Semantics of Multimodal Network Models

A multimodal network (MMN) is a novel graph-theoretic formalism designed to capture the structure of biological networks and to represent relationships derived from multiple biological databases. MMNs generalize the standard notions of graphs and hypergraphs, which are the bases of current diagrammatic representations of biological phenomena, and incorporate the concept of mode. Each vertex of an MMN is a biological entity, a biot, while each modal hyperedge is a typed relationship, where the type is given by the mode of the hyperedge. The semantics of each modal hyperedge e is given through denotational semantics, where a valuation function f_{e} defines the relationship among the values of the vertices incident on e. The meaning of an MMN is denoted in terms of the semantics of a hyperedge sequence. A companion paper defines MMNs and concentrates on the structural aspects of MMNs. This paper develops MMN denotational semantics when used as a representation of the semantics of biological networks and discusses applications of MMNs in managing complex biological data.     

Mismatch negativity in school-age children to speech stimuli that are just perceptibly different

The mismatch negativity event-related potential (MMN) was elicited in normal school-age children in response to just perceptibly different variants of the speech phoneme /da/. A significant MMN was measured in each subject tested. Child and adult MMNs were similar with respect to peak latency and duration. Measures of MMN magnitude (peak-to-peak amplitude and area) were significantly larger in children than in adults. The results of the present study indicate that the MMN can be elicited in response to minimal acoustic stimulus differences in complex speech signals in school-age children. The results support the feasibility of using the MMN as a tool in the study of deficient auditory perception in children.     

Integration of consonant and pitch processing as revealed by the absence of additivity in mismatch negativity

Consonants, unlike vowels, are thought to be speech specific and therefore no interactions would be expected between consonants and pitch, a basic element for musical tones. The present study used an electrophysiological approach to investigate whether, contrary to this view, there is integrative processing of consonants and pitch by measuring additivity of changes in the mismatch negativity (MMN) of evoked potentials. The MMN is elicited by discriminable variations occurring in a sequence of repetitive, homogeneous sounds. In the experiment, event-related potentials (ERPs) were recorded while participants heard frequently sung consonant-vowel syllables and rare stimuli deviating in either consonant identity only, pitch only, or in both dimensions. Every type of deviation elicited a reliable MMN. As expected, the two single-deviant MMNs had similar amplitudes, but that of the double-deviant MMN was also not significantly different from them. This absence of additivity in the double-deviant MMN suggests that consonant and pitch variations are processed, at least at a pre-attentive level, in an integrated rather than independent way. Domain-specificity of consonants may depend on higher-level processes in the hierarchy of speech perception.     

Differences in Mismatch Responses to Vowels and Musical Intervals: MEG Evidence

We investigated the electrophysiological response to matched two-formant vowels and two-note musical intervals, with the goal of examining whether music is processed differently from language in early cortical responses. Using magnetoencephalography (MEG), we compared the mismatch-response (MMN/MMF, an early, pre-attentive difference-detector occurring approximately 200 ms post-onset) to musical intervals and vowels composed of matched frequencies. Participants heard blocks of two stimuli in a passive oddball paradigm in one of three conditions: sine waves, piano tones and vowels. In each condition, participants heard two-formant vowels or musical intervals whose frequencies were 11, 12, or 24 semitones apart. In music, 12 semitones and 24 semitones are perceived as highly similar intervals (one and two octaves, respectively), while in speech 12 semitones and 11 semitones formant separations are perceived as highly similar (both variants of the vowel in ‘cut’). Our results indicate that the MMN response mirrors the perceptual one: larger MMNs were elicited for the 12–11 pairing in the music conditions than in the language condition; conversely, larger MMNs were elicited to the 12–24 pairing in the language condition that in the music conditions, suggesting that within 250 ms of hearing complex auditory stimuli, the neural computation of similarity, just as the behavioral one, differs significantly depending on whether the context is music or speech.     

A sparse neural code for some speech sounds but not for others

The precise neural mechanisms underlying speech sound representations are still a matter of debate. Proponents of 'sparse representations' assume that on the level of speech sounds, only contrastive or otherwise not predictable information is stored in long-term memory. Here, in a passive oddball paradigm, we challenge the neural foundations of such a 'sparse' representation; we use words that differ only in their penultimate consonant ("coronal" [t] vs. "dorsal" [k] place of articulation) and for example distinguish between the German nouns Latz ([lats]; bib) and Lachs ([laks]; salmon). Changes from standard [t] to deviant [k] and vice versa elicited a discernible Mismatch Negativity (MMN) response. Crucially, however, the MMN for the deviant [lats] was stronger than the MMN for the deviant [laks]. Source localization showed this difference to be due to enhanced brain activity in right superior temporal cortex. These findings reflect a difference in phonological 'sparsity': Coronal [t] segments, but not dorsal [k] segments, are based on more sparse representations and elicit less specific neural predictions; sensory deviations from this prediction are more readily 'tolerated' and accordingly trigger weaker MMNs. The results foster the neurocomputational reality of 'representationally sparse' models of speech perception that are compatible with more general predictive mechanisms in auditory perception.     

