Impact of analysis methods on the reproducibility and reliability of resting-state networks

Dev-Atlas: A reference atlas of functional brain networks for typically developing adolescents

It is well accepted that the brain is functionally organized into multiple networks and extensive literature has demonstrated that the organization of these networks shows major changes during adolescence. Yet, there is limited option for a reference functional brain atlas derived from typically-developing adolescents, which is problematic as the reliable identification of functional brain networks crucially depends on the use of such reference functional atlases. In this context, we utilized resting-state functional MRI data from 1391 typically-developing youth aged 8-17 years to create an adolescent-specific reference atlas of functional brain networks. We further investigated the impact of age and sex on these networks. Using a multiscale individual component clustering algorithm, we identified 24 reliable functional brain networks, classified within six domains: Default-Mode (5 networks), Control (4 networks), Salience (3 networks), Attention (4 networks), Somatomotor (5 networks), and Visual (3 networks). We identified reliable and large effects of age on the spatial topography of these majority of networks, as well as on the functional network connectivity. Sex effects were not as widespread. We created a novel brain atlas, named Dev-Atlas, focused on a typically-developing sample, with the hope that this atlas can be used in future developmental neuroscience studies.

Exploring the neural and behavioral correlates of cognitive telerehabilitation in mild cognitive impairment with three distinct approaches

Background

Currently, the impact of drug therapies on neurodegenerative conditions is limited. Therefore, there is a strong clinical interest in non-pharmacological interventions aimed at preserving functionality, delaying disease progression, reducing disability, and improving quality of life for both patients and their caregivers. This longitudinal multicenter Randomized Controlled Trial (RCT) applies three innovative cognitive telerehabilitation (TR) methods to evaluate their impact on brain functional connectivity reconfigurations and on the overall level of cognitive and everyday functions.

Methods

We will include 110 participants with mild cognitive impairment (MCI). Fifty-five participants will be randomly assigned to the intervention group who will receive cognitive TR via three approaches, namely: (a) Network-based Cognitive Training (NBCT), (b) Home-based Cognitive Rehabilitation (HomeCoRe), or (c) Semantic Memory Rehabilitation Training (SMRT). The control group (n = 55) will receive an unstructured home-based cognitive stimulation. The rehabilitative program will last either 4 (NBTC) or 6 weeks (HomeCoRe and SMRT), and the control condition will be adapted to each TR intervention. The effects of TR will be tested in terms of Δ connectivity change, obtained from high-density electroencephalogram (HD-EEG) or functional magnetic resonance imaging at rest (rs-fMRI), acquired before (T0) and after (T1) the intervention. All participants will undergo a comprehensive neuropsychological assessment at four time-points: baseline (T0), within 2 weeks (T1), and after 6 (T2) and 12 months (T3) from the end of TR.

Discussion

The results of this RCT will identify a potential association between improvement in performance induced by individual cognitive TR approaches and modulation of resting-state brain connectivity. The knowledge gained with this study might foster the development of novel TR approaches underpinned by established neural mechanisms to be validated and implemented in clinical practice.Clinical trial registration: [https://classic.clinicaltrials.gov/ct2/show/NCT06278818], identifier [NCT06278818].

Functional MRI and Diffusion Tensor Imaging in Migraine: A Review of Migraine Functional and White Matter Microstructural Changes

Migraine is a complex and heterogenous disorder whose disease mechanisms remain disputed. This narrative review summarizes functional MRI (fMRI) and diffusion tensor imaging (DTI) findings and interprets their association with migraine symptoms and subtype to support and expand our current understanding of migraine pathophysiology. Our PubMed search evaluated and included fMRI and DTI studies involving comparisons between migraineurs vs healthy controls, migraineurs with vs without aura, and episodic vs chronic migraineurs. Migraineurs demonstrate changes in functional connectivity (FC) and regional activation in numerous pain-related networks depending on migraine phase, presence of aura, and chronicity. Changes to diffusion indices are observed in major cortical white matter tracts extending to the brainstem and cerebellum, more prominent in chronic migraine and associated with FC changes. Reported changes in FC and regional activation likely relate to pain processing and sensory hypersensitivities. Diffuse white matter microstructural changes in dysfunctional cortical pain and sensory pathways complement these functional differences. Interpretations of reported fMRI and DTI measure trends have not achieved a clear consensus due to inconsistencies in the migraine neuroimaging literature. Future fMRI and DTI studies should establish and implement a uniform methodology that reproduces existing results and directly compares migraineurs with different subtypes. Combined fMRI and DTI imaging may provide better pathophysiological explanations for nonspecific FC and white matter microstructural differences.

