Learning,visualizing and exploring 16S rRNA structure using an attention-based deep neural network

Recurrent neural networks with memory and attention mechanisms are widely used in natural language processing because they can capture short and long term sequential information for diverse tasks. We propose an integrated deep learning model for microbial DNA sequence data, which exploits convolutional neural networks, recurrent neural networks, and attention mechanisms to predict taxonomic classifications and sample-associated attributes, such as the relationship between the microbiome and host phenotype, on the read/sequence level. In this paper, we develop this novel deep learning approach and evaluate its application to amplicon sequences. We apply our approach to short DNA reads and full sequences of 16S ribosomal RNA (rRNA) marker genes, which identify the heterogeneity of a microbial community sample. We demonstrate that our implementation of a novel attention-based deep network architecture, Read2Pheno, achieves read-level phenotypic prediction. Training Read2Pheno models will encode sequences (reads) into dense, meaningful representations: learned embedded vectors output from the intermediate layer of the network model, which can provide biological insight when visualized. The attention layer of Read2Pheno models can also automatically identify nucleotide regions in reads/sequences which are particularly informative for classification. As such, this novel approach can avoid pre/post-processing and manual interpretation required with conventional approaches to microbiome sequence classification. We further show, as proof-of-concept, that aggregating read-level information can robustly predict microbial community properties, host phenotype, and taxonomic classification, with performance at least comparable to conventional approaches. An implementation of the attention-based deep learning network is available at https://github.com/EESI/sequence_attention (a python package) and https://github.com/EESI/seq2att (a command line tool).     

Letter Position Coding Across Modalities: The Case of Braille Readers

Background

The question of how the brain encodes letter position in written words has attracted increasing attention in recent years. A number of models have recently been proposed to accommodate the fact that transposed-letter stimuli like jugde or caniso are perceptually very close to their base words.

Methodology

Here we examined how letter position coding is attained in the tactile modality via Braille reading. The idea is that Braille word recognition may provide more serial processing than the visual modality, and this may produce differences in the input coding schemes employed to encode letters in written words. To that end, we conducted a lexical decision experiment with adult Braille readers in which the pseudowords were created by transposing/replacing two letters.

Principal Findings

We found a word-frequency effect for words. In addition, unlike parallel experiments in the visual modality, we failed to find any clear signs of transposed-letter confusability effects. This dissociation highlights the differences between modalities.

Conclusions

The present data argue against models of letter position coding that assume that transposed-letter effects (in the visual modality) occur at a relatively late, abstract locus.     

The distribution of baboon species and a new population of gelada baboons along the Wabi-Shebeli river,Ethiopia

We conducted an extensive survey in search of hybrid baboons betweenPapio hamadryas andP. anubis along the Wabi-Shebeli river at the border of the Arusi and Bale Regions, Ethiopia. We made inquiries of villagers on the roadsides concerning the existence of baboon species. We also conducted direct observations at several sites. There are three routes which lead to the north bank of the Wabi-Shebeli river (Arusi Region), and we found hybrid baboons on the bank of the Wabi-Shebeli river in two routes among the three. We found hamadryas baboons in all of the three routes at the cliff areas. There are two routes which lead to the south bank of the Wabi-Shebeli river (Bale Region). We conducted a survey on one of the two. We found hamadryas baboons at the cliff areas of the route. We observed a population of gelada baboons along the cliff extending over 20 km along the north bank of the Wabi-Shebeli river (Arusi Region). This area is far to the south of the known distribution range of gelada baboons (Yalden et al., 1977). The gelada baboons of this area appeared to represent a different form (subspecies?) from those at Debre Sina (Showa Region) based on our observations in both areas. We reached the conclusion that the distributions of baboon species along the Wabi-Shebeli river may have been strongly affected by the intensive cultivation on the plateau of the highland. The distribution patterns of the three baboon species,P. anubis, P. hamadryas, andTheropithecus gelada, appeared to be influenced by their individual adaptabilities to the cliff environment. Hamadryas baboons were distributed continuously along the cliff and the narrow lowland of the Wabi-Shebeli river. Anubis baboons were distributed discontinuously on the cliffs, and their populations tended to be small and isolated. These anubis baboons were strongly hybridized with hamadryas baboons.     

