Light Through the Fog: Using Precision fMRI Data to Disentangle the Neural Substrates of Cognitive Control

Precision fMRI reveals that the language network exhibits adult-like left-hemispheric lateralization by 4 years of age

Left hemisphere damage in adulthood often leads to linguistic deficits, but many cases of early damage leave linguistic processing preserved, and a functional language system can develop in the right hemisphere. To explain this early apparent equipotentiality of the two hemispheres for language, some have proposed that the language system is bilateral during early development and only becomes left-lateralized with age. We examined language lateralization using functional magnetic resonance imaging with two large pediatric cohorts (total n=273 children ages 4-16; n=107 adults). Strong, adult-level left-hemispheric lateralization (in activation volume and response magnitude) was evident by age 4. Thus, although the right hemisphere can take over language function in some cases of early brain damage, and although some features of the language system do show protracted development (magnitude of language response and strength of inter-regional correlations in the language network), the left-hemisphere bias for language is robustly present by 4 years of age. These results call for alternative accounts of early equipotentiality of the two hemispheres for language.

Tracking Components of Bilingual Language Control in Speech Production: An fMRI Study Using Functional Localizers

When bilingual speakers switch back to speaking in their native language (L1) after having used their second language (L2), they often experience difficulty in retrieving words in their L1. This phenomenon is referred to as the L2 after-effect. We used the L2 after-effect as a lens to explore the neural bases of bilingual language control mechanisms. Our goal was twofold: first, to explore whether bilingual language control draws on domain-general or language-specific mechanisms; second, to investigate the precise mechanism(s) that drive the L2 after-effect. We used a precision fMRI approach based on functional localizers to measure the extent to which the brain activity that reflects the L2 after-effect overlaps with the language network (Fedorenko et al., 2010) and the domain-general multiple demand network (Duncan, 2010), as well as three task-specific networks that tap into interference resolution, lexical retrieval, and articulation. Forty-two Polish-English bilinguals participated in the study. Our results show that the L2 after-effect reflects increased engagement of domain-general but not language-specific resources. Furthermore, contrary to previously proposed interpretations, we did not find evidence that the effect reflects increased difficulty related to lexical access, articulation, and the resolution of lexical interference. We propose that difficulty of speech production in the picture naming paradigm-manifested as the L2 after-effect-reflects interference at a nonlinguistic level of task schemas or a general increase of cognitive control engagement during speech production in L1 after L2.

Human striatal association megaclusters

The striatum receives projections from multiple regions of the cerebral cortex consistent with the role of the basal ganglia in diverse motor, affective, and cognitive functions. Within the striatum, the caudate receives projections from association cortex, including multiple distinct regions of prefrontal cortex. Building on recent insights about the details of how juxtaposed cortical networks are specialized for distinct aspects of higher-order cognition, we revisited caudate organization using within-individual precision neuroimaging initially in two intensively scanned individuals (each scanned 31 times). Results revealed that the caudate has side-by-side regions that are coupled to at least five distinct distributed association networks, paralleling the organization observed in the cerebral cortex. We refer to these spatial groupings of regions as striatal association megaclusters. Correlation maps from closely juxtaposed seed regions placed within the megaclusters recapitulated the five distinct cortical networks, including their multiple spatially distributed regions. Striatal association megaclusters were explored in 15 additional participants (each scanned at least 8 times), finding that their presence generalizes to new participants. Analysis of the laterality of the regions within the megaclusters further revealed that they possess asymmetries paralleling their cortical counterparts. For example, caudate regions linked to the language network were left lateralized. These results extend the general notion of parallel specialized basal ganglia circuits with the additional discovery that, even within the caudate, there is fine-grained separation of multiple distinct higher-order networks that reflects the organization and lateralization found in the cerebral cortex.NEW & NOTEWORTHY An individualized precision neuroimaging approach reveals juxtaposed zones of the caudate that are coupled with five distinct networks in association cortex. The organization of these caudate zones recapitulates organization observed in the cerebral cortex and extends the notion of specialized basal ganglia circuits.

