Distinguishing multi-voxel patterns and mean activation: Why, how, and what does it tell us?

The introduction of multi-voxel pattern analysis (MVPA) to the functional magnetic resonance imaging (fMRI) community has brought a deeper appreciation for the diverse forms of information that can be present within fMRI activity. The conclusions drawn from MVPA investigations are frequently influenced by both the ability to decode information from multi-voxel patterns and mean activation levels. In practice, MVPA studies vary widely in why and how they test for differing overall response levels. In this article, I examine the place of univariate information in MVPA investigations. I first discuss the variety of interpretations given to finding univariate response differences. I go on to discuss some of the analysis approaches used to investigate and compare univariate and multivariate sources of information, which can illuminate their respective contributions. It will be important for the MVPA and general fMRI community to continue to discuss and debate these important issues.     

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。     

利用脑成像多体素模式分析解码认知的神经表征:原理和应用

多体素模式分析(multi-voxel pattern analysis,MVPA)是一种基于机器学习理论发展出来的新的功能磁共振数据分析技术。MVPA技术通过训练分类器,对由不同认知状态引起的多体素信号模式进行分类。与传统的基于单个体素的分析方法相比,该技术可更敏感地检测脑对认知状态的表征,并使得从神经信号解码认知状态成为可能。文章介绍MVPA技术的基本原理、分析步骤以及可以用MVPA来解决的科学问题,并对应用中可能面临的问题提供了参考建议。     

Extending problem-solving procedures through reflection

A large-sample (n = 75) fMRI study guided the development of a theory of how people extend their problem-solving procedures by reflecting on them. Both children and adults were trained on a new mathematical procedure and then were challenged with novel problems that required them to change and extend their procedure to solve these problems. The fMRI data were analyzed using a combination of hidden Markov models (HMMs) and multi-voxel pattern analysis (MVPA). This HMM–MVPA analysis revealed the existence of 4 stages: Encoding, Planning, Solving, and Responding. Using this analysis as a guide, an ACT-R model was developed that improved the performance of the HMM–MVPA and explained the variation in the durations of the stages across 128 different problems. The model assumes that participants can reflect on declarative representations of the steps of their problem-solving procedures. A Metacognitive module can hold these steps, modify them, create new declarative steps, and rehearse them. The Metacognitive module is associated with activity in the rostrolateral prefrontal cortex (RLPFC). The ACT-R model predicts the activity in the RLPFC and other regions associated with its other cognitive modules (e.g., vision, retrieval). Differences between children and adults seemed related to differences in background knowledge and computational fluency, but not to the differences in their capability to modify procedures.     

人声加工的神经机制

人声是人类听觉环境中最熟知和重要的声音, 传递着大量社会相关信息。与视觉人脸加工类似, 大脑对人声也有着特异性加工。研究者使用电生理、脑成像等手段找到了对人声有特异性反应的脑区, 即颞叶人声加工区(TVA), 并发现非人类动物也有类似的特异性加工区域。人声加工主要涉及言语、情绪和身份信息的加工, 分别对应于三条既相互独立又相互作用的神经通路。研究者提出了双通路模型、多阶段模型和整合模型分别对人声的言语、情绪和身份加工进行解释。未来研究需要进一步讨论人声加工的特异性能否由特定声学特征的选择性加工来解释, 并深入探究特殊人群(如自闭症和精神分裂症患者)的人声加工的神经机制。     

A connectionist model of category learning by individuals with high-functioning autism spectrum disorder

Individuals with autism spectrum disorder (ASD) show atypical patterns of learning and generalization. We explored the possible impacts of autism-related neural abnormalities on perceptual category learning using a neural network model of visual cortical processing. When applied to experiments in which children or adults were trained to classify complex two-dimensional images, the model can account for atypical patterns of perceptual generalization. This is only possible, however, when individual differences in learning are taken into account. In particular, analyses performed with a self-organizing map suggested that individuals with high-functioning ASD show two distinct generalization patterns: one that is comparable to typical patterns, and a second in which there is almost no generalization. The model leads to novel predictions about how individuals will generalize when trained with simplified input sets and can explain why some researchers have failed to detect learning or generalization deficits in prior studies of category learning by individuals with autism. On the basis of these simulations, we propose that deficits in basic neural plasticity mechanisms may be sufficient to account for the atypical patterns of perceptual category learning and generalization associated with autism, but they do not account for why only a subset of individuals with autism would show such deficits. If variations in performance across subgroups reflect heterogeneous neural abnormalities, then future behavioral and neuroimaging studies of individuals with ASD will need to account for such disparities.     