Representation and Processing of Lexical Tone and Tonal Variants: Evidence from the Mismatch Negativity

Pronunciation variation is ubiquitous in the speech signal. Different models of lexical representation have been put forward to deal with speech variability, which differ in the level as well as the nature of mental representation. We present the first mismatch negativity (MMN) study investigating the effect of allophonic variation on the mental representation and neural processing of lexical tones. Native speakers of Standard Chinese (SC) participated in an oddball electroencephalography (EEG) experiment. All stimuli have the same segments (ma) but different lexical tones: level [T1], rising [T2], and dipping [T3]. In connected speech with a T3T3 sequence, the first T3 may undergo allophonic change and is produced with a rising pitch contour (T3V), similar to the lexical T2 pitch contour. Four oddball conditions were constructed (T1/T3, T3/T1, T2/T3, T3/T2; standard/deviant). All four conditions elicited MMN effects, with the T1–T3 pair eliciting comparable MMNs, but the T2–T3 pair asymmetrical MMN effects. There were significantly greater and earlier MMN effects in the T2/T3 condition than that in the reversed T3/T2 condition. Furthermore, the T3/T2 condition showed more rightward MMN effects than the T2/T3 condition and the T1–T3 pair. Such asymmetries suggest co-activation of long-term memory representations of both T3 and T3V when T3 serves as the standard. The acoustic similarity between the activated T3V (by the standard T3) and the incoming deviant stimulus T2 induces acoustic processing of the tonal contrast in the T3/T2 condition, similar to that of within-category lexical tone processing, which is in contrast to the processing of between-category lexical tones observed in the T2/T3, T1/T3, and T3/T1 conditions.     

Effects of Acoustic Context on the Perceptual Differences between Spatial Sound Stimuli

We studied the effects of the acoustic context on active and passive discrimination of moving sound signals. Different contexts were created by reversing the role of standard and deviant stimuli in the oddball blocks, while their acoustical features were kept the same. Three types of sounds were used as standard or deviant stimuli in different blocks: stationary midline noises and two (smooth and abrupt) moving sounds moving to the left or right of the midline. Auditory event-related potentials (ERPs) were recorded during passive listening (the sound stimulation ignored), and mismatch negativity potentials (MMNs) were obtained. Active discrimination of sound movements was measured by the hit rate (percent correct responses) and false alarm rate, as well as the reaction time. The influence of the stimulus context on active and passive discrimination of the moving sound stimuli was reflected in the phenomenon known as the effect of deviance direction. The hit rate and MMN amplitude were higher when the deviant moved faster than the standard. The MMN amplitude was more responsive to the velocity of sound stimuli than the hit rate and false alarm rate. The psychophysical measurements in the reversed contexts suggest that smooth and abrupt sound movements may belong to the same perceptual category (moving sounds), while the stationary stimuli form another perceptual category.     

A visualization method for multi-relation in dataset

To express the multi-relation in complex dataset, proposed a fast visualization method based on the continuous Catmull–Rom curve. The method expresses the entity data and the taxonomic relation in dataset using the hypergraph. The nodes are designed to represent the entity data in dataset. The hyperedge is introduced to link all the relative nodes in one multi-relation. The hyperedge can be visualized by the curve style or the regional style according to different requirements. The nodes in one hyperedge are linked by a continuous curve in curve style, while all the nodes are surrounded by one closed region in regional style. The Catmull–Rom algorithm is adopted to produce the continuous curve between the adjacent nodes in curve style. The nodes are regarded as the control points in the interpolating processing of Catmull–Rom algorithm. The peripheral points are obtained by extending along the tendency line of the hyperedge in regional style. The closed curve connecting these peripheral points forms a closed region expressing the hyperedge. The color selected to stain the hyperedge based on color wheel. The selected colors for each hyperedge can maximize the visual differentiation in the processing of coloring each hyperedge. The experimental results denoted that the method can achieve the intuitive and accurate expression for the multi-relation in complex dataset. The method can visualize the common scale dataset the in real time.     