Organization of the human cerebral cortex estimated within individuals: networks, global topography, and function

The cerebral cortex is populated by specialized regions that are organized into networks. Here we estimated networks from functional MRI (fMRI) data in intensively sampled participants. The procedure was developed in two participants (scanned 31 times) and then prospectively applied to 15 participants (scanned 8-11 times). Analysis of the networks revealed a global organization. Locally organized first-order sensory and motor networks were surrounded by spatially adjacent second-order networks that linked to distant regions. Third-order networks possessed regions distributed widely throughout association cortex. Regions of distinct third-order networks displayed side-by-side juxtapositions with a pattern that repeated across multiple cortical zones. We refer to these as supra-areal association megaclusters (SAAMs). Within each SAAM, two candidate control regions were adjacent to three separate domain-specialized regions. Response properties were explored with task data. The somatomotor and visual networks responded to body movements and visual stimulation, respectively. Second-order networks responded to transients in an oddball detection task, consistent with a role in orienting to salient events. The third-order networks, including distinct regions within each SAAM, showed two levels of functional specialization. Regions linked to candidate control networks responded to working memory load across multiple stimulus domains. The remaining regions dissociated across language, social, and spatial/episodic processing domains. These results suggest that progressively higher-order networks nest outward from primary sensory and motor cortices. Within the apex zones of association cortex, there is specialization that repeatedly divides domain-flexible from domain-specialized regions. We discuss implications of these findings, including how repeating organizational motifs may emerge during development.NEW & NOTEWORTHY The organization of cerebral networks was estimated within individuals with intensive, repeat sampling of fMRI data. A hierarchical organization emerged in each individual that delineated first-, second-, and third-order cortical networks. Regions of distinct third-order association networks consistently exhibited side-by-side juxtapositions that repeated across multiple cortical zones, with clear and robust functional specialization among the embedded regions.

Two common and distinct forms of variation in human functional brain networks

The cortex has a characteristic layout with specialized functional areas forming distributed large-scale networks. However, substantial work shows striking variation in this organization across people, which relates to differences in behavior. While most previous work treats individual differences as linked to boundary shifts between the borders of regions, here we show that cortical 'variants' also occur at a distance from their typical position, forming ectopic intrusions. Both 'border' and 'ectopic' variants are common across individuals, but differ in their location, network associations, properties of subgroups of individuals, activations during tasks, and prediction of behavioral phenotypes. Border variants also track significantly more with shared genetics than ectopic variants, suggesting a closer link between ectopic variants and environmental influences. This work argues that these two dissociable forms of variation-border shifts and ectopic intrusions-must be separately accounted for in the analysis of individual differences in cortical systems across people.

Lateral frontoparietal functional connectivity based on individual sulcal morphology

A salient neuroanatomical feature of the human brain is its pronounced cortical folding, and there is mounting evidence that sulcal morphology is relevant to functional brain architecture and cognition. Recent studies have emphasized putative tertiary sulci (pTS): small, shallow, late-developing, and evolutionarily new sulci that have been posited to serve as functional landmarks in association cortices. A fruitful approach to characterizing brain architecture has been to delineate regions based on transitions in fMRI-based functional connectivity profiles; however, exact regional boundaries can change depending on the data used to generate the parcellation. As sulci are fixed neuroanatomical structures, here, we propose to anchor functional connectivity to individual-level sulcal anatomy. We characterized fine-grained patterns of functional connectivity across 42 sulci in lateral prefrontal (LPFC) and lateral parietal cortices (LPC) in a pediatric sample (N = 43; 20 female; ages 7-18). Further, we test for relationships between pTS morphology and functional network architecture, focusing on depth as a defining characteristic of these shallow sulci, and one that has been linked to variability in cognition. We find that 1) individual sulci have distinct patterns of connectivity, but nonetheless cluster together into groups with similar patterns - in some cases with distant rather than neighboring sulci, 2) there is moderate agreement in cluster assignments at the group and individual levels, underscoring the need for individual-level analyses, and 3) across individuals, greater depth was associated with higher network centrality for several pTS. These results highlight the importance of considering individual sulcal morphology for understanding functional brain organization.

Significance Statement

A salient, and functionally relevant, feature of the human brain is its pronounced cortical folding. However, the links between sulcal anatomy and brain function are still poorly understood - particularly for small, shallow, individually variable sulci in association cortices. Here, we explore functional connectivity among individually defined sulci in lateral prefrontal and parietal regions. We find that individual sulci have distinct patterns of connectivity but nonetheless cluster together into groups with similar connectivity - in some cases spanning lateral prefrontal and parietal sulci. We further show that the network centrality of specific sulci is positively associated with their depth, thereby helping to bridge the gap between individual differences in brain anatomy and functional networks leveraging the sulcal anatomy of the individual.