Brain connectivity under light sedation with midazolam and ketamine during task performance and the periodic experience of pain: Examining concordance between different approaches for seed-based connectivity analysis

This work focused on functional connectivity changes under midazolam and ketamine sedation during performance of a memory task, with the periodic experience of pain. To maximize ability to compare to previous and future work, we performed secondary region of interest (ROI)-to-ROI functional connectivity analyses on these data, using two granularities of scale for ROIs. These findings are compared to the results of a previous seed-to-voxel analysis methodology, employed in the primary analysis. Healthy adult volunteers participated in this randomized crossover 3 T functional MRI study under no drug, followed by subanesthetic doses of midazolam or ketamine achieving minimal sedation. Periodic painful stimulation was delivered while subjects repeatedly performed a memory-encoding task. Atlas-based and network-level ROIs were used from within Conn Toolbox (ver 18). Timing of experimental task events was regressed from the data to assess drug-induced changes in background connectivity, using ROI-to-ROI methodology. Compared to saline, ROI-to-ROI connectivity changes under ketamine did not survive correction for multiple comparisons, thus data presented is from 16 subjects in a paired analysis between saline and midazolam. In both ROI-to-ROI analyses, the predominant direction of change was towards increased connectivity under midazolam, compared to saline. These connectivity increases occurred between functionally-distinct brain areas, with a posterior-predominant spatial distribution that included many long-range connectivity changes. During performance of an experimental task that involved periodic painful stimulation, compared to saline, low-dose midazolam was associated with robust increases in functional connectivity. This finding was concordant across different seed-based analyses for midazolam, but not ketamine. The neuroimaging drug trial from which this data was drawn was pre-registered (NCT-02515890) prior to enrollment of the first subject.

Disrupted dynamic functional network connectivity in fetal alcohol spectrum disorders

BACKGROUND

Prenatal alcohol exposure (PAE) can result in harmful and long-lasting neurodevelopmental changes. Children with PAE or a fetal alcohol spectrum disorder (FASD) have decreased white matter volume and resting-state spectral power compared to typically developing controls (TDC) and impaired resting-state static functional connectivity. The impact of PAE on resting-state dynamic functional network connectivity (dFNC) is unknown.

METHODS

Using eyes-closed and eyes-open magnetoencephalography (MEG) resting-state data, global dFNC statistics and meta-states were examined in 89 children aged 6-16 years (51 TDC, 38 with FASD). Source analyzed MEG data were used as input to group spatial independent component analysis to derive functional networks from which the dFNC was calculated.

RESULTS

During eyes-closed, relative to TDC, participants with FASD spent a significantly longer time in state 2, typified by anticorrelation (i.e., decreased connectivity) within and between default mode network (DMN) and visual network (VN), and state 4, typified by stronger internetwork correlation. The FASD group exhibited greater dynamic fluidity and dynamic range (i.e., entered more states, changed from one meta-state to another more often, and traveled greater distances) than TDC. During eyes-open, TDC spent significantly more time in state 1, typified by positive intra- and interdomain connectivity with modest correlation within the frontal network (FN), while participants with FASD spent a larger fraction of time in state 2, typified by anticorrelation within and between DMN and VN and strong correlation within and between FN, attention network, and sensorimotor network.

CONCLUSIONS

There are important resting-state dFNC differences between children with FASD and TDC. Participants with FASD exhibited greater dynamic fluidity and dynamic range and spent more time in states typified by anticorrelation within and between DMN and VN, and more time in a state typified by high internetwork connectivity. Taken together, these network aberrations indicate that prenatal alcohol exposure has a global effect on resting-state connectivity.

