An evaluation of the use of multidimensional scaling for understanding brain connectivity

Mapping relational links between motor imagery, action observation, action-related language, and action execution

Actions can be physically executed, observed, imagined, or simply thought about. Unifying mental processes, such as simulation, emulation, or predictive processing, are thought to underlie different action types, whether they are mental states, as in the case of motor imagery and action observation, or involve physical execution. While overlapping brain activity is typically observed across different actions which indicates commonalities, research interest is also concerned with investigating the distinct functional components of these action types. Unfortunately, untangling subtleties associated with the neurocognitive bases of different action types is a complex endeavour due to the high dimensional nature of their neural substrate (e.g., any action process is likely to activate multiple brain regions thereby having multiple dimensions to consider when comparing across them). This has impeded progress in action-related theorising and application. The present study addresses this challenge by using the novel approach of multidimensional modeling to reduce the high-dimensional neural substrate of four action-related behaviours (motor imagery, action observation, action-related language, and action execution), find the least number of dimensions that distinguish or relate these action types, and characterise their neurocognitive relational links. Data for the model comprised brain activations for action types from whole-brain analyses reported in 53 published articles. Eighty-two dimensions (i.e., 82 brain regions) for the action types were reduced to a three-dimensional model, that mapped action types in ordination space where the greater the distance between the action types, the more dissimilar they are. A series of one-way ANOVAs and post-hoc comparisons performed on the mean coordinates for each action type in the model showed that across all action types, action execution and concurrent action observation (AO)-motor imagery (MI) were most neurocognitively similar, while action execution and AO were most dissimilar. Most action types were similar on at least one neurocognitive dimension, the exception to this being action-related language. The import of the findings are discussed in terms of future research and implications for application.

Auditory-prefrontal axonal connectivity in the macaque cortex: quantitative assessment of processing streams

Primate sensory systems subserve complex neurocomputational functions. Consequently, these systems are organised anatomically in a distributed fashion, commonly linking areas to form specialised processing streams. Each stream is related to a specific function, as evidenced from studies of the visual cortex, which features rather prominent segregation into spatial and non-spatial domains. It has been hypothesised that other sensory systems, including auditory, are organised in a similar way on the cortical level. Recent studies offer rich qualitative evidence for the dual stream hypothesis. Here we provide a new paradigm to quantitatively uncover these patterns in the auditory system, based on an analysis of multiple anatomical studies using multivariate techniques. As a test case, we also apply our assessment techniques to more ubiquitously-explored visual system. Importantly, the introduced framework opens the possibility for these techniques to be applied to other neural systems featuring a dichotomised organisation, such as language or music perception.

Seeing through the eyes of anxious individuals: an investigation of anxiety-related interpretations of emotional expressions

The interpretation of emotional states is necessary for successful social communication. Often individuals interpret emotional expressions intuitively and without full cognitive awareness. The aim of the present study was to test whether anxiety would influence affect interpretation in the manner suggested by interpretation bias-the tendency to interpret ambiguous cues in a threatening way. Interpretation of social cues was assessed with the similarity rating task (simtask) in two studies (n1 = 116, n2 = 76). The similarity ratings were analysed with a multidimensional scaling (MDS) approach, and the effects of anxiety on the interpretation of emotional expressions were analysed with multilevel modelling. The results of both studies showed evidence for an anxiety-related interpretation bias. High-anxious individuals tended to interpret milder threats as more negative than low-anxious individuals did. The consequences for anxiety research are discussed.

The history of CoCoMac

CoCoMac, the "Collation of Connectivity Data for the Macaque" is a relational database system which presently constitutes the largest electronic repository of published neuroanatomical connectivity data. Developed since 1996, CoCoMac comprises approximately 40,000 experimental findings on anatomical connections in the macaque brain, as derived from neuroanatomical tract tracing studies. In this historical review, I describe the origin and the history of CoCoMac from a personal perspective, illustrate the principles of its structure and outline the impact it has had on systems neuroscience, in particular as a prelude to the "Human Connectome" research programme.

The intrinsic connectome of the rat amygdala

The connectomes of nervous systems or parts there of are becoming important subjects of study as the amount of connectivity data increases. Because most tract-tracing studies are performed on the rat, we conducted a comprehensive analysis of the amygdala connectome of this species resulting in a meta-study. The data were imported into the neuroVIISAS system, where regions of the connectome are organized in a controlled ontology and network analysis can be performed. A weighted digraph represents the bilateral intrinsic (connections of regions of the amygdala) and extrinsic (connections of regions of the amygdala to non-amygdaloid regions) connectome of the amygdala. Its structure as well as its local and global network parameters depend on the arrangement of neuronal entities in the ontology. The intrinsic amygdala connectome is a small-world and scale-free network. The anterior cortical nucleus (72 in- and out-going edges), the posterior nucleus (45), and the anterior basomedial nucleus (44) are the nuclear regions that posses most in- and outdegrees. The posterior nucleus turns out to be the most important nucleus of the intrinsic amygdala network since its Shapley rate is minimal. Within the intrinsic amygdala, regions were determined that are essential for network integrity. These regions are important for behavioral (processing of emotions and motivation) and functional (memory) performances of the amygdala as reported in other studies.

