Right hemisphere semantic processing of visual words in an aphasic patient: an fMRI study

This study was designed to identify the neural network supporting the semantic processing of visual words in a patient with large-scale damage to left-hemisphere (LH) language structures. Patient GP, and a control subject, RT, performed semantic and orthographic tasks while brain-activation patterns were recorded using functional magnetic resonance imaging. In RT, the semantic-orthographic comparison activated LH perisylvian and extrasylvian temporal regions comparable to the network of areas activated by non-brain-damaged subjects in other neuroimaging studies of semantic discrimination. In GP, the same comparison activated homologous right-hemisphere regions, demonstrating the ability of the right hemisphere to subserve visual lexicosemantic processes. The results are discussed within the context of the normal right hemisphere's capacity for semantic processing of visual words. Examining results from functional neuroimaging studies on recovery in the context of innate hemispheric abilities may enable reconciliation of disparate claims about mechanisms supporting recovery from aphasia.     

Systemic and local administration of estradiol into the prefrontal cortex or hippocampus differentially alters working memory

The influence of estradiol on learning and memory is dependent on a number of factors. The effects of physiological levels of estradiol on the acquisition of a spatial working memory task mediated by the prefrontal cortex (PFC) and the hippocampus were examined in Experiment 1. Ovariectomized Long-Evans rats received daily injections of estradiol or vehicle were tested on the win-shift version of the radial arm maze. A high dose of estradiol benzoate (5 microg) enhanced acquisition of the task, whereas a low dose of estradiol (0.3 microg) increased the number of errors committed over 17 days of testing. Experiment 2 was conducted to examine site-specific influences of estradiol on spatial working memory in well-trained rats. Saline and estradiol cyclodextrin (0.1 and 0.9 microg) were infused into the prelimbic region of the PFC or dorsal hippocampus 40 min prior to testing on the win-shift task. Infusions of estradiol into both brain areas attenuated saline-infusion disruptions in working memory. Specifically, the higher dose of estradiol facilitated working memory when infused into the PFC, whereas the lower dose of estradiol facilitated performance when infused into the dorsal hippocampus. Moreover, working memory was significantly impaired 24 h after infusions of estradiol into the dorsal hippocampus but not the PFC. These data provide further evidence for the notion that estradiol can dose-dependently alter memory processes and suggest that facilitation or disruptions of working memory by estradiol are site- and time-specific.     

The role of phonological activation in the visual semantic retrieval of Chinese characters

The Stroop paradigm was used to examine the role of phonological activation in semantic access and its development in reading Chinese characters. Subjects (age 7-23 years) of different reading ability were asked to name the display color of Chinese characters. The characters were displayed in four different colors: red, yellow, blue and green. There were five types of relationships between a character and its display color: semantically congruent, phonologically congruent, semantically incongruent, phonologically incongruent and neutral. In addition to the classical Stroop effects, interference and facilitation effects from the homophones of color characters were also observed. The younger children and those with lower reading ability exhibited stronger Stroop effects. These findings suggest that phonological codes are activated automatically in Chinese character recognition. Furthermore, there is more phonological activation in the semantic retrieval of children in lower grades and those with lower reading ability.     

Coding infant engagement in the Face-to-Face Still-Face paradigm using deep neural networks

BackgroundThe Face-to-Face Still-Face (FFSF) task is a validated and commonly used observational measure of mother-infant socio-emotional interactions. With the ascendence of deep learning-based facial emotion recognition, it is possible that common complex tasks, such as the coding of FFSF videos, could be coded with a high degree of accuracy by deep neural networks (DNNs). The primary objective of this study was to test the accuracy of four DNN image classification models against the coding of infant engagement conducted by two trained independent manual raters.Methods68 mother-infant dyads completed the FFSF task at three timepoints. Two trained independent raters undertook second-by-second manual coding of infant engagement into one of four classes: 1) positive affect, 2) neutral affect, 3) object/environment engagement, and 4) negative affect.ResultsTraining four different DNN models on 40,000 images, we achieved a maximum accuracy of 99.5% on image classification of infant frames taken from recordings of the FFSF task with a maximum inter-rater reliability (Cohen's κ-value) of 0.993.LimitationsThis study inherits all sampling and experimental limitations of the original study from which the data was taken, namely a relatively small and primarily White sample.ConclusionsBased on the extremely high classification accuracy, these findings suggest that DNNs could be used to code infant engagement in FFSF recordings. DNN image classification models may also have the potential to improve the efficiency of coding all observational tasks with applications across multiple fields of human behavior research.     

探究事件相关脑电/脑磁信号中的神经表征模式：基于分类解码和表征相似性分析的方法

探究不同心智活动下的神经表征差异, 是认知神经科学关注的核心问题之一。早期的脑电/脑磁分析方法主要关注组平均后的神经响应水平, 这要求在关注的时间进程上, 各个被试在相同刺激条件下事件相关电位/事件相关磁场的振幅大小和方向、以及地形图分布和极性均要有较高的一致性。近些年来, 研究者们将功能性磁共振成像研究中常用到的两种技术——机器学习中的分类算法(即基于分类的解码)和表征相似性分析——引入到了脑电/脑磁数据分析中。这两种新技术可以克服传统脑电/脑磁数据基于具体电压/磁感应强度波形平均分析的缺点, 具有在个体水平上探究神经表征编码的特点, 为人们探究大脑在不同时间进程上如何对特定的神经表征信息进行动态编码提供了新的思路。两种技术基于不同的方法学原理来抽提个体间一致的脑认知加工机制, 还为脑电/脑磁研究开展跨时域、跨任务、跨模态、跨群体比较不同认知过程中的表征差异提供了更多新颖的途径。我们首先通过与传统的脑电/脑磁分析方法进行比较, 系统性介绍了基于分类的解码和表征相似性分析的原理和操作流程, 之后对两种方法的应用场景进行了梳理, 并在最后对未来可供研究的方向提出了我们的见解。     