Multivariate patterns of brain-behavior associations across the adult lifespan

The nature of brain-behavior covariations with increasing age is poorly understood. In the current study, we used a multivariate approach to investigate the covariation between behavioral-health variables and brain features across adulthood. We recruited healthy adults aged 20–73 years-old (29 younger, mean age = 25.6 years; 30 older, mean age = 62.5 years), and collected structural and functional MRI (s/fMRI) during a resting-state and three tasks. From the sMRI, we extracted cortical thickness and subcortical volumes; from the fMRI, we extracted activation peaks and functional network connectivity (FNC) for each task. We conducted canonical correlation analyses between behavioral-health variables and the sMRI, or the fMRI variables, across all participants. We found significant covariations for both types of neuroimaging phenotypes (ps = 0.0004) across all individuals, with cognitive capacity and age being the largest opposite contributors. We further identified different variables contributing to the models across phenotypes and age groups. Particularly, we found behavior was associated with different neuroimaging patterns between the younger and older groups. Higher cognitive capacity was supported by activation and FNC within the executive networks in the younger adults, while it was supported by the visual networks’ FNC in the older adults. This study highlights how the brain-behavior covariations vary across adulthood and provides further support that cognitive performance relies on regional recruitment that differs between older and younger individuals.

Using Network Parcels and Resting-State Networks to Estimate Correlates of Mood Disorder and Related Research Domain Criteria Constructs of Reward Responsiveness and Inhibitory Control

BACKGROUND

Resting-state graph-based network edges can be powerful tools for identification of mood disorders. We address whether these edges can be integrated with Research Domain Criteria (RDoC) constructs for accurate identification of mood disorder-related markers, while minimizing active symptoms of disease.

METHODS

We compared 132 individuals with currently remitted or euthymic mood disorder with 65 healthy comparison participants, ages 18-30 years. Subsets of smaller brain parcels, combined into three prominent networks and one network of parcels overlapping across these networks, were used to compare edge differences between groups. Consistent with the RDoC framework, we evaluated individual differences with performance measure regressors of inhibitory control and reward responsivity. Within an omnibus regression model, we predicted edges related to diagnostic group membership, performance within both RDoC domains, and relevant interactions.

RESULTS

There were several edges of mood disorder group, predominantly of greater connectivity across networks, different than those related to individual differences in inhibitory control and reward responsivity. Edges related to diagnosis and inhibitory control did not align well with prior literature, whereas edges in relation to reward responsivity constructs showed greater alignment with prior literature. Those edges in interaction between RDoC constructs and diagnosis showed a divergence for inhibitory control (negative interactions in default mode) relative to reward (positive interactions with salience and emotion network).

CONCLUSIONS

In conclusion, there is evidence that prior simple network models of mood disorders are currently of insufficient biological or diagnostic clarity or that parcel-based edges may be insufficiently sensitive for these purposes.

Weaker Connectivity of the Cortical Networks Is Linked with the Uncharacteristic Gait in Youth with Cerebral Palsy

Cerebral palsy (CP) is the most prevalent pediatric neurologic impairment and is associated with major mobility deficiencies. This has led to extensive investigations of the sensorimotor network, with far less research focusing on other major networks. The aim of this study was to investigate the functional connectivity (FC) of the main sensory networks (i.e., visual and auditory) and the sensorimotor network, and to link FC to the gait biomechanics of youth with CP. Using resting-state functional magnetic resonance imaging, we first identified the sensorimotor, visual and auditory networks in youth with CP and neurotypical controls. Our analysis revealed reduced FC among the networks in the youth with CP relative to the controls. Notably, the visual network showed lower FC with both the sensorimotor and auditory networks. Furthermore, higher FC between the visual and sensorimotor cortices was associated with larger step length (r = 0.74, p_FDR = 0.04) in youth with CP. These results confirm that CP is associated with functional brain abnormalities beyond the sensorimotor network, suggesting abnormal functional integration of the brain’s motor and primary sensory systems. The significant association between abnormal visuo-motor FC and gait could indicate a link with visuomotor disorders in this patient population.

Recognition memory performance can be estimated based on brain activation networks

BACKGROUND

Recognition memory is an essential ability for functioning in everyday life. Establishing robust brain networks linked to recognition memory performance can help to understand the neural basis of recognition memory itself and the interindividual differences in recognition memory performance.

METHODS

We analysed behavioural and whole-brain fMRI data from 1'410 healthy young adults during the testing phase of a picture-recognition task. Using independent component analysis (ICA), we decomposed the fMRI contrast for previously seen vs. new (old-new) pictures into networks of brain activity. This was done in two independent samples (training sample: N = 645, replication sample: N = 665). Next, we investigated the relationship between the identified brain networks and interindividual differences in recognition memory performance by conducting a prediction analysis. We estimated the prediction accuracy in a third independent sample (test sample: N = 100).