Processing Faces and Facial Expressions

This paper reviews processing of facial identity and expressions. The issue of independence of these two systems for these tasks has been addressed from different approaches over the past 25 years. More recently, neuroimaging techniques have provided researchers with new tools to investigate how facial information is processed in the brain. First, findings from traditional approaches to identity and expression processing are summarized. The review then covers findings from neuroimaging studies on face perception, recognition, and encoding. Processing of the basic facial expressions is detailed in light of behavioral and neuroimaging data. Whereas data from experimental and neuropsychological studies support the existence of two systems, the neuroimaging literature yields a less clear picture because it shows considerable overlap in activation patterns in response to the different face-processing tasks. Further, activation patterns in response to facial expressions support the notion of involved neural substrates for processing different facial expressions.     

Identifying bilingual semantic neural representations across languages

The goal of the study was to identify the neural representation of a noun’s meaning in one language based on the neural representation of that same noun in another language. Machine learning methods were used to train classifiers to identify which individual noun bilingual participants were thinking about in one language based solely on their brain activation in the other language. The study shows reliable (p < .05) pattern-based classification accuracies for the classification of brain activity for nouns across languages. It also shows that the stable voxels used to classify the brain activation were located in areas associated with encoding information about semantic dimensions of the words in the study. The identification of the semantic trace of individual nouns from the pattern of cortical activity demonstrates the existence of a multi-voxel pattern of activation across the cortex for a single noun common to both languages in bilinguals.     

情绪变化的自动化加工：来自EMMN的启示

人脑如何自动化加工瞬息万变的情绪信息?研究者们在借鉴听觉通道的失匹配负波(mismatch negativity, MMN)研究的基础上, 进一步发展出了表情失匹配负波(expression mismatch negativity, EMMN), 以此作为视觉情绪信息前注意加工的重要指标。与以往的一般视觉线索的视觉失匹配负波(visual mismatch negativity, vMMN)有所区别, EMMN研究专注于人脑如何自动化加工瞬息万变的情绪信息。当前的研究主要探讨了不同类型的面部表情、不同性别、高低流体智力个体的EMMN差异, 以及自闭症、抑郁症、精神分裂症等异常个体EMMN的特点。此外, 从预测编码的角度阐释了EMMN的机制。今后的研究有必要聚焦EMMN在临床诊断和治疗中的应用, 考察不同情绪线索EMMN的特点, 并进一步揭示EMMN的神经机制。     

自闭症谱系障碍人群词义加工的脑激活模式：基于fMRI研究的元分析

本研究筛选了11项采用功能性磁共振成像技术探究言语自闭症人群词义加工的研究, 探讨了该人群与典型人群脑激活模式的差异是否具有跨研究的稳定性。结果表明, 差异的脑激活模式稳定存在, 且表现为主要涉及左额上回的典型脑区激活不足。该结果为言语ASD人群语言加工的神经机制提供了来自词义加工的跨研究激活证据, 在明确“减弱的额叶激活”这一稳定差异表现的基础上, 强调了针对不同语言加工任务开展元分析研究的必要性。     

Longitudinal relations between psychological distress and moderate-to-vigorous physical activity: A latent change score approach

ObjectivesThe effect of physical inactivity on mental health risk is well established; however, less is known about about how psychological distress might deter participation in physical activity. Guided by advancements in the treatment of longitudinal data, the aim of this study was to examine patterns and predictors of change in moderate-to-vigorous physical activity (MVPA) and psychological distress (e.g., feeling nervous, worthless).Design methodAustralian adults (4944 females, M_age = 34.63 years ±5.34; 4322 males, M_age = 37.51 years ±6.14) provided baseline data as part of the Longitudinal Study of Australian Children (LSAC) and were followed for measurements every two years for 10 years.ResultsLatent change score analyses revealed support for a reciprocal effects model, whereby change in MVPA and psychological distress occurred as a function of individuals’ prior levels of, and/or prior change in these variables.ConclusionsThis investigation is the first to document that changes in MVPA and psychological distress are coupled temporally. Notably, we observed that individuals’ distress levels at a given time point predicted subsequent change on both MVPA and distress; a finding which provides novel and important insight into how adults’ activity levels and psychological distress fluctuate relative to one another.     

