Editorial for First Edition of Special Issue “Brain Functional Connectivity: Prediction, Dynamics, and Modeling”

1. Introduction

The brain is one of the most complex and mysterious systems in the known world. Understanding its functional architecture, the dynamic interplay between distinct regions, is fundamental to deciphering the neural code of cognition and behavior [1]. This connectivity can be studied in both the frequency and time domains using a robust methodological toolkit that includes coherence, correlation, and modern artificial neural networks [2]. Revealing these functional networks is more than a mapping exercise; it is essential for uncovering the mechanisms that underlie information processing and decision-making during cognitive tasks [3].

The implications of this knowledge extend far beyond the laboratory, addressing practical and profoundly challenging problems in healthcare, clinical medicine, biomedical engineering, brain–machine interfaces, and cognitive sciences [4]. From identifying biomarkers for neurological disorders to designing adaptive neural interfaces, research in functional connectivity sits at the exciting intersection of discovery and application [5].

This Special Issue provides a multifaceted snapshot of contemporary brain connectivity research through a curated collection of 13 contributions. The compilation includes two comprehensive reviews, ten original research papers (encompassing a case report), and one perspective study, each addressing distinct aspects of the field. These works collectively advance the discipline by bridging theoretical modeling, cutting-edge experimental designs, and advanced analytical techniques applied to data from EEG, fMRI, MEG, and PET. The articles highlight not only technical and methodological sophistication but also important conceptual shifts, thereby offering an integrative view of how functional connectivity is transforming our comprehension of brain dynamics in health and disease.

The contributions assembled here exemplify the richness and real-world relevance of applied neuroscience. From the laboratory to the clinic, and from theoretical models to practical interfaces, they illustrate how the study of the connected brain is increasingly embedded in solutions for health, technology, and understanding human cognition.

2. Reviews

The first review article in this issue (Contribution 1), presented by Bubliková et al., focuses on the structural connectivity of the substantia nigra, a pivotal nucleus within the brain’s motor and dopaminergic circuits. Using diffusion tensor imaging and tractography, the authors provide a thorough synthesis of current knowledge concerning the substantia nigra’s anatomical pathways in humans. The review outlines both classical and newly described connections, including direct pars compacta projections to the thalamus, cortical inputs, and links to limbic and hippocampal regions. This mapping reinforces the view of the substantia nigra as a functionally diversified hub, with substantial implications for understanding Parkinson’s disease and related neuropsychiatric conditions [6]. By systematically addressing a gap in the literature, the article establishes a valuable framework for future investigations into dopaminergic circuitry and its role in both motor and non-motor functions.

In the second article (Contribution 2), Andreou et al. present a timely review on the shared social cognitive dysfunctions in Autism Spectrum Disorder (ASD) and schizophrenia, with a focus on the mirror neuron system (MNS) as a potential neural substrate. The authors synthesize recent neuroimaging and neurophysiological evidence, examining how altered functionality and activation within the MNS, a network of neurons responsive during both action observation and execution, may underlie deficits in imitation, empathy, and theory of mind common to both disorders. Their analysis critically evaluates whether MNS dysfunction represents a transdiagnostic mechanism linking social impairment across diagnostic boundaries, while also considering methodological challenges in measuring MNS activity in clinical populations [7]. By bridging psychiatric and neuroscience perspectives, this review offers a nuanced framework for understanding social cognition deficits and suggests pathways for future translational research aimed at biomarkers and targeted interventions [8].

3. Original Research

The ten original research papers in this Special Issue are organized into four thematic groups, each reflecting a key domain of contemporary functional connectivity research: (i) cognitive dynamics (4 papers), (ii) prediction models (2 papers), (iii) brain disorders (3 papers) and (iv) applications in sport and exercise programs (1 paper). This diversity not only illustrates the richness and methodological breadth of the field but also its expanding relevance across basic science, clinical translation, and applied human performance.

3.1. Cognitive Dynamics

A central theme across several contributions is the investigation of cognitive dynamics, specifically, how emotional and contextual factors modulate cognitive functions and decision-making processes [9]. In an electrophysiological study, Rovelli et al. (Contribution 3) demonstrate that decision contexts with higher socio-emotional salience elicit faster, emotionally guided choices. Their work underscores the value of integrating behavioral, experiential, and neural measures to characterize how individuals adaptively regulate decision-making under socially evaluative stress. Furthermore, the authors highlight the utility of dual-paradigm experimental designs in advancing both theory and application in cognitive–affective neuroscience.