Cognitive and Behavioral Contributions to Depression Severity,Quality of Life,and Functioning Among People Living With HIV in South Africa

Cognitive-behavioral treatments for depression typically address both behavioral (e.g., activation) and cognitive (e.g., rumination) components, and consequently improve quality of life (QOL) and function in high-resource settings. However, little is known about the cross-cultural applicability and relative contribution of these components to depression symptom severity, QOL, and functional impairment in South Africa and other resource-limited global settings with high HIV prevalence rates.Persons with HIV (N = 274) from a peri-urban community outside Cape Town, South Africa, were administered multiple measures of depression (Hamilton Depression Scale, Centre for Epidemiological Studies Depression Scale, South African Depression Scale), cognitive and behavioral components related to depression (Ruminative Response Scale, Behavioral Activation for Depression Scale), and measures of QOL and functioning (Sheehan Disability Scale, Quality of Life Enjoyment and Satisfaction Scale—Short Form). Multiple linear regression models were fit to assess the relative contribution of behavioral and cognitive components to depression severity, QOL, and functional impairment in this population.Models accounting for age and sex revealed that lower levels of behavioral activation (BA) were significantly associated with all measures of depression, as well as with QOL and functional impairment (all ps < .01). Rumination was associated with all measures of depression (all ps < .01), but not with QOL or functional impairment.The consistent and unique association of BA with depression, QOL, and functional impairment bolsters its importance as a treatment target for this population.     

连绵词语音类似的启动效应

研究采用快速启动范型考察了视觉连绵词加工中语音的自动激活.两个实验发现,无论是首字同音还是尾字同音启动都可以促进对连绵词的命名反应,且基本不受目标词熟悉性(词频高低)及真假词的影响.可以推断,至少在连绵词--这种单词素词的词汇通达中存在语音的自动激活,这种语音的激活可以说是前词的装配与词水平上的改变.     

Eighteen-month-olds integrate verbal cues into their action processing: Evidence from ERPs and mu power

Behavioral research has shown that infants use both behavioral cues and verbal cues when processing the goals of others’ actions. For instance, 18-month-olds selectively imitate an observed goal-directed action depending on its (in)congruence with a model’s previous verbal announcement of a desired action goal. This EEG-study analyzed the electrophysiological underpinnings of these behavioral findings on the two functional levels of conceptual action processing and motor activation. Mid-latency mean negative ERP amplitude and mu-frequency band power were analyzed while 18-month-olds (N = 38) watched videos of an adult who performed one out of two potential actions on a novel object. In a within-subjects design, the action demonstration was preceded by either a congruent or an incongruent verbally announced action goal (e.g., “up” or “down” and upward movement). Overall, ERP negativity did not differ between conditions, but a closer inspection revealed that in two subgroups, about half of the infants showed a broadly distributed increased mid-latency ERP negativity (indicating enhanced conceptual action processing) for either the congruent or the incongruent stimuli, respectively. As expected, mu power at sensorimotor sites was reduced (indicating enhanced motor activation) for congruent relative to incongruent stimuli in the entire sample. Both EEG correlates were related to infants’ language skills. Hence, 18-month-olds integrate action-goal-related verbal cues into their processing of others’ actions, at the functional levels of both conceptual processing and motor activation. Further, cue integration when inferring others’ action goals is related to infants’ language proficiency.     

Brain tumor detection using fusion of hand crafted and deep learning features

The perilous disease in the worldwide now a days is brain tumor. Tumor affects the brain by damaging healthy tissues or intensifying intra cranial pressure. Hence, rapid growth in tumor cells may lead to death. Therefore, early brain tumor diagnosis is a more momentous task that can save patient from adverse effects. In the proposed work, the Grab cut method is applied for accurate segmentation of actual lesion symptoms while Transfer learning model visual geometry group (VGG-19) is fine-tuned to acquire the features which are then concatenated with hand crafted (shape and texture) features through serial based method. These features are optimized through entropy for accurate and fast classification and fused vector is supplied to classifiers. The presented model is tested on top medical image computing and computer-assisted intervention (MICCAI) challenge databases including multimodal brain tumor segmentation (BRATS) 2015, 2016, and 2017 respectively. The testing results with dice similarity coefficient (DSC) achieve 0.99 on BRATS 2015, 1.00 on BRATS 2015 and 0.99 on BRATS 2017 respectively.     

Hierarchical growing grid networks for skeleton based action recognition

In this paper, a novel cognitive architecture for action recognition is developed by applying layers of growing grid neural networks. Using these layers makes the system capable of automatically arranging its representational structure. In addition to the expansion of the neural map during the growth phase, the system is provided with a prior knowledge of the input space, which increases the processing speed of the learning phase. Apart from two layers of growing grid networks the architecture is composed of a preprocessing layer, an ordered vector representation layer and a one-layer supervised neural network. These layers are designed to solve the action recognition problem. The first-layer growing grid receives the input data of human actions and the neural map generates an action pattern vector representing each action sequence by connecting the elicited activation of the trained map. The pattern vectors are then sent to the ordered vector representation layer to build the time-invariant input vectors of key activations for the second-layer growing grid. The second-layer growing grid categorizes the input vectors to the corresponding action clusters/sub-clusters and finally the one-layer supervised neural network labels the shaped clusters with action labels. Three experiments using different datasets of actions show that the system is capable of learning to categorize the actions quickly and efficiently. The performance of the growing grid architecture is compared with the results from a system based on Self-Organizing Maps, showing that the growing grid architecture performs significantly superior on the action recognition tasks.