降低决策冲动性的方法及其神经机制

大量实验研究显示, 成瘾患者、多动症患者和病态赌博者在决策中表现出冲动偏好。目前研究冲动性决策偏好的经典实验范式是延迟折扣任务。通过延迟折扣任务, 心理学家揭示了冲动性决策偏好的神经机制并构建了多种理论, 包括单一系统评价理论、双系统评价理论、自我控制理论和自我参照加工理论。基于这些理论, 已经发展出了多种降低决策冲动性的方法, 包括想象未来具体事件、工作记忆训练、预先承诺以及提高血液中葡萄糖水平, 其背后的神经机制是通过认知控制和价值表征理论来实现的。未来研究方向需集中于价值计算、价值系统和多体素模式分析(MVPA)在改善决策冲动性以及内在机制研究上的应用。     

老年人整体运动敏感性降低的脑灰质体积变化基础

本研究使用多体素模式分析方法和随机点阵范式,首次探索了老年人整体运动敏感性（GMS）下降和其脑灰质体积下降间的关系。结果发现,老年人GMS显著低于年轻人,老年人V5/MT、V3和额顶等脑区灰质体积显著小于年轻人;V5/MT和V3区的灰质体积信号可以有效预测个体的GMS。结果提示,V5/MT区和V3区,特别是V3区的灰质体积下降可能是老年人GMS下降的主要原因;而基于“去分化假设”,全脑范围内其余脑区的灰质体积下降可能也在一定程度上与老年人GMS下降有关。     

Phases of learning: How skill acquisition impacts cognitive processing

This fMRI study examines the changes in participants’ information processing as they repeatedly solve the same mathematical problem. We show that the majority of practice-related speedup is produced by discrete changes in cognitive processing. Because the points at which these changes take place vary from problem to problem, and the underlying information processing steps vary in duration, the existence of such discrete changes can be hard to detect. Using two converging approaches, we establish the existence of three learning phases. When solving a problem in one of these learning phases, participants can go through three cognitive stages: Encoding, Solving, and Responding. Each cognitive stage is associated with a unique brain signature. Using a bottom-up approach combining multi-voxel pattern analysis and hidden semi-Markov modeling, we identify the duration of that stage on any particular trial from participants brain activation patterns. For our top-down approach we developed an ACT-R model of these cognitive stages and simulated how they change over the course of learning. The Solving stage of the first learning phase is long and involves a sequence of arithmetic computations. Participants transition to the second learning phase when they can retrieve the answer, thereby drastically reducing the duration of the Solving stage. With continued practice, participants then transition to the third learning phase when they recognize the problem as a single unit and produce the answer as an automatic response. The duration of this third learning phase is dominated by the Responding stage.     

Practice strategies of musicians modulate neural processing and the learning of sound-patterns

Previous studies suggest that pre-attentive auditory processing of musicians differs depending on the strategies used in music practicing and performance. This study aimed at systematically revealing whether there are differences in auditory processing between musicians preferring and not-preferring aural strategies such as improvising, playing by ear, and rehearsing by listening to records. Participants were assigned to aural and non-aural groups according to how much they employ aural strategies, as determined by a questionnaire. The change-related mismatch negativity (MMN) component of event-related brain potentials (ERPs) was used to probe pre-attentive neural discrimination of simple sound features and melody-like patterns. Further, the musicians' behavioral accuracy in sound perception was tested with a discrimination task and the AMMA musicality test. The data indicate that practice strategies do not affect musicians' pre-attentive neural discrimination of changes in simple sound features but do modulate the speed of neural discrimination of interval and contour changes within melody-like patterns. Moreover, while the aural and non-aural groups did not differ in their initial neural accuracy for discriminating melody-like patterns, they differed after a focused training session. A correlation between behavioral and neural measures was also obtained. Taken together, these results suggest that auditory processing of musicians who prefer aural practice strategies differs in melodic contour and interval processing and perceptual learning, rather than in simple sound processing, in comparison to musicians preferring other practice strategies.     