Another compelling contribution to cognitive dynamics comes from Angioletti et al. (Contribution 4), who investigated functional connectivity during a negotiation process using electrophysiological correlates in dyads engaged in a shared decision-making task. Employing EEG hyperscanning, the authors analyzed both single-brain and inter-brain neural activity during a structured negotiation paradigm. Their findings reveal the dual nature of negotiation as both a cooperative endeavor and a cognitively demanding process, requiring emotional alignment and strategic adaptation between partners. By capturing the neural signatures of mutual influence and coordination, this study advances our understanding of the neurophysiological foundations of social negotiation and offers novel insights into how inter-brain dynamics underpin real-world collaborative decision-making. While the study demonstrates that negotiation elicits measurable inter-brain synchrony, a critical question remains: what is the precise functional role of this neural alignment? Is it a marker of successful communication, a mechanism for emotional contagion, a facilitator of mutual understanding, or simply an epiphenomenon of shared attention? Future research could dissect these possibilities by correlating specific patterns of inter-brain connectivity with behavioral outcomes such as deal quality, joint payoff, or post-negotiation trust [10]. This would move the field from demonstrating that brains synchronize to explaining why and when such synchronization is beneficial or costly.

Carvalho et al. (Contribution 5) employ advanced tensor decomposition techniques to identify individual brain fingerprints with high accuracy using resting-state fMRI data. The concept of a functional connectome (FC) fingerprint, a unique, stable neural signature identifiable across scanning sessions, has gained substantial empirical support over the past decade through studies using fMRI [11,12] and EEG [13,14]. In this work, the authors implement a Tucker decomposition framework for FC fingerprinting, enabling robust identification of an individual’s functional connectome from a set of repeated fMRI sessions. Their results demonstrate the high potential of tensor-based methods to uncover individualized neural signatures, offering a mathematically elegant and computationally efficient approach to person-level brain characterization. Perhaps individualized connectome fingerprints may serve as sensitive baselines for detecting subtle pathological deviations, offering a pathway to personalized biomarkers for neurological or psychiatric disorders. The tensor methods showcased here could enhance our ability to detect early signs of Alzheimer’s disease, depression, or other conditions by quantifying deviations from a person’s own neural baseline [15].

The case study by Pisarchik et al. (Contribution 6) introduces a novel methodological framework for constructing both graphs and hypergraphs of functional brain connectivity from single-subject EEG data during endogenous selective visual attention. The authors generate a comprehensive set of connectivity networks using multiple complementary measures (coherence, cross-correlation, and mutual information), thereby capturing linear, nonlinear, and statistical dependencies between brain regions. This multi-metric approach provides a richer, more nuanced characterization of the functional relationships underpinning attentional control. The framework’s ability to model higher-order interactions via hypergraphs is particularly notable, as it moves beyond traditional pairwise connections to capture more complex network dynamics. These methodological innovations have direct translational potential for attention monitoring systems and the clinical assessment of attention-deficit disorders [16]. The work is timely, as recent research using effective connectivity methods like Granger causality has begun to dissect the directed networks of endogenous attention [17]. The methodology presented in contribution 6 is well-suited to extend such investigations by providing tools to model both the strength and the higher-order structure of these directed and undirected functional networks. In summary, this case study provides a valuable blueprint for single-subject network neuroscience, demonstrating how advanced graph and hypergraph theory can be applied to EEG data to reveal the complex, multi-faceted nature of functional connectivity during a fundamental cognitive process.

3.2. Prediction Models

In the realm of prediction models, Mizrahi et al. (Contribution 7) present a novel neurocomputational approach that employs EEG recordings to predict individual attachment styles during performance on the Secretary Problem—a canonical optimal stopping problem in decision theory that formalizes the trade-off between exploration and exploitation [18]. Classically, the Secretary Problem prescribes a strategy of observing the first ≈37% of options to establish a benchmark, followed by a commitment to the first subsequent option that exceeds it, thus advocating for “informed decisiveness” over exhaustive search [19]. The authors leverage this paradigm to examine whether neural signals during decision-making can reflect stable interpersonal dispositions. Using graph-based EEG features and machine learning classifiers, Mizrahi et al. successfully distinguish between attachment styles, suggesting that neural signatures of exploratory and decisional phases may encode dispositional tendencies toward anxiety or avoidance in interpersonal contexts. Their findings indicate that future research should employ larger and more diverse samples to validate and refine these predictive models. If replicated in multi-site cohorts, such graph-based EEG markers could inform personalized psychological interventions by providing an objective, neural measure of attachment-related vulnerabilities [20]. This study highlights the potential of EEG-based prediction to bridge cognitive neuroscience, computational psychiatry, and personality research, offering a pathway toward more individualized assessment and intervention in both clinical and non-clinical settings.