neuroVIISAS: approaching multiscale simulation of the rat connectome

neuroVIISAS is a generic platform which allows the integration of neuroontologies, mapping functions for brain atlas development, and connectivity data administration; all of which are required for the analysis of structurally and neurobiologically realistic simulations of networks. What makes neuroVIISAS unique is the ability to integrate neuroontologies, image stacks, mappings, visualizations, analyzes and simulations to use them for modelling and simulations. Based on the analysis of over 2020 tracing studies, atlas terminologies and registered histological stacks of images, neuroVIISAS permits the definition of neurobiologically realistic networks that are transferred to the simulation engine NEST. The analysis on a local and global level, the visualization of connectivity data and the results of simulations offer new possibilities to study structural and functional relationships of neural networks. This paper describes the major components and techniques of how to analyse, visualize and simulate with neuroVIISAS shown on a model network at a coarse CNS level (106 regions, 1566 connections) out of 13681 regions and 134043 connections of the left and right part of the CNS. This network of major components of the left and right hemisphere has small-world properties of the Watts-Strogatz model. Furthermore, synchronized subpopulations, oscillations of rate distributions and a time shift of population activities of the left and right hemisphere were observed in the neurocomputational simulations. In summary, a generic platform has been developed that realizes data-analysis-visualization integration for the exploration of network dynamics on multiple levels.

A weighted and directed interareal connectivity matrix for macaque cerebral cortex

Retrograde tracer injections in 29 of the 91 areas of the macaque cerebral cortex revealed 1,615 interareal pathways, a third of which have not previously been reported. A weight index (extrinsic fraction of labeled neurons [FLNe]) was determined for each area-to-area pathway. Newly found projections were weaker on average compared with the known projections; nevertheless, the 2 sets of pathways had extensively overlapping weight distributions. Repeat injections across individuals revealed modest FLNe variability given the range of FLNe values (standard deviation <1 log unit, range 5 log units). The connectivity profile for each area conformed to a lognormal distribution, where a majority of projections are moderate or weak in strength. In the G_{29 × 29} interareal subgraph, two-thirds of the connections that can exist do exist. Analysis of the smallest set of areas that collects links from all 91 nodes of the G_{29 × 91} subgraph (dominating set analysis) confirms the dense (66%) structure of the cortical matrix. The G_{29 × 29} subgraph suggests an unexpectedly high incidence of unidirectional links. The directed and weighted G_{29 × 91} connectivity matrix for the macaque will be valuable for comparison with connectivity analyses in other species, including humans. It will also inform future modeling studies that explore the regularities of cortical networks.

Comparison of Methods for Collecting and Modeling Dissimilarity Data: Applications to Complex Sound Stimuli

Sorting procedures are frequently adopted as an alternative to dissimilarity ratings to measure the dissimilarity of large sets of stimuli in a comparatively short time. However, systematic empirical research on the consequences of this experiment-design choice is lacking. We carried out a behavioral experiment to assess the extent to which sorting procedures compare to dissimilarity ratings in terms of efficiency, reliability, and accuracy, and the extent to which data from different data-collection methods are redundant and are better fit by different distance models. Participants estimated the dissimilarity of either semantically charged environmental sounds or semantically neutral synthetic sounds. We considered free and hierarchical sorting and derived indications concerning the properties of constrained and truncated hierarchical sorting methods from hierarchical sorting data. Results show that the higher efficiency of sorting methods comes at a considerable cost in terms of data reliability and accuracy. This loss appears to be minimized with truncated hierarchical sorting methods that start from a relatively low number of groups of stimuli. Finally, variations in data-collection method differentially affect the fit of various distance models at the group-average and individual levels. On the basis of these results, we suggest adopting sorting as an alternative to dissimilarity-rating methods only when strictly necessary. We also suggest analyzing the raw behavioral dissimilarities, and avoiding modeling them with one single distance model.

Multidimensional scaling of multiplex data: human milk cytokines

The purpose of this study was to use multidimensional scaling (MDS) and cluster-analytic techniques to examine how cytokine levels from a large multiplex assay of human milk samples covary. Milk samples were collected at 4-6 weeks postpartum from 57 women and were assayed by Luminex multiplex technology for 20 cytokines, chemokines, and growth factors. The MDS was applied to a proximity-score matrix based on these values. A three-dimensional (3D) space was sufficient to accommodate the configuration of relationships. Cytokines that covaried in their concentrations were assigned similar coordinates and plotted close together in 3D space. Several clusters of cytokines were identified. Since very little is known about the origins and functions of cytokines in milk, this approach may provide new clues that will guide future explorations of origins and functional relationships of the separate clusters. This analytical tool may provide a new approach to understanding the physiology of milk cytokines and may be generalizable to multiplex data in general.