Theory of interface: Category theory,directed networks and evolution of biological networks

Biological networks have two modes. The first mode is static: a network is a passage on which something flows. The second mode is dynamic: a network is a pattern constructed by gluing functions of entities constituting the network. In this paper, first we discuss that these two modes can be associated with the category theoretic duality (adjunction) and derive a natural network structure (a path notion) for each mode by appealing to the category theoretic universality. The path notion corresponding to the static mode is just the usual directed path. The path notion for the dynamic mode is called lateral path which is the alternating path considered on the set of arcs. Their general functionalities in a network are transport and coherence, respectively. Second, we introduce a betweenness centrality of arcs for each mode and see how the two modes are embedded in various real biological network data. We find that there is a trade-off relationship between the two centralities: if the value of one is large then the value of the other is small. This can be seen as a kind of division of labor in a network into transport on the network and coherence of the network. Finally, we propose an optimization model of networks based on a quality function involving intensities of the two modes in order to see how networks with the above trade-off relationship can emerge through evolution. We show that the trade-off relationship can be observed in the evolved networks only when the dynamic mode is dominant in the quality function by numerical simulations. We also show that the evolved networks have features qualitatively similar to real biological networks by standard complex network analysis.     

Uncovering the Molecular Mechanism of Actions between Pharmaceuticals and Proteins on the AD Network

This study begins with constructing the mini metabolic networks (MMNs) of beta amyloid (Aβ) and acetylcholine (ACh) which stimulate the Alzheimer’s Disease (AD). Then we generate the AD network by incorporating MMNs of Aβ and ACh, and other MMNs of stimuli of AD. The panel of proteins contains 49 enzymes/receptors on the AD network which have the 3D-structure in PDB. The panel of drugs is formed by 5 AD drugs and 5 AD nutraceutical drugs, and 20 non-AD drugs. All of these complexes formed by these 30 drugs and 49 proteins are transformed into dyadic arrays. Utilizing the prior knowledge learned from the drug panel, we propose a statistical classification (dry-lab). According to the wet-lab for the complex of amiloride and insulin degrading enzyme, and the complex of amiloride and neutral endopeptidase, we are confident that this dry-lab is reliable. As the consequences of the dry-lab, we discover many interesting implications. Especially, we show that possible causes of Tacrine, donepezil, galantamine and huperzine A cannot improve the level of ACh which is against to their original design purpose but they still prevent AD to be worse as Aβ deposition appeared. On the other hand, we recommend Miglitol and Atenolol as the safe and potent drugs to improve the level of ACh before Aβ deposition appearing. Moreover, some nutrients such as NADH and Vitamin E should be controlled because they may harm health if being used in wrong way and wrong time. Anyway, the insights shown in this study are valuable to be developed further.     

Analyzing molecular reaction networks

Pathways are typically the central concept in the analysis of biochemical reaction networks. A pathway can be interpreted as a chain of enzymatical reactions performing a specific biological function. A common way to study metabolic networks are minimal pathways that can operate at steady state called elementary modes. The theory of chemical organizations has recently been used to decompose biochemical networks into algebraically closed and self-maintaining subnetworks termed organizations. The aim of this paper is to elucidate the relation between these two concepts. Whereas elementary modes represent the boundaries of the potential behavior of the network, organizations define metabolite compositions that are likely to be present in biological feasible situations. Hence, steady state organizations consist of combinations of elementary modes. On the other hand, it is possible to assign a unique (and possibly empty) set of organizations to each elementary mode, indicating the metabolites accompanying the active pathway in a feasible steady state.     

Combination therapeutics in complex diseases

The biological redundancies in molecular networks of complex diseases limit the efficacy of many single drug therapies. Combination therapeutics, as a common therapeutic method, involve pharmacological intervention using several drugs that interact with multiple targets in the molecular networks of diseases and may achieve better efficacy and/or less toxicity than monotherapy in practice. The development of combination therapeutics is complicated by several critical issues, including identifying multiple targets, targeting strategies and the drug combination. This review summarizes the current achievements in combination therapeutics, with a particular emphasis on the efforts to develop combination therapeutics for complex diseases.     