In the next paper, Vakorin et al. (Contribution 8) introduce a novel method for estimating neurobiological brain age, a biomarker reflecting deviation from typical aging trajectories, using resting-state MEG data. Their approach moves beyond conventional univariate analyses by examining cross-frequency coupling (CFC) within functionally defined MRI networks, capturing the dynamic interplay between neuronal oscillations as a marker of neurophysiological integrity [21]. Unlike models that emphasize static baseline activity, their framework highlights neural capacity: the brain’s ability to flexibly coordinate oscillatory activity across frequencies. The results suggest that such capacity-sensitive indices may offer a more nuanced and state-dependent perspective on brain aging, potentially detecting early or subtle deviations associated with cognitive resilience or decline [22]. This approach may capture aspects of neural communication efficiency and cognitive reserve that are not evident in slower-changing structural measures [23]. An important question thus arises: Are CFC-derived brain-age estimates more predictive of functional outcomes, such as memory decline, processing speed, or daily functioning, than traditional neuroimaging-based predictions? A critical nuance in brain-age research lies in distinguishing pathological from normative variability. The brain age gap—the difference between estimated brain age and chronological age—has been shown to predict risks of cognitive decline, mental health disorders, and mortality [24]. However, not all individuals with an “older” brain age exhibit cognitive impairment, and vice versa. Future models could therefore benefit from integrating behavioral, lifestyle, and genetic data to differentiate benign variations in neural aging from deviations signaling increased risk for neurodegenerative disease. Such a multimodal approach aligns with the growing emphasis on precision aging research [25].

3.3. Brain Disorders

Five contributions in this Special Issue investigate alterations in functional brain connectivity across major neuropsychiatric and neurological disorders, including Autism Spectrum Disorder (ASD) and schizophrenia (review Contribution 2), Alzheimer’s disease (AD) (Contribution 9), epilepsy (Contributions 10 and 11), and cognitive depression (Contribution 12). This collective focus underscores the critical role of network-based approaches in elucidating the pathophysiological signatures of these complex conditions.

In one such study, Motta et al. (contribution 9) apply novel ¹⁸F-Fluorodeoxyglucose Positron Emission Tomography (¹⁸F-FDG PET) measures (ventricular uptake, cortical uptake, and their ratio) to probe metabolic dysfunction and aberrant glucose dynamics in patients with AD. These innovative indices reflect not only regional hypometabolism but also broader disturbances in cerebral energy homeostasis. The authors demonstrate that these PET-derived measures are significantly associated with key biological factors, including age, blood–brain barrier integrity, and mitochondrial dysfunction [26]. Notably, distinct metabolic patterns emerged across different Apolipoprotein E (APOE) genotypes, highlighting how genetic risk modulates brain energy metabolism in AD [27]. This work advances the use of quantitative PET biomarkers to capture the multifaceted metabolic alterations underlying AD progression. The findings by Motta et al. invite a broader discussion on the potential of metabolic network biomarkers for early detection and personalized prognosis in neurodegenerative disease. Future research integrating these PET measures with functional connectivity data from fMRI or MEG could offer a more comprehensive model linking metabolic deficits to network-level dysregulation [28]. Furthermore, longitudinal studies are needed to determine whether these biomarkers can track disease progression or response to emerging metabolic and neuroprotective therapies [29].

In the next contribution, Amoiridou et al. (Contribution 10) investigate brain network topological variations between patients with temporal lobe epilepsy (TLE) and extratemporal lobe epilepsy (ETLE). Research into the distinctive functional connectivity signatures of epilepsy subtypes is crucial for understanding their divergent neural mechanisms and clinical trajectories [30,31,32]. While recent studies have increasingly applied graph theory and machine learning to EEG data [33,34], the authors adopt a complementary approach using resting-state functional MRI (rs-fMRI). Their methodology is notable for its comprehensive comparison of connectivity measures: they construct individual brain graphs using three distinct metrics (undirected Pearson correlation, nonlinear undirected mutual information, and directed Granger causality) for TLE patients, ETLE patients, and healthy controls. This multi-method design allows for a robust characterization of both static and directed network properties. The graph analysis revealed that TLE patients exhibit more disassortative networks at lower density levels compared to ETLE patients, indicating differences in how hubs connect within each subtype. Furthermore, the comparison of global centralization features across groups at varying density thresholds highlighted distinct network integration and segregation profiles. These findings suggest that the brain network organization in TLE and ETLE is fundamentally shaped by the unique pathophysiology of each epilepsy type. The identification of such distinct topological signatures offers a promising path toward developing type-specific neuroimaging biomarkers for diagnosis, prognosis, and potentially for guiding targeted treatment strategies [35].

