Search

To study the human mind is to consider the nature of associations-how are they learned, what are their constituent parts, and how can they be severed or adjusted? The manipulation of associations stands as a pillar of statistical learning (SL) research, which strongly suggests that processes as diverse as word segmentation, learning of grammatical patterns, and event perception can be explained by the learner's sensitivity to simple temporal dependencies (among other regularities). Used to determine the edges of a network, associations are similarly crucial to consider when quantifying the graph-theoretical properties of various cognitive systems. With this point of convergence in mind, the present work reaffirms the unique value of network science in illuminating the broad-level architectures of complex cognitive systems. However, I also describe how insights from the SL literature, coupled with insights from psycholinguistics more broadly, offer a strong theoretical backbone upon which we can develop and study networks that reflect, as closely as possible, the psychological realities of learning.

Cross-linguistic similarity is a term so broad and multi-faceted that it is not easily defined. The degree of overlap between languages is known to affect lexical competition during online processing and production, and its relevance for second language acquisition has also been established. Nevertheless, determining what makes two languages similar (or not) increases in complexity when multiple levels of the language hierarchy (e.g., phonology, syntax) are considered. How can we feasibly account for the patterns of convergence and divergence at each level of representation, as well as the interactions between them? The growing field of network science brings new methodologies to bear on this longstanding question. Below, we summarize current network science approaches to modeling language structure and discuss implications for understanding various linguistic processes. Critically, we stress the particular value of multilayer techniques, unique and powerful in their ability to simultaneously accommodate an array of node-to-node (or word-to-word) relationships.

Human learners are adept at grasping the complex relationships underlying incoming sequential input. In the present work, we formalize complex relationships as graph structures derived from temporal associations in motor sequences. Next, we explore the extent to which learners are sensitive to key variations in the topological properties inherent to those graph structures. Participants performed a probabilistic motor sequence task in which the order of button presses was determined by the traversal of graphs with modular, lattice-like, or random organization. Graph nodes each represented a unique button press and edges represented a transition between button presses. Results indicate that learning, indexed here by participants' response times, was strongly mediated by the graph's meso-scale organization, with modular graphs being associated with shorter response times than random and lattice graphs. Moreover, variations in a node's number of connections (degree) and a node's role in mediating long-distance communication (betweenness centrality) impacted graph learning, even after accounting for level of practice on that node. These results demonstrate that the graph architecture underlying temporal sequences of stimuli fundamentally constrains learning, and moreover that tools from network science provide a valuable framework for assessing how learners encode complex, temporally structured information., 29 pages, 4 figures

Behavioral evidence has shown that humans automatically develop internal representations adapted to the temporal and spatial statistics of the environment. Building on prior fMRI studies that have focused on statistical learning of temporal sequences, we investigated the neural substrates and mechanisms underlying statistical learning from scenes with a structured spatial layout. Our goals were twofold: (1) to determine discrete brain regions in which degree of learning (i.e., behavioral performance) was a significant predictor of neural activity during acquisition of spatial regularities and (2) to examine how connectivity between this set of areas and the rest of the brain changed over the course of learning. Univariate activity analyses indicated a diffuse set of dorsal striatal and occipitoparietal activations correlated with individual differences in participants' ability to acquire the underlying spatial structure of the scenes. In addition, bilateral medial-temporal activation was linked to participants' behavioral performance, suggesting that spatial statistical learning recruits additional resources from the limbic system. Connectivity analyses examined, across the time course of learning, psychophysiological interactions with peak regions defined by the initial univariate analysis. Generally, we find that task-based connectivity with these regions was significantly greater in early relative to later periods of learning. Moreover, in certain cases, decreased task-based connectivity between time points was predicted by overall posttest performance. Results suggest a narrowing mechanism whereby the brain, confronted with a novel structured environment, initially boosts overall functional integration and then reduces interregional coupling over time.

Successful knowledge acquisition requires a cognitive system that is both sensitive to statistical information and able to distinguish among multiple structures (i.e., to detect pattern shifts and form distinct representations). Extensive behavioral evidence has highlighted the importance of cues to structural change, demonstrating how, without them, learners fail to detect pattern shifts and are biased in favor of early experience. Here, we seek a neural account of the mechanism underpinning this primacy effect in learning. During fMRI scanning, adult participants were presented with two artificial languages: a familiar language (L1) on which they had been pretrained followed by a novel language (L2). The languages were composed of the same syllable inventory organized according to unique statistical structures. In the absence of cues to the transition between languages, posttest familiarity judgments revealed that learners on average more accurately segmented words from the familiar language compared with the novel one. Univariate activation and functional connectivity analyses showed that participants with the strongest learning of L1 had decreased recruitment of fronto-subcortical and posterior parietal regions, in addition to a dissociation between downstream regions and early auditory cortex. Participants with a strong new language learning capacity (i.e., higher L2 scores) showed the opposite trend. Thus, we suggest that a bias toward neural efficiency, particularly as manifested by decreased sampling from the environment, accounts for the primacy effect in learning. Potential implications of this hypothesis are discussed, including the possibility that “inefficient” learning systems may be more sensitive to structural changes in a dynamic environment.