Convergence of internal and external structure for the California Child Q-set

The language of personality traits includes single-word trait descriptors, and longer phrases or sentences. Evidence has accumulated that abstract, semantic relationships among single words have the same underlying structure as the empirical relationships when words are applied to individuals. The present study examines whether these two kinds of structure are also isomorphic for longer trait descriptors. Empirical descriptions and judgements of semantic similarity were collected among the descriptors comprising the California Child Q-set, or CCQ, and analysed with multidimensional scaling. Canonical correlation showed the solutions to be closely related to one another, and to independent sets of ratings available for the CCQ items. Informants' similarity judgements were not affected by the context in which they were made. The dominant dimensions of the solutions reproduce dimensions found previously for the single-word personality lexicon, indicating the two trait-descriptive languages to be closely parallel.

A multidimensional analysis of the epistemic origins of nursing theories, models, and frameworks

The purpose of this article is to introduce our notion of epistemic space and to demonstrate its utility for understanding the origins and trajectories of nursing theory in the 20th century using multidimensional scaling (MDS). A literature review was conducted on primary and secondary sources written by and about 20 nurse theorists to investigate whether or not they cited 129 different scholars in the fields of anthropology, biology, nursing, philosophy, psychology, and sociology. Seventy-four scholars were identified as having been cited by at least two nurse theorists (319 citations total). Proximity scores, quantifying the similarity among nurse theorists based on proportions of shared citations, were calculated and analyzed using MDS. The emergent model of epistemic space that accommodated these similarities among nurse theorists revealed the systematic influence of scholars from various fields, notably psychology, biology, and philosophy. We believe that this schema and resulting taxonomy will prove useful for furthering our understanding of the relationships among nursing theories and theories in other fields of science.

Multidimensional Scaling Methods for Many-Object Sets: A Review

Given a set of dissimilarities data between n objects, multidimensional scaling is the problem of reconstructing a geometrical pattern of these objects, using n points, so that between-points distance corresponds to between-objects dissimilarity. Often, the collection of input data requires rating the dissimilarities between all n(n - 1)/2 possible pairs of stimuli. When the number of stimuli is large, say n $ 30, the number of pairs to be compared becomes very large and the similarity task inefficient. Hence a question of major importance is how to increase the efficiency of the similarity task while maintaining satisfactory scaling solutions. This article reviews the main similarity task methods suitable for a large objects set.

Analysis of the connectional organization of neural systems associated with the hippocampus in rats

The hippocampus of the rat enjoys a central significance for researchers interested in the neural mechanisms of memory and spatial information processing. Many of the theoretical models advanced to explain function in this system, however, do not reflect the wealth of information on the connectivity of these structures, and employ greatly simplified treatments of its complex connectivity. We were interested in whether a more analytical approach, which begins with analysis of the connectivity of the system, might provide insights complementary to those derived by synthetic models. Accordingly, we collated detailed neuroanatomical information about the connectivity of the hippocampal system in the rat, and analysed the resulting data. Analyses of connectivity based on a variety of different analytical techniques have recently been used to elucidate the global organization of other systems in the macaque and cat, and have given rise to successful predictions. We applied non-metric multidimensional scaling and non-parametric cluster analysis to our summary matrix of connection data. The analyses produced organizational schemes that were consistent with known physiological properties and provided the basis for making tentative predictions of the further structures that may contain 'place' and 'head-direction' cells, which structures we identify. The consistency between the analyses of connectivity and the distribution of physiological properties across the system suggests that functional relationships are constrained by the organization of the connectivity of the system, and so that structure and function are linked at the systems level.

Anatomical connectivity defines the organization of clusters of cortical areas in the macaque monkey and the cat

The number of different cortical structures in mammalian brains and the number of extrinsic fibres linking these regions are both large. As with any complex system, systematic analysis is required to draw reliable conclusions about the organization of the complex neural networks comprising these numerous elements. One aspect of organization that has long been suspected is that cortical networks are organized into 'streams' or 'systems'. Here we report computational analyses capable of showing whether clusters of strongly interconnected areas are aspects of the global organization of cortical systems in macaque and cat. We used two different approaches to analyse compilations of corticocortical connection data from the macaque and the cat. The first approach, optimal set analysis, employed an explicit definition of a neural 'system' or 'stream', which was based on differential connectivity. We defined a two-component cost function that described the cost of the global cluster arrangement of areas in terms of the areas' connectivity within and between candidate clusters. Optimal cluster arrangements of cortical areas were then selected computationally from the very many possible arrangements, using an evolutionary optimization algorithm. The second approach, non-parametric cluster analysis (NPCA), grouped cortical areas on the basis of their proximity in multidimensional scaling representations. We used non-metric multidimensional scaling to represent the cortical connectivity structures metrically in two and five dimensions. NPCA then analysed these representations to determine the nature of the clusters for a wide range of different cluster shape parameters. The results from both approaches largely agreed. They showed that macaque and cat cortices are organized into densely intra-connected clusters of areas, and identified the constituent members of the clusters. These clusters reflected functionally specialized sets of cortical areas, suggesting that structure and function are closely linked at this gross, systems level.