RESULTS

We identified 12 robust and replicable brain networks using two independent samples. Based on the activity of those networks we could successfully estimate interindividual differences in recognition memory performance with high accuracy in a third independent sample (r = 0.5, p = 1.29 × 10^-07).

CONCLUSION

Given the robustness of the ICA decomposition as well as the high prediction estimate, the identified brain networks may be considered as potential biomarkers of recognition memory performance in healthy young adults and can be further investigated in the context of health and disease.

Atlas55+: Brain Functional Atlas of Resting-State Networks for Late Adulthood

Currently, several human brain functional atlases are used to define the spatial constituents of the resting-state networks (RSNs). However, the only brain atlases available are derived from samples of young adults. As brain networks are continuously reconfigured throughout life, the lack of brain atlases derived from older populations may influence RSN results in late adulthood. To address this gap, the aim of the study was to construct a reliable brain atlas derived only from older participants. We leveraged resting-state functional magnetic resonance imaging data from three cohorts of healthy older adults (total N = 563; age = 55-95 years) and a younger-adult cohort (N = 128; age = 18-35 years). We identified the major RSNs and their subdivisions across all older-adult cohorts. We demonstrated high spatial reproducibility of these RSNs with an average spatial overlap of 67%. Importantly, the RSNs derived from the older-adult cohorts were spatially different from those derived from the younger-adult cohort (P = 2.3 × 10-3). Lastly, we constructed a novel brain atlas, called Atlas55+, which includes the consensus of the major RSNs and their subdivisions across the older-adult cohorts. Thus, Atlas55+ provides a reliable age-appropriate template for RSNs in late adulthood and is publicly available. Our results confirm the need for age-appropriate functional atlases for studies investigating aging-related brain mechanisms.

Functional connectivity of the default mode, dorsal attention and fronto-parietal executive control networks in glial tumor patients

PURPOSE

Resting state functional magnetic resonance imaging (rsfMRI) is an emerging tool to explore the functional connectivity of different brain regions. We aimed to assess the disruption of functional connectivity of the Default Mode Network (DMN), Dorsal Attention Network(DAN) and Fronto-Parietal Network (FPN) in patients with glial tumors.

METHODS

rsfMRI data acquired on 3T-MR of treatment-naive glioma patients prospectively recruited (2015-2019) and matched controls from the 1000 functional-connectomes-project were analyzed using the CONN functional toolbox. Seed-Based Connectivity Analysis (SBCA) and Independent Component Analysis (ICA, with 10 to 100 components) were performed to study reliably the three networks of interest.

RESULTS

35 patients with gliomas (17 WHO grade I-II, 18 grade III-IV) and 70 controls were included. Global increased DMN connectivity was consistently found with SBCA and ICA in patients compared to controls (Cluster1: Precuneus, height: p < 10^-6; Cluster2: subcallosum; height: p < 10^-5). However, an area of decreased connectivity was found in the posterior corpus callosum, particularly in high-grade gliomas (height: p < 10^-5). The DAN demonstrated small areas of increased connectivity in frontal and occipital regions (height: p < 10^-6). For the FPN, increased connectivity was noted in the precuneus, posterior cingulate gyrus, and frontal cortex. No difference in the connectivity of the networks of interest was demonstrated between low- and high-grade gliomas, as well as when stratified by their IDH1-R132H (isocitrate dehydrogenase) mutation status.

CONCLUSION

Altered functional connectivity is reliably found with SBCA and ICA in the DMN, DAN, and FPN in glioma patients, possibly explained by decreased connectivity between the cerebral hemispheres across the corpus callosum due to disruption of the connections.

Scanning Conditions in Functional Connectivity Magnetic Resonance Imaging: How to Standardise Resting-State for Optimal Data Acquisition and Visualisation?