Enhancing cognitive control with transcranial magnetic stimulation in subject-specific frontoparietal networks

BACKGROUND

Cognitive control processes, including those involving frontoparietal networks, are highly variable between individuals, posing challenges to basic and clinical sciences. While distinct frontoparietal networks have been associated with specific cognitive control functions such as switching, inhibition, and working memory updating functions, there have been few basic tests of the role of these networks at the individual level.

METHODS

To examine the role of cognitive control at the individual level, we conducted a within-subject excitatory transcranial magnetic stimulation (TMS) study in 19 healthy individuals that targeted intrinsic ("resting") frontoparietal networks. Person-specific intrinsic networks were identified with resting state functional magnetic resonance imaging scans to determine TMS targets. The participants performed three cognitive control tasks: an adapted Navon figure-ground task (requiring set switching), n-back (working memory), and Stroop color-word (inhibition).

OBJECTIVE

Hypothesis: We predicted that stimulating a network associated with externally oriented control [the "FPCN-B" (fronto-parietal control network)] would improve performance on the set switching and working memory task relative to a network associated with attention (the Dorsal Attention Network, DAN) and cranial vertex in a full within-subjects crossover design.

RESULTS

We found that set switching performance was enhanced by FPCN-B stimulation along with some evidence of enhancement in the higher-demand n-back conditions.

CONCLUSION

Higher task demands or proactive control might be a distinguishing role of the FPCN-B, and personalized intrinsic network targeting is feasible in TMS designs.

Why the Single-N Design Should Be the Default in Affective Neuroscience

Many studies in affective neuroscience rely on statistical procedures designed to estimate population averages and base their main conclusions on group averages. However, the obvious unit of analysis in affective neuroscience is the individual, not the group, because emotions are individual phenomena that typically vary across individuals. Conclusions based on group averages may therefore be misleading or wrong, if interpreted as statements about emotions of an individual, or meaningless, if interpreted as statements about the group, which has no emotions. We therefore advocate the Single-N design as the default strategy in research on emotions, testing one or several individuals extensively with the primary purpose of obtaining results at the individual level. In neuroscience, the equivalent to the Single-N design is deep imaging, the emerging trend of extensive measurements of activity in single brains. Apart from the fact that individuals react differently to emotional stimuli, they also vary in shape and size of their brains. Group-based analysis of brain imaging data therefore refers to an "average brain" that was activated in a way that may not be representative of the physiology of any of the tested individual brains, nor of how these brains responded to the experimental stimuli. Deep imaging avoids such group-averaging artifacts by simply focusing on the individual brain. This methodological shift toward individual analysis has already opened new research areas in fields like vision science. Inspired by this, we call for a corresponding shift in affective neuroscience, away from group averages, and toward experimental designs targeting the individual.

Driving and suppressing the human language network using large language models

Transformer models such as GPT generate human-like language and are highly predictive of human brain responses to language. Here, using fMRI-measured brain responses to 1,000 diverse sentences, we first show that a GPT-based encoding model can predict the magnitude of brain response associated with each sentence. Then, we use the model to identify new sentences that are predicted to drive or suppress responses in the human language network. We show that these model-selected novel sentences indeed strongly drive and suppress activity of human language areas in new individuals. A systematic analysis of the model-selected sentences reveals that surprisal and well-formedness of linguistic input are key determinants of response strength in the language network. These results establish the ability of neural network models to not only mimic human language but also noninvasively control neural activity in higher-level cortical areas, like the language network.

Human Striatal Association Megaclusters

The striatum receives projections from multiple regions of the cerebral cortex consistent with its role in diverse motor, affective, and cognitive functions. Supporting cognitive functions, the caudate receives projections from cortical association regions. Building on recent insights about the details of how multiple cortical networks are specialized for distinct aspects of higher-order cognition, we revisited caudate organization using within-individual precision neuroimaging (n=2, each participant scanned 31 times). Detailed analysis revealed that the caudate has side-by-side zones that are coupled to at least Give distinct distributed association networks, paralleling the specialization observed in the cerebral cortex. Examining correlation maps from closely juxtaposed seed regions in the caudate recapitulated the Give distinct cerebral networks including their multiple spatially distributed regions. These results extend the general notion of parallel specialized basal ganglia circuits, with the additional discovery that even within the caudate, there is Gine-grained separation of multiple distinct higher-order networks.