In a related network-focused study, Evans et al. (Contribution 11) investigate resting-state EEG functional connectivity between key large-scale brain networks: the dorsal attention network (DAN), the ventral attention network (VAN), and the salience network (SN), to assess their potential as electrophysiological biomarkers for depression subtypes. Their work aligns with the growing consensus that major depressive disorder (MDD) is a heterogeneous syndrome, necessitating biologically defined subtypes for more precise diagnosis and treatment [36]. The authors examined connectivity patterns across multiple frequency bands. Their analysis revealed that connectivity in the beta and gamma bands was significantly associated with the Anhedonia and Cognitive depression subtypes across and within all three networks. In contrast, alpha band connectivity showed no significant associations, and only a single significant finding emerged for the Mood or Somatic subtypes. This frequency- and subtype-specific pattern underscores the distinct neurophysiological profiles underlying different clinical presentations of depression. These results provide compelling support for the conceptualization of depression as a heterogeneous condition with identifiable neural signatures [37,38]. By identifying novel electrophysiological signatures, specifically beta/gamma hyperconnectivity in attention and salience networks for the Anhedonia and Cognitive subtypes, this study moves beyond symptom-based classification toward a more objective, network-based taxonomy. Such an approach holds significant promise for developing personalized neurobiomarkers and guiding subtype-specific therapeutic interventions [39,40].

3.4. Applications in Sport and Exercise Programs

A pilot study by Poinsard et al. (Contribution 12) bridges functional connectivity research with applied sports science. The authors investigate the neural and physiological responses of six trained male cyclists during two distinct exercise protocols: a traditional Incremental Exercise Test (IET) and a Self-Paced $\dot{\mathrm{V}}$O₂max (SPV) test. Neural activity was continuously recorded via EEG, alongside comprehensive physiological monitoring of gas exchange, heart rate, stroke volume, and power output. The IET followed a conventional fixed-intensity ramp, while the SPV protocol allowed participants to self-regulate intensity using ratings of perceived exertion (RPE), a method grounded in psychophysiological models of exercise regulation [41]. The EEG analysis revealed divergent neural efficiency profiles between the protocols. During the SPV test, beta band power spectral densities increased initially but stabilized after approximately 80% of the test duration, suggesting a plateau in neural demand and effective cognitive management of effort. In contrast, the IET elicited a continuous, monotonic increase in beta activity, indicating escalating neural resource allocation and a higher likelihood of premature central fatigue [42]. These neural patterns were paralleled by differences in perceived exertion and performance sustainability. The findings provide novel evidence that self-paced exercise, guided by RPE, optimizes neural efficiency and delays the onset of fatigue compared to externally imposed, fixed-intensity protocols. This supports the psychobiological model of endurance performance, which emphasizes the role of the brain in pacing and fatigue management [43]. Consequently, this study highlights the potential of EEG-derived metrics as objective biomarkers for optimizing athletic training regimens and underscores the value of neuroergonomic approaches in sports science [44].

4. Perspective Study

The Special Issue concludes with a perspective study by Falsaperla et al. (Contribution 13), which explores the evolving paradigm of precision medicine for epileptic and developmental encephalopathies (E/DEs). These severe, genetically heterogeneous conditions are characterized by early-onset seizures and profound developmental impairments. The authors provide a comprehensive overview of the field, emphasizing the critical role of genetic insights in designing targeted therapies and advocating for a multidisciplinary clinical framework. They also candidly address the significant barriers to widespread implementation, including diagnostic delays, limited accessibility to genetic testing, and a paucity of robust clinical evidence for novel interventions. Falsaperla et al. envision a future for encephalopathy management centered on therapies that move beyond symptomatic control to directly address underlying genetic and molecular pathophysiology, offering a more effective and individualized standard of care. Building upon this perspective, a critical and complementary research frontier lies in the development of methods for the prediction and preemptive control of seizures. The future of epilepsy management will likely integrate precision medicine with advanced neuroengineering. This involves leveraging extreme event theory [45] and modern machine learning techniques [46] to analyze neural data for early seizure prediction. Such systems could enable timely intervention by sending inhibitory signals via closed-loop neuromodulation devices. Promising modalities for this intervention include Deep Brain Stimulation (DBS) [47,48] and Vagus Nerve Stimulation (VNS) [49,50], which are evolving from open-loop to responsive, adaptive systems. The convergence of genetic precision, predictive analytics, and responsive neuromodulation heralds a transformative era where encephalopathy management is not only personalized but also proactive, aiming to prevent seizures before they clinically manifest.

5. Conclusions

This Special Issue presents a compelling cross-section of how functional connectivity research is transforming our understanding of the brain. The collected papers bridge fundamental discovery with real-world application, revealing the neural architecture of decision-making, social interaction, cognitive endurance, and clinical disorders. By weaving together methodologies from electrophysiology to artificial intelligence, the work showcased here moves beyond correlation toward prediction and mechanism. Ultimately, these contributions affirm that the study of connected brain networks is not merely an academic pursuit but a foundational tool for advancing healthcare, technology, and our fundamental understanding of human experience.