Evidence for the late MMN as a neurophysiological endophenotype for dyslexia

Dyslexia affects 5-10% of school-aged children and is therefore one of the most common learning disorders. Research on auditory event related potentials (AERP), particularly the mismatch negativity (MMN) component, has revealed anomalies in individuals with dyslexia to speech stimuli. Furthermore, candidate genes for this disorder were found through molecular genetic studies. A current challenge for dyslexia research is to understand the interaction between molecular genetics and brain function, and to promote the identification of relevant endophenotypes for dyslexia. The present study examines MMN, a neurophysiological correlate of speech perception, and its potential as an endophenotype for dyslexia in three groups of children. The first group of children was clinically diagnosed with dyslexia, whereas the second group of children was comprised of their siblings who had average reading and spelling skills and were therefore "unaffected" despite having a genetic risk for dyslexia. The third group consisted of control children who were not related to the other groups and were also unaffected. In total, 225 children were included in the study. All children showed clear MMN activity to/da/-/ba/contrasts that could be separated into three distinct MMN components. Whilst the first two MMN components did not differentiate the groups, the late MMN component (300-700 ms) revealed significant group differences. The mean area of the late MMN was attenuated in both the dyslexic children and their unaffected siblings in comparison to the control children. This finding is indicative of analogous alterations of neurophysiological processes in children with dyslexia and those with a genetic risk for dyslexia, without a manifestation of the disorder. The present results therefore further suggest that the late MMN might be a potential endophenotype for dyslexia.     

A Parallel Implementation of the Network Identification by Multiple Regression (NIR) Algorithm to Reverse-Engineer Regulatory Gene Networks

The reverse engineering of gene regulatory networks using gene expression profile data has become crucial to gain novel biological knowledge. Large amounts of data that need to be analyzed are currently being produced due to advances in microarray technologies. Using current reverse engineering algorithms to analyze large data sets can be very computational-intensive. These emerging computational requirements can be met using parallel computing techniques. It has been shown that the Network Identification by multiple Regression (NIR) algorithm performs better than the other ready-to-use reverse engineering software. However it cannot be used with large networks with thousands of nodes - as is the case in biological networks - due to the high time and space complexity. In this work we overcome this limitation by designing and developing a parallel version of the NIR algorithm. The new implementation of the algorithm reaches a very good accuracy even for large gene networks, improving our understanding of the gene regulatory networks that is crucial for a wide range of biomedical applications.     

Differentiation of schizophrenia patients from healthy subjects by mismatch negativity and neuropsychological tests

Background

Schizophrenia is a heterogeneous disorder with diverse presentations. The current and the proposed DSM-V diagnostic system remains phenomenologically based, despite the fact that several neurobiological and neuropsychological markers have been identified. A multivariate approach has better diagnostic utility than a single marker method. In this study, the mismatch negativity (MMN) deficit of schizophrenia was first replicated in a Han Chinese population, and then the MMN was combined with several neuropsychological measurements to differentiate schizophrenia patients from healthy subjects.

Methodology/Principal Findings

120 schizophrenia patients and 76 healthy controls were recruited. Each subject received examinations for duration MMN, Continuous Performance Test, Wisconsin Card Sorting Test, and Wechsler Adult Intelligence Scale Third Edition (WAIS-III). The MMN was compared between cases and controls, and important covariates were investigated. Schizophrenia patients had significantly reduced MMN amplitudes, and MMN decreased with increasing age in both patient and control groups. None of the neuropsychological indices correlated with MMN. Predictive multivariate logistic regression models using the MMN and neuropsychological measurements as predictors were developed. Four predictors, including MMN at electrode FCz and three scores from the WAIS-III (Arithmetic, Block Design, and Performance IQ) were retained in the final predictive model. The model performed well in differentiating patients from healthy subjects (percentage of concordant pairs: 90.5%).

Conclusions/Significance

MMN deficits were found in Han Chinese schizophrenia patients. The multivariate approach combining biomarkers from different modalities such as electrophysiology and neuropsychology had a better diagnostic utility.     

The role and mechanism-of-action of Sema3E and Plexin-D1 in vascular and neural development

Class 3 secreted semaphorins (Sema3A–3G) participate in many aspects of axon guidance through holoreceptor complexes that include Neuropilin-1 (Npn-1) or Neuropilin-2 and one of the four class A plexin proteins. However, unlike other Sema3 family proteins, Sema3E directly binds to Plexin-D1 without neuropilins. Its biological function was first explored in intersomitic vessel formation and since its initial discovery, Sema3E–Plexin-D1 signaling has been found to participate in the many biological systems in addition to vascular development, via seemingly different mode of actions. For example, temporal and spatial control of ligand vs. receptor results in two different mechanisms governing vascular patterning. Interactions with other transmembrane proteins such as neuropilin and VEGFR2 result in different axonal behaviors. Ligand receptor localization on pre- vs. post-synaptic neurons is used to control different types of synapse formation. Perhaps different downstream effectors will also result in different functional outcomes. Given the limited number of ligands and receptors in the genome and their multifunctional nature, we expect that more modes of action will be discovered in the future. In this review, we highlight current advances on the mechanisms of how Sema3E–Plexin-D1 interaction shapes the networks of multiple biological systems, in particular the vascular and nervous systems.     

