监督机器学习算法介绍

摘要：目前,多体素模式分析（MVPA）日渐普遍地应用于脑影像研究。近些年,机器学习的模式分类等算法在MVPA方法中被广泛应用,因其具有能够抽取高维数据模式,提高数据利用率的优点。其中一种典型的应用是利用解码的思想来解决神经表征问题,本文主要介绍了利用基于Python语言的工具库中有监督学习算法分析数据的过程。除介绍Nilearn结合Scikit-learn分析数据的步骤外,还比较不同算法的效率,为算法的选择及参数设备提供具体参考。     

脑成像技术及其在决策研究中的应用

先进的无创神经影像技术(如EEG和fMRI)允许研究者直接观察被试在完成多种知觉、运动和认知任务时的大脑活动。将脑功能成像与严密的实验设计和数据分析方法结合起来, 我们可以考察大脑不同脑区的功能以及它们之间的交互作用。随着脑功能成像技术在研究人类决策行为中的日益成功运用, 一个被称为神经经济学的新兴领域正在逐渐形成和发展起来。本文中首先对脑成像技术进行一个总体介绍, 重点在于探讨近年来在多体素分析和多模态数据整合的最新进展。接下来, 我们以风险决策、跨时间选择以及社会决策领域的几个研究为例, 阐述神经影像技术如何能加深和拓展我们对人类决策的认识。最后, 我们讨论了神经经济学中研究中面临的一些挑战以及未来的研究方向。     

Moving beyond pure signal-detection models: comment on Wixted (2007)

The dual-process signal-detection (DPSD) model assumes that recognition memory is based on recollection of qualitative information or on a signal-detection-based familiarity process. The model has proven useful for understanding results from a wide range of memory research, including behavioral, neuropsychological, electrophysiological, and neuroimaging studies. However, a number of concerns have been raised about the model over the years, and it has been suggested that an unequal-variance signal-detection (UVSD) model that incorporates separate recollection and familiarity processes (J. T. Wixted, 2007) may provide an equally good, or even better, account of the data. In this article, the authors show that the results of studies that differentiate these models support the predictions of the DPSD model and indicate that recognition does not reflect the summing of 2 signal-detection processes, as the new UVSD model assumes. In addition, the assumptions of the DPSD model are clarified in order to address some of the common misconceptions about the model. Although important challenges remain, hybrid models such as this provide a more useful framework within which to understand human memory than do pure signal-detection models.     

Determining informative priors for cognitive models

The development of cognitive models involves the creative scientific formalization of assumptions, based on theory, observation, and other relevant information. In the Bayesian approach to implementing, testing, and using cognitive models, assumptions can influence both the likelihood function of the model, usually corresponding to assumptions about psychological processes, and the prior distribution over model parameters, usually corresponding to assumptions about the psychological variables that influence those processes. The specification of the prior is unique to the Bayesian context, but often raises concerns that lead to the use of vague or non-informative priors in cognitive modeling. Sometimes the concerns stem from philosophical objections, but more often practical difficulties with how priors should be determined are the stumbling block. We survey several sources of information that can help to specify priors for cognitive models, discuss some of the methods by which this information can be formalized in a prior distribution, and identify a number of benefits of including informative priors in cognitive modeling. Our discussion is based on three illustrative cognitive models, involving memory retention, categorization, and decision making.     

Bayesian Fundamentalism or Enlightenment? On the explanatory status and theoretical contributions of Bayesian models of cognition