Functional connectivity magnetic resonance imaging (fcMRI), performed during resting wakefulness without tasks or stimulation, is a non-invasive technique to assess and visualise functional brain networks in vivo. Acquisition of resting-state imaging data has become increasingly common in longitudinal studies to investigate brain health and disease. However, the scanning protocols vary considerably across different institutions creating challenges for comparability especially for the interpretation of findings in patient cohorts and establishment of diagnostic or prognostic imaging biomarkers. The aim of this chapter is to discuss the effect of two experimental conditions (i.e. a low cognitive demand paradigm and a pure resting-state fcMRI) on the reproducibility of brain networks between a baseline and a follow-up session, 30 (±5) days later, acquired from 12 right-handed volunteers (29 ± 5 yrs). A novel method was developed and used for a direct statistical comparison of the test-retest reliability using 28 well-established functional brain networks. Overall, both scanning conditions produced good levels of test-retest reliability. While the pure resting-state condition showed higher test-retest reliability for 18 of the 28 analysed networks, the low cognitive demand paradigm produced higher test-retest reliability for 8 of the 28 brain networks (i.e. visual, sensorimotor and frontal areas); in 2 of the 28 brain networks no significant changes could be detected. These results are relevant to planning of longitudinal studies, as higher test-retest reliability generally increases statistical power. This work also makes an important contribution to neuroimaging where optimising fcMRI experimental scanning conditions, and hence data visualisation of brain function, remains an on-going topic of interest. In this chapter, we provide a full methodological explanation of the two paradigms and our analysis so that readers can apply them to their own scanning protocols.

A longitudinal investigation of structural brain changes during second language learning

Few studies have examined the time course of second language (L2) induced neuroplasticity or how individual differences may be associated with brain changes. The current longitudinal structural magnetic resonance imaging study examined changes in cortical thickness (CT) and gray matter volume (GMV) across two semesters of L2 Spanish classroom learning. Learners' lexical processing was assessed via a language decision task containing English and Spanish words. Our findings indicated that (1) CT increased in the left anterior cingulate cortex (ACC) and right middle temporal gyrus (MTG) after L2 learning, (2) CT in the right MTG increased in individuals who were better able to discriminate between native language and L2 words, and (3) CT in the left ACC was correlated with functional connectivity between the ACC and MTG. These findings indicate that L2 lexical development is associated with functional and structural changes in brain regions important for cognitive control and semantic processing.

Evaluation of the spatial variability in the major resting-state networks across human brain functional atlases

The human brain is intrinsically organized into resting‐state networks (RSNs). Currently, several human brain functional atlases are used to define the spatial constituents of these RSNs. However, there are significant concerns about interatlas variability. In response, we undertook a quantitative comparison of the five major RSNs (default mode [DMN], salience, central executive, sensorimotor, and visual networks) across currently available brain functional atlases (n = 6) in which we demonstrated that (a) similarity between atlases was modest and positively linked to the size of the sample used to construct them; (b) across atlases, spatial overlap among major RSNs ranged between 17 and 76% (mean = 39%), which resulted in variability in their functional connectivity; (c) lower order RSNs were generally spatially conserved across atlases; (d) among higher order RSNs, the DMN was the most conserved across atlases; and (e) voxel‐wise flexibility (i.e., the likelihood of a voxel to change network assignment across atlases) was high for subcortical regions and low for the sensory, motor and medial prefrontal cortices, and the precuneus. In order to facilitate RSN reproducibility in future studies, we provide a new freely available Consensual Atlas of REsting‐state Networks, based on the most reliable atlases.

Comparison of Resting-State Brain Activation Detected by BOLD, Blood Volume and Blood Flow

Resting-state brain activity has been widely investigated using blood oxygenation level dependent (BOLD) contrast techniques. However, BOLD signal changes reflect a combination of the effects of cerebral blood flow (CBF), cerebral blood volume (CBV), as well as the cerebral metabolic rate of oxygen (CMRO₂). In this study, resting-state brain activation was detected and compared using the following techniques: (a) BOLD, using a gradient-echo echo planar imaging (GE-EPI) sequence; (b) CBV-weighted signal, acquired using gradient and spin echo (GRASE) based vascular space occupancy (VASO); and (c) CBF, using pseudo-continuous arterial spin labeling (pCASL). Reliable brain networks were detected using VASO and ASL, including sensorimotor, auditory, primary visual, higher visual, default mode, salience and left/right executive control networks. Differences between the resting-state activation detected with ASL, VASO and BOLD could potentially be due to the different temporal signal-to-noise ratio (tSNR) and the short post-labeling delay (PLD) in ASL, along with differences in the spin-echo readout of VASO. It is also possible that the dynamics of spontaneous fluctuations in BOLD, CBV and CBF could differ due to biological reasons, according to their location within the brain.