DL-ADR: a novel deep learning model for classifying genomic variants into adverse drug reactions

Background

Genomic variations are associated with the metabolism and the occurrence of adverse reactions of many therapeutic agents. The polymorphisms on over 2000 locations of cytochrome P450 enzymes (CYP) due to many factors such as ethnicity, mutations, and inheritance attribute to the diversity of response and side effects of various drugs. The associations of the single nucleotide polymorphisms (SNPs), the internal pharmacokinetic patterns and the vulnerability of specific adverse reactions become one of the research interests of pharmacogenomics. The conventional genomewide association studies (GWAS) mainly focuses on the relation of single or multiple SNPs to a specific risk factors which are a one-to-many relation. However, there are no robust methods to establish a many-to-many network which can combine the direct and indirect associations between multiple SNPs and a serial of events (e.g. adverse reactions, metabolic patterns, prognostic factors etc.). In this paper, we present a novel deep learning model based on generative stochastic networks and hidden Markov chain to classify the observed samples with SNPs on five loci of two genes (CYP2D6 and CYP1A2) respectively to the vulnerable population of 14 types of adverse reactions.

Methods

A supervised deep learning model is proposed in this study. The revised generative stochastic networks (GSN) model with transited by the hidden Markov chain is used. The data of the training set are collected from clinical observation. The training set is composed of 83 observations of blood samples with the genotypes respectively on CYP2D6*2, *10, *14 and CYP1A2*1C, *1 F. The samples are genotyped by the polymerase chain reaction (PCR) method. A hidden Markov chain is used as the transition operator to simulate the probabilistic distribution. The model can perform learning at lower cost compared to the conventional maximal likelihood method because the transition distribution is conditional on the previous state of the hidden Markov chain. A least square loss (LASSO) algorithm and a k-Nearest Neighbors (kNN) algorithm are used as the baselines for comparison and to evaluate the performance of our proposed deep learning model.

Results

There are 53 adverse reactions reported during the observation. They are assigned to 14 categories. In the comparison of classification accuracy, the deep learning model shows superiority over the LASSO and kNN model with a rate over 80 %. In the comparison of reliability, the deep learning model shows the best stability among the three models.

Conclusions

Machine learning provides a new method to explore the complex associations among genomic variations and multiple events in pharmacogenomics studies. The new deep learning algorithm is capable of classifying various SNPs to the corresponding adverse reactions. We expect that as more genomic variations are added as features and more observations are made, the deep learning model can improve its performance and can act as a black-box but reliable verifier for other GWAS studies.

     

Predicting enhancers with deep convolutional neural networks

Background

With the rapid development of deep sequencing techniques in the recent years, enhancers have been systematically identified in such projects as FANTOM and ENCODE, forming genome-wide landscapes in a series of human cell lines. Nevertheless, experimental approaches are still costly and time consuming for large scale identification of enhancers across a variety of tissues under different disease status, making computational identification of enhancers indispensable.

Results

To facilitate the identification of enhancers, we propose a computational framework, named DeepEnhancer, to distinguish enhancers from background genomic sequences. Our method purely relies on DNA sequences to predict enhancers in an end-to-end manner by using a deep convolutional neural network (CNN). We train our deep learning model on permissive enhancers and then adopt a transfer learning strategy to fine-tune the model on enhancers specific to a cell line. Results demonstrate the effectiveness and efficiency of our method in the classification of enhancers against random sequences, exhibiting advantages of deep learning over traditional sequence-based classifiers. We then construct a variety of neural networks with different architectures and show the usefulness of such techniques as max-pooling and batch normalization in our method. To gain the interpretability of our approach, we further visualize convolutional kernels as sequence logos and successfully identify similar motifs in the JASPAR database.

Conclusions

DeepEnhancer enables the identification of novel enhancers using only DNA sequences via a highly accurate deep learning model. The proposed computational framework can also be applied to similar problems, thereby prompting the use of machine learning methods in life sciences.

     

Large-scale prediction of protein ubiquitination sites using a multimodal deep architecture

Background

Ubiquitination, which is also called “lysine ubiquitination”, occurs when an ubiquitin is attached to lysine (K) residues in targeting proteins. As one of the most important post translational modifications (PTMs), it plays the significant role not only in protein degradation, but also in other cellular functions. Thus, systematic anatomy of the ubiquitination proteome is an appealing and challenging research topic. The existing methods for identifying protein ubiquitination sites can be divided into two kinds: mass spectrometry and computational methods. Mass spectrometry-based experimental methods can discover ubiquitination sites from eukaryotes, but are time-consuming and expensive. Therefore, it is priority to develop computational approaches that can effectively and accurately identify protein ubiquitination sites.