Dynamical roles of biological regulatory circuits

Regulatory circuits are found at the basis of all non-trivial dynamical properties of biological networks. More specifically, positive circuits are involved in the generation of multiple differentiated states, whereas negative circuits can generate cyclic or homeostatic behaviours. These notions are briefly reviewed, from initial biological formulations to mathematical formalisations, further encompassing their application to the design of synthetic regulatory systems. Finally, current challenges for the analysis of increasingly complex regulatory networks are indicated, as well as prospects for our understanding of development and evolution.     

Frequent Pattern Discovery in Multiple Biological Networks: Patterns and Algorithms

The rapid accumulation of biological network data is creating an urgent need for computational methods capable of integrative network analysis. This paper discusses a suite of algorithms that we have developed to discover biologically significant patterns that appear frequently in multiple biological networks: coherent dense subgraphs, frequent dense vertex-sets, generic frequent subgraphs, differential subgraphs, and recurrent heavy subgraphs. We demonstrate these methods on gene co-expression networks, using the identified patterns to systematically annotate gene functions, map genome to phenome, and perform high-order cooperativity analysis.     

Oscillatory Underpinnings of Mismatch Negativity and Their Relationship with Cognitive Function in Patients with Schizophrenia

Background

Impairments in mismatch negativity (MMN) generation have been consistently reported in patients with schizophrenia. However, underlying oscillatory activity of MMN deficits in schizophrenia and the relationship with cognitive impairments have not been investigated in detail. Time-frequency power and phase analyses can provide more detailed measures of brain dynamics of MMN deficits in schizophrenia.

Method

21 patients with schizophrenia and 21 healthy controls were tested with a roving frequency paradigm to generate MMN. Time-frequency domain power and phase-locking (PL) analysis was performed on all trials using short-time Fourier transforms with Hanning window tapering. A comprehensive battery (CANTAB) was used to assess neurocognitive functioning.

Results

Mean MMN amplitude was significantly lower in patients with schizophrenia (95% CI 0.18 - 0.77). Patients showed significantly lower EEG power (95% CI -1.02 - -0.014) in the ~4-7 Hz frequency range (theta band) between 170 and 210 ms. Patients with schizophrenia showed cognitive impairment in multiple domains of CANTAB. However, MMN impairments in amplitude and power were not correlated with clinical measures, medication dose, social functioning or neurocognitive performance.

Conclusion

The findings from this study suggested that while MMN may be a useful marker to probe NMDA receptor mediated mechanisms and associated impairments in gain control and perceptual changes, it may not be a useful marker in association with clinical or cognitive changes. Trial-by-trial EEG power analysis can be used as a measure of brain dynamics underlying MMN deficits which also can have implications for the use of MMN as a biomarker for drug discovery.     

Processing of complex auditory patterns in musicians and nonmusicians

In the present study we investigated the capacity of the memory store underlying the mismatch negativity (MMN) response in musicians and nonmusicians for complex tone patterns. While previous studies have focused either on the kind of information that can be encoded or on the decay of the memory trace over time, we studied capacity in terms of the length of tone sequences, i.e., the number of individual tones that can be fully encoded and maintained. By means of magnetoencephalography (MEG) we recorded MMN responses to deviant tones that could occur at any position of standard tone patterns composed of four, six or eight tones during passive, distracted listening. Whereas there was a reliable MMN response to deviant tones in the four-tone pattern in both musicians and nonmusicians, only some individuals showed MMN responses to the longer patterns. This finding of a reliable capacity of the short-term auditory store underlying the MMN response is in line with estimates of a three to five item capacity of the short-term memory trace from behavioural studies, although pitch and contour complexity covaried with sequence length, which might have led to an understatement of the reported capacity. Whereas there was a tendency for an enhancement of the pattern MMN in musicians compared to nonmusicians, a strong advantage for musicians could be shown in an accompanying behavioural task of detecting the deviants while attending to the stimuli for all pattern lengths, indicating that long-term musical training differentially affects the memory capacity of auditory short-term memory for complex tone patterns with and without attention. Also, a left-hemispheric lateralization of MMN responses in the six-tone pattern suggests that additional networks that help structuring the patterns in the temporal domain might be recruited for demanding auditory processing in the pitch domain.