Controversies and progress on standardization of large-scale brain network nomenclature

Progress in scientific disciplines is accompanied by standardization of terminology. Network neuroscience, at the level of macroscale organization of the brain, is beginning to confront the challenges associated with developing a taxonomy of its fundamental explanatory constructs. The Workgroup for HArmonized Taxonomy of NETworks (WHATNET) was formed in 2020 as an Organization for Human Brain Mapping (OHBM)-endorsed best practices committee to provide recommendations on points of consensus, identify open questions, and highlight areas of ongoing debate in the service of moving the field toward standardized reporting of network neuroscience results. The committee conducted a survey to catalog current practices in large-scale brain network nomenclature. A few well-known network names (e.g., default mode network) dominated responses to the survey, and a number of illuminating points of disagreement emerged. We summarize survey results and provide initial considerations and recommendations from the workgroup. This perspective piece includes a selective review of challenges to this enterprise, including (1) network scale, resolution, and hierarchies; (2) interindividual variability of networks; (3) dynamics and nonstationarity of networks; (4) consideration of network affiliations of subcortical structures; and (5) consideration of multimodal information. We close with minimal reporting guidelines for the cognitive and network neuroscience communities to adopt.

Brain hubs defined in the group do not overlap with regions of high inter-individual variability

Connector 'hubs' are brain regions with links to multiple networks. These regions are hypothesized to play a critical role in brain function. While hubs are often identified based on group-average functional magnetic resonance imaging (fMRI) data, there is considerable inter-subject variation in the functional connectivity profiles of the brain, especially in association regions where hubs tend to be located. Here we investigated how group hubs are related to locations of inter-individual variability. To answer this question, we examined inter-individual variation at group-level hubs in both the Midnight Scan Club and Human Connectome Project datasets. The top group hubs defined based on the participation coefficient did not overlap strongly with the most prominent regions of inter-individual variation (termed 'variants' in prior work). These hubs have relatively strong similarity across participants and consistent cross-network profiles, similar to what was seen for many other areas of cortex. Consistency across participants was further improved when these hubs were allowed to shift slightly in local position. Thus, our results demonstrate that the top group hubs defined with the participation coefficient are generally consistent across people, suggesting they may represent conserved cross-network bridges. More caution is warranted with alternative hub measures, such as community density (which are based on spatial proximity to network borders) and intermediate hub regions which show higher correspondence to locations of individual variability.

Within-Individual Organization of the Human Cerebral Cortex: Networks, Global Topography, and Function

The human cerebral cortex is populated by specialized regions that are organized into networks. Here we estimated networks using a Multi-Session Hierarchical Bayesian Model (MS-HBM) applied to intensively sampled within-individual functional MRI (fMRI) data. The network estimation procedure was initially developed and tested in two participants (each scanned 31 times) and then prospectively applied to 15 new participants (each scanned 8 to 11 times). Detailed analysis of the networks revealed a global organization. Locally organized first-order sensory and motor networks were surrounded by spatially adjacent second-order networks that also linked to distant regions. Third-order networks each possessed regions distributed widely throughout association cortex. Moreover, regions of distinct third-order networks displayed side-by-side juxtapositions with a pattern that repeated similarly across multiple cortical zones. We refer to these as Supra-Areal Association Megaclusters (SAAMs). Within each SAAM, two candidate control regions were typically adjacent to three separate domain-specialized regions. Independent task data were analyzed to explore functional response properties. The somatomotor and visual first-order networks responded to body movements and visual stimulation, respectively. A subset of the second-order networks responded to transients in an oddball detection task, consistent with a role in orienting to salient or novel events. The third-order networks, including distinct regions within each SAAM, showed two levels of functional specialization. Regions linked to candidate control networks responded to working memory load across multiple stimulus domains. The remaining regions within each SAAM did not track working memory load but rather dissociated across language, social, and spatial / episodic processing domains. These results support a model of the cerebral cortex in which progressively higher-order networks nest outwards from primary sensory and motor cortices. Within the apex zones of association cortex there is specialization of large-scale networks that divides domain-flexible from domain-specialized regions repeatedly across parietal, temporal, and prefrontal cortices. We discuss implications of these findings including how repeating organizational motifs may emerge during development.