Results

The existing computational methods usually require feature engineering, which may lead to redundancy and biased representations. While deep learning is able to excavate underlying characteristics from large-scale training data via multiple-layer networks and non-linear mapping operations. In this paper, we proposed a deep architecture within multiple modalities to identify the ubiquitination sites. First, according to prior knowledge and biological knowledge, we encoded protein sequence fragments around candidate ubiquitination sites into three modalities, namely raw protein sequence fragments, physico-chemical properties and sequence profiles, and designed different deep network layers to extract the hidden representations from them. Then, the generative deep representations corresponding to three modalities were merged to build the final model. We performed our algorithm on the available largest scale protein ubiquitination sites database PLMD, and achieved 66.4% specificity, 66.7% sensitivity, 66.43% accuracy, and 0.221 MCC value. A number of comparative experiments also indicated that our multimodal deep architecture outperformed several popular protein ubiquitination site prediction tools.

Conclusion

The results of comparative experiments validated the effectiveness of our deep network and also displayed that our method outperformed several popular protein ubiquitination site prediction tools. The source codes of our proposed method are available at https://github.com/jiagenlee/deepUbiquitylation.

     

Lewis Carroll's Doublets Net of English Words: Network Heterogeneity in a Complex System

Lewis Carroll''s English word game Doublets is represented as a system of networks with each node being an English word and each connectivity edge confirming that its two ending words are equal in letter length, but different by exactly one letter. We show that this system, which we call the Doublets net, constitutes a complex body of linguistic knowledge concerning English word structure that has computable multiscale features. Distributed morphological, phonological and orthographic constraints and the language''s local redundancy are seen at the node level. Phonological communities are seen at the network level. And a balancing act between the language''s global efficiency and redundancy is seen at the system level. We develop a new measure of intrinsic node-to-node distance and a computational algorithm, called community geometry, which reveal the implicit multiscale structure within binary networks. Because the Doublets net is a modular complex cognitive system, the community geometry and computable multi-scale structural information may provide a foundation for understanding computational learning in many systems whose network structure has yet to be fully analyzed.     

BCDForest: a boosting cascade deep forest model towards the classification of cancer subtypes based on gene expression data

Background

The classification of cancer subtypes is of great importance to cancer disease diagnosis and therapy. Many supervised learning approaches have been applied to cancer subtype classification in the past few years, especially of deep learning based approaches. Recently, the deep forest model has been proposed as an alternative of deep neural networks to learn hyper-representations by using cascade ensemble decision trees. It has been proved that the deep forest model has competitive or even better performance than deep neural networks in some extent. However, the standard deep forest model may face overfitting and ensemble diversity challenges when dealing with small sample size and high-dimensional biology data.

Results

In this paper, we propose a deep learning model, so-called BCDForest, to address cancer subtype classification on small-scale biology datasets, which can be viewed as a modification of the standard deep forest model. The BCDForest distinguishes from the standard deep forest model with the following two main contributions: First, a named multi-class-grained scanning method is proposed to train multiple binary classifiers to encourage diversity of ensemble. Meanwhile, the fitting quality of each classifier is considered in representation learning. Second, we propose a boosting strategy to emphasize more important features in cascade forests, thus to propagate the benefits of discriminative features among cascade layers to improve the classification performance. Systematic comparison experiments on both microarray and RNA-Seq gene expression datasets demonstrate that our method consistently outperforms the state-of-the-art methods in application of cancer subtype classification.

Conclusions

The multi-class-grained scanning and boosting strategy in our model provide an effective solution to ease the overfitting challenge and improve the robustness of deep forest model working on small-scale data. Our model provides a useful approach to the classification of cancer subtypes by using deep learning on high-dimensional and small-scale biology data.