A core question in cognitive science concerns how humans acquire and represent knowledge about their environments. To this end, quantitative theories of learning processes have been formalized in an attempt to explain and predict changes in brain and behavior. We connect here statistical learning approaches in cognitive science, which are rooted in the sensitivity of learners to local distributional regularities, and network science approaches to characterizing global patterns and their emergent properties. We focus on innovative work that describes how learning is influenced by the topological properties underlying sensory input. The confluence of these theoretical approaches and this recent empirical evidence motivate the importance of scaling-up quantitative approaches to learning at both the behavioral and neural levels.

Brain anatomy and physiology support the human ability to navigate a complex space of perceptions and actions. To maneuver across an ever-changing landscape of mental states, the brain invokes cognitive control – a set of dynamic processes that engage and disengage different sets of brain regions to modulate attention, switch between tasks, and inhibit prepotent responses. Current theory suggests that cooperative and competitive interactions between brain areas may mediate processes of network reorganization that support transitions between dynamical states. In this study, we used a quantitative approach to identify distinct topological states of functional interactions and examine how their expression relates to cognitive control processes and behavior. In particular, we acquired fMRI BOLD signal in twenty–eight healthy subjects as they performed two cognitive control tasks – a local-global perception switching task using Navon figures and a Stroop interference task – each with low cognitive control demand and high cognitive control demand conditions. Based on these data, we constructed dynamic functional brain networks and used a parts-based network decomposition technique called non-negative matrix factorization to identify putative cognitive control subgraphs whose temporal expression captured key dynamical states involved in control processes. Our results demonstrate that the temporal expression of these functional subgraphs reflect cognitive demands and are associated with individual differences in task-based performance. These findings offer insight into how coordinated changes in the cooperative and competitive roles of distributed brain networks map trajectories between cognitively demanding brain states.

Network science has emerged as a powerful tool through which we can study the higher-order architectural properties of the world around us. How human learners exploit this information remains an essential question. Here, we focus on the temporal constraints that govern such a process. Participants viewed a continuous sequence of images generated by three distinct walks on a modular network. Walks varied along two critical dimensions: their predictability and the density with which they sampled from communities of images. Learners exposed to walks that richly sampled from each community exhibited a sharp increase in processing time upon entry into a new community. This effect was eliminated in a highly regular walk that sampled exhaustively from images in short, successive cycles (i.e., that increasingly minimized uncertainty about the nature of upcoming stimuli). These results demonstrate that temporal organization plays an essential role in how robustly knowledge of network architecture is acquired., Comment: 22 pages, 2 figures, 1 table, plus supplement

Prior to the advent of fMRI, the primary means of examining the mechanisms underlying learning were restricted to studying human behavior and non-human neural systems. However, recent advances in neuroimaging technology have enabled the concurrent study of human behavior and neural activity. We propose that the integration of behavioral response with brain activity provides a powerful method of investigating the process through which internal representations are formed or changed. Nevertheless, a review of the literature reveals that many fMRI studies of learning either (1) focus on outcome rather than process or (2) are built on the untested assumption that learning unfolds uniformly over time. We discuss here various challenges faced by the field and highlight studies that have begun to address them. In doing so, we aim to encourage more research that examines the process of learning by considering the interrelation of behavioral measures and fMRI recording during learning.

Functional magnetic resonance imaging (fMRI) was used to assess neural activation as participants learned to segment continuous streams of speech containing syllable sequences varying in their transi- tional probabilities. Speech streams were presented in four runs, each followed by a behavioral test to measure the extent of learning over time. Behavioral performance indicated that participants could dis- criminate statistically coherent sequences (words) from less coherent sequences (partwords). Individual rates of learning, defined as the difference in ratings for words and partwords, were used as predictors of neural activation to ask which brain areas showed activity associated with these measures. Results showed significant activity in the pars opercularis and pars triangularis regions of the left inferior frontal gyrus (LIFG). The relationship between these findings and prior work on the neural basis of statistical learning is discussed, and parallels to the frontal/subcortical network involved in other forms of implicit sequence learning are considered.