The prominence of Bayesian modeling of cognition has increased recently largely because of mathematical advances in specifying and deriving predictions from complex probabilistic models. Much of this research aims to demonstrate that cognitive behavior can be explained from rational principles alone, without recourse to psychological or neurological processes and representations. We note commonalities between this rational approach and other movements in psychology - namely, Behaviorism and evolutionary psychology - that set aside mechanistic explanations or make use of optimality assumptions. Through these comparisons, we identify a number of challenges that limit the rational program's potential contribution to psychological theory. Specifically, rational Bayesian models are significantly unconstrained, both because they are uninformed by a wide range of process-level data and because their assumptions about the environment are generally not grounded in empirical measurement. The psychological implications of most Bayesian models are also unclear. Bayesian inference itself is conceptually trivial, but strong assumptions are often embedded in the hypothesis sets and the approximation algorithms used to derive model predictions, without a clear delineation between psychological commitments and implementational details. Comparing multiple Bayesian models of the same task is rare, as is the realization that many Bayesian models recapitulate existing (mechanistic level) theories. Despite the expressive power of current Bayesian models, we argue they must be developed in conjunction with mechanistic considerations to offer substantive explanations of cognition. We lay out several means for such an integration, which take into account the representations on which Bayesian inference operates, as well as the algorithms and heuristics that carry it out. We argue this unification will better facilitate lasting contributions to psychological theory, avoiding the pitfalls that have plagued previous theoretical movements.     

基于坐标的脑成像元分析方法:激活似然性评估

随着脑成像研究数量的增长,研究者迫切需要一种能够将众多研究进行整合的元分析方法。本文在简单回顾了以往脑成像元分析方法的基础上,介绍了激活似然性评估法(ALE)的操作实施、优势特征、更新完善,以及其发展与应用。该元分析方法将每一个激活的坐标看作一种分布的可能性,通过特定的算法与拟合获得激活似然值,最后,将不同文献汇聚计算而来的ALE值与虚无分布获得的ALE值进行对比,以确定激活的可能性分布。近十年来,激活似然性评估法几经修正,已被广泛运用于各领域元分析的研究中。     

Computer-augmented decision making

The principal judgmental components of multiattribute decision making are examined here with specific reference to how these components can be captured electronically. Once captured, a function, rule, or algorithm may be executed for the integration of this information and the selection of the optimal alternative(s). Two kinds of algorithms are discussed: one based on linear models, the other on fuzzy-set theory and ratio scaling. With on-line support and certain assumptions about human biases (which lead to nonoptimal decisions), the quality of decisions can be enhanced considerably. The principal concerns are with end-user acceptance of computer augmented decisions.     

Electrophysiology in neuropsychiatric research: a network perspective

A growing number of anatomic and physiologic studies have shown that parallel sensory and motor information processing occurs in multiple cortical areas. These findings challenge the traditional model of brain processing, which states that the brain is a collection of physically discrete processing modules that pass information to each other by neuronal impulses in a stepwise manner. New concepts based on neural network models suggest that the brain is a dynamically shifting collection of interpenetrating, distributed, and transient neural networks. Neither of these models is necessarily mutually exclusive, but each gives different perspectives on the brain that might be complementary. Each model has its own research methodology, with functional magnetic resonance imaging supporting notions of modular processing, and electrophysiology (eg, electroencephalography) emphasizing the network model. These two technologies might be combined fruitfully in the near future to provide us with a better understanding of the brain. However, this common enterprise can succeed only when the inherent limitations and advantages of both models and technologies are known. After a general introduction about electrophysiology as a research tool and its relation to the network model, several practical examples are given on the generation of pathophysiologic models and disease classification, intermediate phenotyping for genetic investigations, and pharmacodynamic modeling. Finally, proposals are made about how to integrate electrophysiology and neuroimaging methods.     

Brain mechanisms for predictive control by switching internal models: implications for higher-order cognitive functions

Humans can guide their actions toward the realization of their intentions. Flexible, rapid and precise realization of intentions and goals relies on the brain learning to control its actions on external objects and to predict the consequences of this control. Neural mechanisms that mimic the input–output properties of our own body and other objects can be used to support prediction and control, and such mechanisms are called internal models. We first summarize functional neuroimaging, behavioral and computational studies of the brain mechanisms related to acquisition, modular organization, and the predictive switching of internal models mainly for tool use. These mechanisms support predictive control and flexible switching of intentional actions. We then review recent studies demonstrating that internal models are crucial for the execution of not only immediate actions but also higher-order cognitive functions, including optimization of behaviors toward long-term goals, social interactions based on prediction of others’ actions and mental states, and language processing. These studies suggest that a concept of internal models can consistently explain the neural mechanisms and computational principles needed for fundamental sensorimotor functions as well as higher-order cognitive functions. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Connectionist models of neuropsychological disorders