     

Cross-modal interaction between visual and olfactory learning in Apis cerana

The power of the small honeybee brain carrying out behavioral and cognitive tasks has been shown repeatedly to be highly impressive. The present study investigates, for the first time, the cross-modal interaction between visual and olfactory learning in Apis cerana. To explore the role and molecular mechanisms of cross-modal learning in A. cerana, the honeybees were trained and tested in a modified Y-maze with seven visual and five olfactory stimulus, where a robust visual threshold for black/white grating (period of 2.8°–3.8°) and relatively olfactory threshold (concentration of 50–25 %) was obtained. Meanwhile, the expression levels of five genes (AcCREB, Acdop1, Acdop2, Acdop3, Actyr1) related to learning and memory were analyzed under different training conditions by real-time RT-PCR. The experimental results indicate that A. cerana could exhibit cross-modal interactions between visual and olfactory learning by reducing the threshold level of the conditioning stimuli, and that these genes may play important roles in the learning process of honeybees.     

Reading between Eye Saccades

Background

Skilled adult readers, in contrast to beginners, show no or little increase in reading latencies as a function of the number of letters in words up to seven letters. The information extraction strategy underlying such efficiency in word identification is still largely unknown, and methods that allow tracking of the letter information extraction through time between eye saccades are needed to fully address this question.

Methodology/Principal Findings

The present study examined the use of letter information during reading, by means of the Bubbles technique. Ten participants each read 5,000 five-letter French words sampled in space-time within a 200 ms window. On the temporal dimension, our results show that two moments are especially important during the information extraction process. On the spatial dimension, we found a bias for the upper half of words. We also show for the first time that letter positions four, one, and three are particularly important for the identification of five-letter words.

Conclusions/Significance

Our findings are consistent with either a partially parallel reading strategy or an optimal serial reading strategy. We show using computer simulations that this serial reading strategy predicts an absence of a word-length effect for words from four- to seven letters in length. We believe that the Bubbles technique will play an important role in further examining the nature of reading between eye saccades.     

Learning Oncogenetic Networks by Reducing to Mixed Integer Linear Programming

Cancer can be a result of accumulation of different types of genetic mutations such as copy number aberrations. The data from tumors are cross-sectional and do not contain the temporal order of the genetic events. Finding the order in which the genetic events have occurred and progression pathways are of vital importance in understanding the disease. In order to model cancer progression, we propose Progression Networks, a special case of Bayesian networks, that are tailored to model disease progression. Progression networks have similarities with Conjunctive Bayesian Networks (CBNs) [1],a variation of Bayesian networks also proposed for modeling disease progression. We also describe a learning algorithm for learning Bayesian networks in general and progression networks in particular. We reduce the hard problem of learning the Bayesian and progression networks to Mixed Integer Linear Programming (MILP). MILP is a Non-deterministic Polynomial-time complete (NP-complete) problem for which very good heuristics exists. We tested our algorithm on synthetic and real cytogenetic data from renal cell carcinoma. We also compared our learned progression networks with the networks proposed in earlier publications. The software is available on the website https://bitbucket.org/farahani/diprog.     

Active Semi-Supervised Learning Method with Hybrid Deep Belief Networks

In this paper, we develop a novel semi-supervised learning algorithm called active hybrid deep belief networks (AHD), to address the semi-supervised sentiment classification problem with deep learning. First, we construct the previous several hidden layers using restricted Boltzmann machines (RBM), which can reduce the dimension and abstract the information of the reviews quickly. Second, we construct the following hidden layers using convolutional restricted Boltzmann machines (CRBM), which can abstract the information of reviews effectively. Third, the constructed deep architecture is fine-tuned by gradient-descent based supervised learning with an exponential loss function. Finally, active learning method is combined based on the proposed deep architecture. We did several experiments on five sentiment classification datasets, and show that AHD is competitive with previous semi-supervised learning algorithm. Experiments are also conducted to verify the effectiveness of our proposed method with different number of labeled reviews and unlabeled reviews respectively.     