Differences in Cortical Structure and Functional MRI Connectivity in High Functioning Autism

Autism spectrum disorders (ASD) represent a complex group of neurodevelopmental conditions characterized by deficits in communication and social behaviors. We examined the functional connectivity (FC) of the default mode network (DMN) and its relation to multimodal morphometry to investigate superregional, system-level alterations in a group of 22 adolescents and young adults with high-functioning autism compared to age-, and intelligence quotient-matched 29 healthy controls. The main findings were that ASD patients had gray matter (GM) reduction, decreased cortical thickness and larger cortical surface areas in several brain regions, including the cingulate, temporal lobes, and amygdala, as well as increased gyrification in regions associated with encoding visual memories and areas of the sensorimotor component of the DMN, more pronounced in the left hemisphere. Moreover, patients with ASD had decreased connectivity between the posterior cingulate cortex, and areas of the executive control component of the DMN and increased FC between the anteromedial prefrontal cortex and areas of the sensorimotor component of the DMN. Reduced cortical thickness in the right inferior frontal lobe correlated with higher social impairment according to the scores of the Autism Diagnostic Interview-Revised (ADI-R). Reduced cortical thickness in left frontal regions, as well as an increased cortical thickness in the right temporal pole and posterior cingulate, were associated with worse scores on the communication domain of the ADI-R. We found no association between scores on the restrictive and repetitive behaviors domain of ADI-R with structural measures or FC. The combination of these structural and connectivity abnormalities may help to explain some of the core behaviors in high-functioning ASD and need to be investigated further.

Adolescent Gender Differences in Cognitive Control Performance and Functional Connectivity Between Default Mode and Fronto-Parietal Networks Within a Self-Referential Context

Ineffective reduction of functional connectivity between the default mode network (DMN) and frontoparietal network (FPN) during cognitive control can interfere with performance in healthy individuals—a phenomenon present in psychiatric disorders, such as depression. Here, this mechanism is studied in healthy adolescents by examining gender differences in task-regressed functional connectivity using functional magnetic resonance imaging (MRI) and a novel task designed to place the DMN—supporting self-referential processing (SRP)—and FPN—supporting cognitive control—into conflict. Compared to boys, girls showed stronger functional connectivity between DMN and FPN during cognitive control in an SRP context (n = 40; boys = 20), a context that also elicited more errors of omission in girls. The gender difference in errors of omission was mediated by higher self-reported co-rumination—the extensive and repetitive discussion of problems and focus on negative feelings with a same-gender peer—by girls, compared to boys. These findings indicate that placing internal and external attentional demands in conflict lead to persistent functional connectivity between FPN and DMN in girls, but not boys; however, deficits in performance during this context were explained by co-rumination, such that youth with higher co-rumination displayed the largest performance deficits. Previous research shows that co-rumination predicts depressive symptoms during adolescence; thus, gender differences in the mechanisms involved with transitioning from internal to external processing may be relevant for understanding heightened vulnerability for depression in adolescent girls.

The Effect of Low-Frequency Physiological Correction on the Reproducibility and Specificity of Resting-State fMRI Metrics: Functional Connectivity, ALFF, and ReHo

The resting-state fMRI (rs-fMRI) signal is affected by a variety of low-frequency physiological phenomena, including variations in cardiac-rate (CRV), respiratory-volume (RVT), and end-tidal CO₂ (PETCO₂). While these effects have become better understood in recent years, the impact that their correction has on the quality of rs-fMRI measurements has yet to be clarified. The objective of this paper is to investigate the effect of correcting for CRV, RVT and PETCO₂ on the rs-fMRI measurements. Nine healthy subjects underwent a test-retest rs-fMRI acquisition using repetition times (TRs) of 2 s (long-TR) and 0.323 s (short-TR), and the data were processed using eight different physiological correction strategies. Subsequently, regional homogeneity (ReHo), amplitude of low-frequency fluctuation (ALFF), and resting-state connectivity of the motor and default-mode networks are calculated for each strategy. Reproducibility is calculated using intra-class correlation and the Dice Coefficient, while the accuracy of functional-connectivity measures is assessed through network separability, sensitivity and specificity. We found that: (1) the reproducibility of the rs-fMRI measures improved significantly after correction for PETCO₂; (2) separability of functional networks increased after PETCO₂ correction but was not affected by RVT and CRV correction; (3) the effect of physiological correction does not depend on the data sampling-rate; (4) the effect of physiological processes and correction strategies is network-specific. Our findings highlight limitations in our understanding of rs-fMRI quality measures, and underscore the importance of using multiple quality measures to determine the optimal physiological correction strategy.