Artificial neural networks ('connectionist models') embody aspects of real neuronal systems. But does studying the breakdown of performance in such models help us to understand cognitive impairments in humans following brain damage? Here we review recent attempts to capture different neuropsychological disorders using connectionist models with simulated lesions. We show how such lesion studies can be used to evaluate some of the standard assumptions made in neuropsychological research, concerning both double dissociations and associations between patterns of impairment. We also illustrate how lesioned models, like humans, can sometimes be more impaired on the easier of two tasks and demonstrate that connectionist models can incorporate forms of internal structure. Finally we discuss the utility of the models for understanding and predicting the effectiveness of different rehabilitation strategies. Future questions concern the role and possible development of internal structure within these models, whether the models can be generalized to larger-scale simulations, and whether they can accommodate higher-order linguistic disorders.     

Some problems for representations of brain organization based on activation in functional imaging

Functional brain imaging has overshadowed traditional lesion studies in becoming the dominant approach to the study of brain-behavior relationships. The proponents of functional imaging studies frequently argue that this approach provides an advantage over lesion studies by observing normal brain activity in vivo without the disruptive effects of brain damage. However, the enthusiastic onslaught of brain images, frequently presented as veridical representations of mental function, has sometimes overwhelmed some basic facts about brain organization repeatedly observed over more than a century. In particular, the lateralization of speech and language to the left cerebral hemisphere in over 90% of the right-handed population does not appear to have been taken as a serious constraint in the interpretation of imaging results in studies of these functions. This paper reviews a number of areas in which standard activation assumptions yield results that are at odds with clinical experience. The activation approach will be contrasted with a performance-based analysis of functional image data, which, at least in the case of speech production, yields results in better agreement with lesion data. Functional imaging represents enormous opportunities for understanding brain-behavior relationships, but at the present level of understanding of what is being represented in such images, it is premature to adhere to a single approach based on the strong but questionable assumptions inherent in most activation studies.     

汉语语调早期加工的脑机制

采用ERP方法, 结合LORETA源定位分析技术, 考察了汉语语调早期自动加工的脑机制。结果发现： (1)对于附着在二声调汉字后的语调, 无论以单字词的形式, 还是以句子的形式呈现, 都没有诱发MMN; (2)当过滤掉实验刺激的言语信息后, 词音高和句子音高条件均诱发了MMN, 且这两个条件下的MMN波幅差异不显著; (3)LORETA分析发现, 句子音高条件在右半球顶叶的多个区域存在显著激活。该结果拓展了以往的研究结论, 为声学假设提供了证据。     

神经成像数据的元分析

随着高空间分辨率神经成像技术如fMRI和PET的普及, 神经成像研究报告的数量增长迅猛。文献的积累为研究者提供了大量的数据, 研究者可以通过对文献的分析来验证研究结论以及提出新的假设。由于神经成像研究的主要目的之一在于寻求认知过程与脑区的空间位置对应关系, 基于坐标的元分析方法满足了这种需求, 成为神经成像数据元分析中主导的方法。其中, 激活可能性估计法(Activation Likelihood Estimation, ALE)由于方法上的合理性和使用上的便利, 成为当前使用最广泛的基于坐标的元分析方法。本文首先介绍了ALE方法的基本原理, 并在此基础上讨论了神经成像数据元分析的两种主要思路：寻找多个研究的一致性以及寻找脑区激活的调节变量。此外, 文章还介绍了新近流行的脑连通性元分析模型(MACM), 即使用元分析方法进行功能连通性分析。最后, 文章讨论了当前神经成像数据元分析的发展趋势。