Non-literal language processing is jointly supported by the language and theory of mind networks: Evidence from a novel meta-analytic fMRI approach

Going beyond the literal meaning of language is key to communicative success. However, the mechanisms that support non-literal inferences remain debated. Using a novel meta-analytic approach, we evaluate the contribution of linguistic, social-cognitive, and executive mechanisms to non-literal interpretation. We identified 74 fMRI experiments (n = 1,430 participants) from 2001 to 2021 that contrasted non-literal language comprehension with a literal control condition, spanning ten phenomena (e.g., metaphor, irony, indirect speech). Applying the activation likelihood estimation approach to the 825 activation peaks yielded six left-lateralized clusters. We then evaluated the locations of both the individual-study peaks and the clusters against probabilistic functional atlases (cf. anatomical locations, as is typically done) for three candidate brain networks-the language-selective network (Fedorenko, Behr, & Kanwisher, 2011), which supports language processing, the Theory of Mind (ToM) network (Saxe & Kanwisher, 2003), which supports social inferences, and the domain-general Multiple-Demand (MD) network (Duncan, 2010), which supports executive control. These atlases were created by overlaying individual activation maps of participants who performed robust and extensively validated 'localizer' tasks that selectively target each network in question (n = 806 for language; n = 198 for ToM; n = 691 for MD). We found that both the individual-study peaks and the ALE clusters fell primarily within the language network and the ToM network. These results suggest that non-literal processing is supported by both i) mechanisms that process literal linguistic meaning, and ii) mechanisms that support general social inference. They thus undermine a strong divide between literal and non-literal aspects of language and challenge the claim that non-literal processing requires additional executive resources.

Pedunculopontine-induced cortical decoupling as the neurophysiological locus of dissociation

Mounting evidence suggests an association between aberrant sleep phenomena and dissociative experiences. However, no wake-sleep boundary theory provides a compelling explanation of dissociation or specifies its physiological substrates. We present a theoretical account of dissociation that integrates theories and empirical results from multiple lines of research concerning the domain of dissociation and the regulation of rapid eye movement (REM) sleep. This theory posits that individual differences in the circuitry governing the REM sleep promoting Pedunculopontine Nucleus and Laterodorsal Tegmental Nucleus determine the degree of similarity in the cortical connectivity profiles of wakefulness and REM sleep. We propose that a latent trait characterized by elevated dissociative experiences emerges from the decoupling of frontal executive regions due to a REM sleep-like aminergic/cholinergic balance. The Pedunculopontine-Induced Cortical Decoupling Account of Dissociation (PICDAD) suggests multiple fruitful lines of inquiry and provides novel insights. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Generalizing the control architecture of the lateral prefrontal cortex

Cognitive control guides non-habitual, goal directed behaviors allowing us to flexibly adapt to ongoing demands. Previous work has suggested that multiple cognitive control processes exist that can be classed according to their action on present-oriented/external information versus future-oriented/internal information. These processes can be mapped onto the lateral prefrontal cortex (LPFC) such that increasingly rostral areas are involved in increasingly future-oriented/internal control processes. Whether and how such processes are organized to support goal-directed behavior remains unclear. On the one hand, the LPFC may flexibly adapt based upon demands. On the other hand, there may be a consistent control architecture such as a control hierarchy that generalizes across demands. Previous work using fMRI in humans during a comprehensive control task that engaged several control processes at once found that an area in mid-LPFC consistently exerted widespread influence throughout the LPFC. These data suggested that the mid-LPFC forms an apex of a putative control hierarchy. However, whether such an architecture generalizes across tasks remains to be tested. Here, we utilized a modified comprehensive control task designed to alter how control processes influence one another to test the generalizability of the LPFC control architecture. Univariate fMRI activations revealed distinct control-related activations relative to past work. Despite these changes, effective connectivity modeling revealed a directed architecture similar to previous findings with the mid-LPFC exerting the most widespread influences throughout LPFC. These results suggest that the fundamental control architecture of the LPFC is relatively fixed, and that different demands are accommodated through modulations of this fixed architecture.