Effects of dexamphetamine-induced dopamine release on resting-state network connectivity in recreational amphetamine users and healthy controls

Dexamphetamine (dAMPH) is not only used for the treatment of attention deficit hyperactivity disorder (ADHD), but also as a recreational drug. Acutely, dAMPH induces release of predominantly dopamine (DA) in the striatum, and in the cortex both DA and noradrenaline. Recent animal studies have shown that chronic dAMPH administration can induce changes in the DA system following long-term exposure, as evidenced by reductions in DA transporters, D2/3 receptors and endogenous DA levels. However, only a limited number of studies have investigated the effects of dAMPH in the human brain. We used a combination of resting-state functional magnetic resonance imaging (rs-fMRI) and [(123)I]IBZM single-photon emission computed tomography (SPECT) (to assess baseline D2/3 receptor binding and DA release) in 15 recreational AMPH users and 20 matched healthy controls to investigate the short-, and long-term effects of AMPH before and after an acute intravenous challenge with dAMPH. We found that acute dAMPH administration reduced functional connectivity in the cortico-striatal-thalamic network. dAMPH-induced DA release, but not DA D2/3 receptor binding, was positively associated with connectivity changes in this network. In addition, acute dAMPH reduced connectivity in default mode networks and salience-executive-networks networks in both groups. In contrast to our hypothesis, no significant group differences were found in any of the rs-fMRI networks investigated, possibly due to lack of sensitivity or compensatory mechanisms. Our findings thus support the use of ICA-based resting-state functional connectivity as a tool to investigate acute, but not chronic, alterations induced by dAMPH on dopaminergic processing in the striatum.

Identification of autism spectrum disorder using deep learning and the ABIDE dataset

The goal of the present study was to apply deep learning algorithms to identify autism spectrum disorder (ASD) patients from large brain imaging dataset, based solely on the patients brain activation patterns. We investigated ASD patients brain imaging data from a world-wide multi-site database known as ABIDE (Autism Brain Imaging Data Exchange). ASD is a brain-based disorder characterized by social deficits and repetitive behaviors. According to recent Centers for Disease Control data, ASD affects one in 68 children in the United States. We investigated patterns of functional connectivity that objectively identify ASD participants from functional brain imaging data, and attempted to unveil the neural patterns that emerged from the classification. The results improved the state-of-the-art by achieving 70% accuracy in identification of ASD versus control patients in the dataset. The patterns that emerged from the classification show an anticorrelation of brain function between anterior and posterior areas of the brain; the anticorrelation corroborates current empirical evidence of anterior-posterior disruption in brain connectivity in ASD. We present the results and identify the areas of the brain that contributed most to differentiating ASD from typically developing controls as per our deep learning model.

Functional and Structural Network Recovery after Mild Traumatic Brain Injury: A 1-Year Longitudinal Study

Brain connectivity after mild traumatic brain injury (mTBI) has not been investigated longitudinally with respect to both functional and structural networks together within the same patients, crucial to capture the multifaceted neuropathology of the injury and to comprehensively monitor the course of recovery and compensatory reorganizations at macro-level. We performed a prospective study with 49 mTBI patients at an average of 5 days and 1 year post-injury and 49 healthy controls. Neuropsychological assessments as well as resting-state functional and diffusion-weighted magnetic resonance imaging were obtained. Functional and structural connectome analyses were performed using network-based statistics. They included a cross-sectional group comparison and a longitudinal analysis with the factors group and time. The latter tracked the subnetworks altered at the early phase and, in addition, included a whole-brain group × time interaction analysis. Finally, we explored associations between the evolution of connectivity and changes in cognitive performance. The early phase of mTBI was characterized by a functional hypoconnectivity in a subnetwork with a large overlap of regions involved within the classical default mode network. In addition, structural hyperconnectivity in a subnetwork including central hub areas such as the cingulate cortex was found. The impaired functional and structural subnetworks were strongly correlated and revealed a large anatomical overlap. One year after trauma and compared to healthy controls we observed a partial normalization of both subnetworks along with a considerable compensation of functional and structural connectivity subsequent to the acute phase. Connectivity changes over time were correlated with improvements in working memory, divided attention, and verbal recall. Neuroplasticity-induced recovery or compensatory processes following mTBI differ between brain regions with respect to their time course and are not fully completed 1 year after trauma.