Word learning and the cerebral hemispheres: from serial to parallel processing of written words

Reading familiar words differs from reading unfamiliar non-words in two ways. First, word reading is faster and more accurate than reading of unfamiliar non-words. Second, effects of letter length are reduced for words, particularly when they are presented in the right visual field in familiar formats. Two experiments are reported in which right-handed participants read aloud non-words presented briefly in their left and right visual fields before and after training on those items. The non-words were interleaved with familiar words in the naming tests. Before training, naming was slow and error prone, with marked effects of length in both visual fields. After training, fewer errors were made, naming was faster, and the effect of length was much reduced in the right visual field compared with the left. We propose that word learning creates orthographic word forms in the mid-fusiform gyrus of the left cerebral hemisphere. Those word forms allow words to access their phonological and semantic representations on a lexical basis. But orthographic word forms also interact with more posterior letter recognition systems in the middle/inferior occipital gyri, inducing more parallel processing of right visual field words than is possible for any left visual field stimulus, or for unfamiliar non-words presented in the right visual field.     

Epidemiology and Characterization of Seizures in a Pedigreed Baboon Colony

This study evaluated the incidence, prevalence, and clinical features of seizures in a pedigreed captive colony of baboons. The association of seizures with subspecies, age, sex, and various clinical features was assessed. Records for 1527 captive, pedigreed baboons were reviewed, and 3389 events were identified in 1098 baboons. Of these events, 1537 (45%) represented witnessed seizures, whereas the remaining 1852 presented with craniofacial trauma or episodic changes in behavior that were suggestive, but not diagnostic, of seizure activity. Seizures were generalized myoclonic or tonic–clonic, with two thirds of the events witnessed in the morning. Seizure onset occurred in adolescence (age, 5 y), with an average of 3 seizures in a lifetime. The incidence and prevalence of seizures were 2.5% and 26%, respectively, whereas the prevalence of recurrent seizures (that is, epilepsy) was 15%. Seizures were more prevalent in male baboons, which tended to present with earlier onset and more seizures over a lifetime than did female baboons. Seizures were equally distributed between the subspecies; age of onset and seizure recurrences did not differ significantly between subspecies. Clinical features including age of onset, characteristics, and diurnal presentation of seizures in baboons suggested similarities to juvenile myoclonic epilepsy in humans. Facial trauma may be useful marker for epilepsy in baboons, but its specificity should be characterized.The Texas Biomedical Research Institute (Texas Biomed; San Antonio, TX) is home to the Southwest National Primate Research Center, which manages the world''s largest baboon colony, currently comprising about 2500 baboons. Almost 2000 baboons, stretching across 5 to 7 generations and consisting of primarily olive baboons (Papio hamadryas anubis, 64%), yellow baboons (P. h. cynocephalus, 4%), and their hybrids (29%), belong to a pedigreed colony that is widely used for genetic research.²³ Baboons are ideal for the development of genetic models of human disorders due to the many genetic, anatomic, biochemical, and physiologic features shared by humans and baboons.^20,21 Researchers at numerous institutions have used baboons as animal models for a broad range of diseases including diabetes, heart disease, osteoporosis, and chronic infectious illnesses.^12,21Baboons are a natural model of idiopathic generalized epilepsy.⁸ The occurrence of seizures among colony baboons has been noted since the inception of our colony at Texas Biomed more than 50 y ago.⁹ The seizures occur spontaneously or are triggered by ketamine (used for sedation) or other stressors, such as handling or fighting among baboons. Often the seizures are not witnessed, but the baboons are found lying prone on the ground, presumptively having fallen from an elevated structure. These baboons often undergo craniofacial trauma, including periorbital lacerations or bruising, injury to the muzzle or mouth, and broken teeth. Nonetheless, most of these baboons are otherwise healthy, without evidence of developmental delay or focal neurologic deficits. Some baboons with seizures have been reported to be congenitally blind or demonstrate congenital brain damage,⁴ whereas others exhibit seizures as a result of head trauma or infectious diseases. For the most part, however, the baboons with seizures have normal brain anatomy.^8,11,16,19The seizures reported in our baboon colony are typically convulsive, either described as brief generalized myoclonic seizures or tonic–clonic seizures,^17,18 similar to those described in red baboons (P. h. papio)⁸ and humans.³ Previous scalp electroencephalographic studies characterizing epilepsy in the baboon colony demonstrated a high prevalence of generalized interictal epileptic discharges.¹⁷ As is true for P. h. papio, epileptic baboons in our colony are photosensitive (that is, seizures in the animals can be triggered by visual stimuli such as intermittent light stimulation).^9,11,17,18Electroclinical findings suggest that baboons provide an ideal animal model for idiopathic generalized epilepsies in humans. However, little is known about the natural history of epilepsy in baboons. In the current study, we present clinical data regarding the incidence, prevalence, and characteristics of the seizures; their age of onset and tendency to recur; and the effects of epidemiologic factors, including age, sex, and subspecies, on their expression.     