Person-specific and precision neuroimaging: Current methods and future directions

Most neuroimaging studies of brain function analyze data in normalized space to identify regions of common activation across participants. These studies treat interindividual differences in brain organization as noise, but this approach can obscure important information about the brain's functional architecture. Recently, a number of studies have adopted a person-specific approach that aims to characterize these individual differences and explore their reliability and implications for behavior. A subset of these studies has taken a precision imaging approach that collects multiple hours of data from each participant to map brain function on a finer scale. In this review, we provide a broad overview of how person-specific and precision imaging techniques have used resting-state measures to examine individual differences in the brain's organization and their impact on behavior, followed by how task-based activity continues to add detail to these discoveries. We argue that person-specific and precision approaches demonstrate substantial promise in uncovering new details of the brain's functional organization and its relationship to behavior in many areas of cognitive neuroscience. We also discuss some current limitations in this new field and some new directions it may take.

Applications of Federated Learning; Taxonomy, Challenges, and Research Trends

The federated learning technique (FL) supports the collaborative training of machine learning and deep learning models for edge network optimization. Although a complex edge network with heterogeneous devices having different constraints can affect its performance, this leads to a problem in this area. Therefore, some research can be seen to design new frameworks and approaches to improve federated learning processes. The purpose of this study is to provide an overview of the FL technique and its applicability in different domains. The key focus of the paper is to produce a systematic literature review of recent research studies that clearly describes the adoption of FL in edge networks. The search procedure was performed from April 2020 to May 2021 with a total initial number of papers being 7546 published in the duration of 2016 to 2020. The systematic literature synthesizes and compares the algorithms, models, and frameworks of federated learning. Additionally, we have presented the scope of FL applications in different industries and domains. It has been revealed after careful investigation of studies that 25% of the studies used FL in IoT and edge-based applications and 30% of studies implement the FL concept in the health industry, 10% for NLP, 10% for autonomous vehicles, 10% for mobile services, 10% for recommender systems, and 5% for FinTech. A taxonomy is also proposed on implementing FL for edge networks in different domains. Moreover, another novelty of this paper is that datasets used for the implementation of FL are discussed in detail to provide the researchers an overview of the distributed datasets, which can be used for employing FL techniques. Lastly, this study discusses the current challenges of implementing the FL technique. We have found that the areas of medical AI, IoT, edge systems, and the autonomous industry can adapt the FL in many of its sub-domains; however, the challenges these domains can encounter are statistical heterogeneity, system heterogeneity, data imbalance, resource allocation, and privacy.

Bilingual language entropy influences executive functions through functional connectivity and signal variability

How bilinguals' experience in managing multiple languages in different communicative contexts influences the cognitive and neural aspects of executive functions remains unclear. Therefore, we examined whether variations in language experience in young adult bilinguals were associated with data-driven brain functional network patterns (connectivity and signal variability) defined by performance during executive control tasks (Stroop and task-switching). Multiple aspects of language experience, such as the extent of balanced bilingualism in language proficiency and usage, and language diversity across social contexts (i.e., language entropy) were assessed. We found that greater language diversity, rather than balanced bilingualism per se, was related to higher brain network specialization and segregation concentrated on the default mode and executive control networks, and lower signal variability, a pattern linked to smaller RT-indices related to executive functioning of shifting, goal-maintenance, and conflict-monitoring. Our findings underscore the important role of language diversity in influencing bilinguals' neurocognitive characteristics of executive functioning.

Neural Coding of Cognitive Control: The Representational Similarity Analysis Approach

Cognitive control relies on distributed and potentially high-dimensional frontoparietal task representations. Yet, the classical cognitive neuroscience approach in this domain has focused on aggregating and contrasting neural measures - either via univariate or multivariate methods - along highly abstracted, 1D factors (e.g., Stroop congruency). Here, we present representational similarity analysis (RSA) as a complementary approach that can powerfully inform representational components of cognitive control theories. We review several exemplary uses of RSA in this regard. We further show that most classical paradigms, given their factorial structure, can be optimized for RSA with minimal modification. Our aim is to illustrate how RSA can be incorporated into cognitive control investigations to shed new light on old questions.