The role of inferior frontal cortex in phonological processing

Recent neuroimaging studies of language processing are examining the neural substrate of phonology because of its critical role in mapping sound information onto higher levels of language processing (e.g., words) as well as providing codes in which verbal information can be temporarily stored in working memory. However, the precise role of the inferior frontal cortex in spoken and written phonological tasks has remained elusive. Although lesion studies have indicated the presence of selective deficits in phonological processing, the location of lesions underlying these impairments has not revealed a consistent pattern. Despite efforts to refine methods and tasks, functional neuroimaging studies have also revealed variability in activation patterns. Reanalysis of evidence from these neuroimaging studies suggests that there are functional subregions within the inferior frontal gyrus that correspond to specific components of phonological processing (e.g., orthographic to phonological conversion in reading, and segmentation in speech).     

Momentary affective response to bouts of moderate-to-vigorous physical activity predicts changes in physical activity and sedentary behavior during behavioral weight loss

BackgroundAffective responses are posited to be key predictors of the uptake and maintenance of health behaviors. However, few studies have examined how individuals’ affective response to physical activity, as well as the degree to which their affect response changes, may predict changes in physical activity and sedentary time during behavioral weight loss treatment.PurposeThe current study examined how baseline momentary affective response (i.e., stress and anxiety) to moderate-to-vigorous physical activity (MVPA) and the degree of pre--post intervention change in this response predicted change in daily sedentary, light, and MVPA time during a three-month internet-based weight loss program.MethodsWomen with overweight/obesity (final N = 37) completed 14-day ecological momentary assessment (EMA) protocols with objective measurement of physical activity (i.e., bout-related MVPA time) before and after the intervention.ResultsWomen who had more reinforcing responses to MVPA (i.e., greater reductions in anxiety and stress response following MVPA bouts) at baseline had greater increases in overall MVPA at the end of the intervention. Those who had greater anxiety reductions after MVPA bouts at baseline also evidenced less sedentary time at the end of the intervention. Changes in affective responses across the intervention were not related to changes in physical activity levels.ConclusionsFindings suggest initial levels of affective reinforcement from MVPA bouts predict future change in MVPA and sedentary time during behavioral weight loss. Future work is needed to examine the utility of more precisely targeting affective responses to physical activity to optimize intervention approaches.     

Neural decoding of positive and negative self-knowledge

A prevalent explanation for the self-reference effect is that self-knowledge is represented by a set of specific brain regions, including anterior cingulate cortex (ACC), middle frontal gyrus (MFG), superior temporal gyrus (STG), precuneus, and inferior parietal lobule (IPL), which enables self-knowledge to be processed in priority than other-knowledge. However, the conventional univariate activation analysis adopted by previous studies could only detect the activation of separate brain regions. The current study mainly investigated the global neural patterns of self-knowledge (relative to other-knowledge) by the multivariate pattern analysis (MVPA). Results obtained in Experiments 1 and 2 were highly consistent, indicating that the core self-network (mainly the ACC) and salience network (mainly the insula) could distinguish self-knowledge from other-knowledge. Furthermore, the neural pattern of positive self-knowledge mainly included the ventral part of ACC, while the neural pattern of negative self-knowledge mainly included the ventral and dorsal parts of ACC and cognitive control network (dorsolateral prefrontal cortex: dlPFC). These findings suggest that the core self-network and salience network are specific to the neural process of self-knowledge. Moreover, both positive and negative self-knowledge are separately driven by different cognitive and neural characteristics.

     

面孔审美加工的神经机制及个体差异

外显和内隐的面孔审美加工的脑成像研究表明,美的面孔比不美的面孔导致眶额皮层、伏隔核、杏仁核等奖赏脑区更大的激活。脑电研究则发现了与面孔审美加工有关的早期负波和晚期正成分。面孔审美加工有关的脑区活动也受到性别、生理周期等个体因素的调节。未来的研究需要比较美的面孔与其它奖赏刺激加工的神经机制,探讨面孔审美加工的不同阶段及时间进程,在面孔知觉的框架下形成合理的面孔审美加工的神经模型。     

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.