Activation of the Left Inferior Frontal Gyrus in the First 200 ms of Reading: Evidence from Magnetoencephalography (MEG)

Background

It is well established that the left inferior frontal gyrus plays a key role in the cerebral cortical network that supports reading and visual word recognition. Less clear is when in time this contribution begins. We used magnetoencephalography (MEG), which has both good spatial and excellent temporal resolution, to address this question.

Methodology/Principal Findings

MEG data were recorded during a passive viewing paradigm, chosen to emphasize the stimulus-driven component of the cortical response, in which right-handed participants were presented words, consonant strings, and unfamiliar faces to central vision. Time-frequency analyses showed a left-lateralized inferior frontal gyrus (pars opercularis) response to words between 100–250 ms in the beta frequency band that was significantly stronger than the response to consonant strings or faces. The left inferior frontal gyrus response to words peaked at ∼130 ms. This response was significantly later in time than the left middle occipital gyrus, which peaked at ∼115 ms, but not significantly different from the peak response in the left mid fusiform gyrus, which peaked at ∼140 ms, at a location coincident with the fMRI–defined visual word form area (VWFA). Significant responses were also detected to words in other parts of the reading network, including the anterior middle temporal gyrus, the left posterior middle temporal gyrus, the angular and supramarginal gyri, and the left superior temporal gyrus.

Conclusions/Significance

These findings suggest very early interactions between the vision and language domains during visual word recognition, with speech motor areas being activated at the same time as the orthographic word-form is being resolved within the fusiform gyrus. This challenges the conventional view of a temporally serial processing sequence for visual word recognition in which letter forms are initially decoded, interact with their phonological and semantic representations, and only then gain access to a speech code.     

Neural mechanisms of human perceptual learning: electrophysiological evidence for a two-stage process

Background

Humans and other animals change the way they perceive the world due to experience. This process has been labeled as perceptual learning, and implies that adult nervous systems can adaptively modify the way in which they process sensory stimulation. However, the mechanisms by which the brain modifies this capacity have not been sufficiently analyzed.

Methodology/Principal Findings

We studied the neural mechanisms of human perceptual learning by combining electroencephalographic (EEG) recordings of brain activity and the assessment of psychophysical performance during training in a visual search task. All participants improved their perceptual performance as reflected by an increase in sensitivity (d'') and a decrease in reaction time. The EEG signal was acquired throughout the entire experiment revealing amplitude increments, specific and unspecific to the trained stimulus, in event-related potential (ERP) components N2pc and P3 respectively. P3 unspecific modification can be related to context or task-based learning, while N2pc may be reflecting a more specific attentional-related boosting of target detection. Moreover, bell and U-shaped profiles of oscillatory brain activity in gamma (30–60 Hz) and alpha (8–14 Hz) frequency bands may suggest the existence of two phases for learning acquisition, which can be understood as distinctive optimization mechanisms in stimulus processing.

Conclusions/Significance

We conclude that there are reorganizations in several neural processes that contribute differently to perceptual learning in a visual search task. We propose an integrative model of neural activity reorganization, whereby perceptual learning takes place as a two-stage phenomenon including perceptual, attentional and contextual processes.     

Music alters visual perception

Background

Visual perception is not a passive process: in order to efficiently process visual input, the brain actively uses previous knowledge (e.g., memory) and expectations about what the world should look like. However, perception is not only influenced by previous knowledge. Especially the perception of emotional stimuli is influenced by the emotional state of the observer. In other words, how we perceive the world does not only depend on what we know of the world, but also by how we feel. In this study, we further investigated the relation between mood and perception.

Methods and Findings

We let observers do a difficult stimulus detection task, in which they had to detect schematic happy and sad faces embedded in noise. Mood was manipulated by means of music. We found that observers were more accurate in detecting faces congruent with their mood, corroborating earlier research. However, in trials in which no actual face was presented, observers made a significant number of false alarms. The content of these false alarms, or illusory percepts, was strongly influenced by the observers'' mood.

Conclusions

As illusory percepts are believed to reflect the content of internal representations that are employed by the brain during top-down processing of visual input, we conclude that top-down modulation of visual processing is not purely predictive in nature: mood, in this case manipulated by music, may also directly alter the way we perceive the world.     

Analysis of Whole-Brain Resting-State fMRI Data Using Hierarchical Clustering Approach

Background

Previous studies using hierarchical clustering approach to analyze resting-state fMRI data were limited to a few slices or regions-of-interest (ROIs) after substantial data reduction.

Purpose

To develop a framework that can perform voxel-wise hierarchical clustering of whole-brain resting-state fMRI data from a group of subjects.

Materials and Methods

Resting-state fMRI measurements were conducted for 86 adult subjects using a single-shot echo-planar imaging (EPI) technique. After pre-processing and co-registration to a standard template, pair-wise cross-correlation coefficients (CC) were calculated for all voxels inside the brain and translated into absolute Pearson''s distances after imposing a threshold CC≥0.3. The group averages of the Pearson''s distances were then used to perform hierarchical clustering with the developed framework, which entails gray matter masking and an iterative scheme to analyze the dendrogram.

Results

With the hierarchical clustering framework, we identified most of the functional connectivity networks reported previously in the literature, such as the motor, sensory, visual, memory, and the default-mode functional networks (DMN). Furthermore, the DMN and visual system were split into their corresponding hierarchical sub-networks.

Conclusion

It is feasible to use the proposed hierarchical clustering scheme for voxel-wise analysis of whole-brain resting-state fMRI data. The hierarchical clustering result not only confirmed generally the finding in functional connectivity networks identified previously using other data processing techniques, such as ICA, but also revealed directly the hierarchical structure within the functional connectivity networks.     

The Alliance Hypothesis for Human Friendship

Background

Exploration of the cognitive systems underlying human friendship will be advanced by identifying the evolved functions these systems perform. Here we propose that human friendship is caused, in part, by cognitive mechanisms designed to assemble support groups for potential conflicts. We use game theory to identify computations about friends that can increase performance in multi-agent conflicts. This analysis suggests that people would benefit from: 1) ranking friends, 2) hiding friend-ranking, and 3) ranking friends according to their own position in partners'' rankings. These possible tactics motivate the hypotheses that people possess egocentric and allocentric representations of the social world, that people are motivated to conceal this information, and that egocentric friend-ranking is determined by allocentric representations of partners'' friend-rankings (more than others'' traits).

Methodology/Principal Findings

We report results from three studies that confirm predictions derived from the alliance hypothesis. Our main empirical finding, replicated in three studies, was that people''s rankings of their ten closest friends were predicted by their own perceived rank among their partners'' other friends. This relationship remained strong after controlling for a variety of factors such as perceived similarity, familiarity, and benefits.

Conclusions/Significance

Our results suggest that the alliance hypothesis merits further attention as a candidate explanation for human friendship.     

Auditory affective norms for German: testing the influence of depression and anxiety on valence and arousal ratings

Background

The study of emotional speech perception and emotional prosody necessitates stimuli with reliable affective norms. However, ratings may be affected by the participants'' current emotional state as increased anxiety and depression have been shown to yield altered neural responding to emotional stimuli. Therefore, the present study had two aims, first to provide a database of emotional speech stimuli and second to probe the influence of depression and anxiety on the affective ratings.

Methodology/Principal Findings

We selected 120 words from the Leipzig Affective Norms for German database (LANG), which includes visual ratings of positive, negative, and neutral word stimuli. These words were spoken by a male and a female native speaker of German with the respective emotional prosody, creating a total set of 240 auditory emotional stimuli. The recordings were rated again by an independent sample of subjects for valence and arousal, yielding groups of highly arousing negative or positive stimuli and neutral stimuli low in arousal. These ratings were correlated with participants'' emotional state measured with the Depression Anxiety Stress Scales (DASS). Higher depression scores were related to more negative valence of negative and positive, but not neutral words. Anxiety scores correlated with increased arousal and more negative valence of negative words.

Conclusions/Significance

These results underscore the importance of representatively distributed depression and anxiety scores in participants of affective rating studies. The LANG-audition database, which provides well-controlled, short-duration auditory word stimuli for the experimental investigation of emotional speech is available in Supporting Information S1.