Alerted default mode: functional connectivity changes in the aftermath of social stress

Stress affects the brain at a network level: the salience network is supposedly upregulated, while at the same time the executive control network is downregulated. While theoretically described, the effects in the aftermath of stress have thus far not been tested empirically. Here, we compared for the first time resting-state functional connectivity in a large sample of healthy volunteers before and after a mild social stressor. Following the theoretical prediction, we focused on connectivity of the salience network (SN), the executive control network (ECN) and the default mode network (DMN). The DMN exhibited increased resting-state functional connectivity following the cyberball task to the key nodes of the SN, namely the dorsal anterior cingulate cortex (dACC) and the anterior insula, as well as sensorimotor regions and higher-order visual areas. We conclude that this increased connectivity of the DMN with key nodes of the SN and regions responsible for preparatory motor activity and visual motion processing indicates a shift towards an ‘alerted default mode’ in the aftermath of stress. This brain response may be triggered or aggravated by (social) stress induced by the cyberball task, enabling individuals to better reorient attention, detect salient external stimuli, and deal with the emotional and affective consequences of stress.

Pharmacologically Induced Sex Hormone Fluctuation Effects on Resting-State Functional Connectivity in a Risk Model for Depression: A Randomized Trial

Women are at relatively greater lifetime risk for depression than men. This elevated risk in women is partly due to heightened risk during time periods characterized by marked fluctuations in sex hormones, including postpartum and perimenopausal periods. How sex hormone fluctuations contribute to heightened risk is not fully understood but may involve intrinsic functional connectivity. We induced a biphasic ovarian sex hormone fluctuation using the gonadotropin-releasing hormone agonist (GnRHa) goserelin to determine, with a randomized placebo-controlled design, intervention effects on or GnRHa-provoked depressive symptoms associations with change in resting-state functional connectivity (rs-FC) in 58 healthy women for six seeds (amygdala, hippocampus, anterior cingulate cortex, dorsal raphe, median raphe, and posterior cingulate cortex). GnRHa intervention did not significantly affect rs-FC in any seeds. Considering the GnRHa group only, the emergence of depressive symptoms following intervention was positively associated with amygdala-right temporal cortex and negatively associated with hippocampus-cingulate rs-FC. A test for mediation suggested that rs-FC changes in these networks marginally mediated the association between decrease in estradiol and increase in depressive symptoms in the GnRHa group (p=0.07). Our findings provide novel evidence-linking changes in rs-FC networks, the emergence of depressive symptoms and sex hormone fluctuations. Notably, we observed evidence that changes in rs-FC may represent a key neurobiological intermediary between molecular changes induced by hormone fluctuations and the emergence of depressive symptoms. Taken together, our findings indicate that sex hormone fluctuations may contribute to heightened risk for developing depressive symptoms by affecting intrinsic functional connectivity of key limbic brain structures.

Large-scale functional network overlap is a general property of brain functional organization: Reconciling inconsistent fMRI findings from general-linear-model-based analyses

Functional magnetic resonance imaging (fMRI) studies regularly use univariate general-linear-model-based analyses (GLM). Their findings are often inconsistent across different studies, perhaps because of several fundamental brain properties including functional heterogeneity, balanced excitation and inhibition (E/I), and sparseness of neuronal activities. These properties stipulate heterogeneous neuronal activities in the same voxels and likely limit the sensitivity and specificity of GLM. This paper selectively reviews findings of histological and electrophysiological studies and fMRI spatial independent component analysis (sICA) and reports new findings by applying sICA to two existing datasets. The extant and new findings consistently demonstrate several novel features of brain functional organization not revealed by GLM. They include overlap of large-scale functional networks (FNs) and their concurrent opposite modulations, and no significant modulations in activity of most FNs across the whole brain during any task conditions. These novel features of brain functional organization are highly consistent with the brain's properties of functional heterogeneity, balanced E/I, and sparseness of neuronal activity, and may help reconcile inconsistent GLM findings.

Resting-state networks in 6-to-10 year old children

Resting-state functional magnetic resonance imaging provides a non-invasive approach to the study of intrinsic functional brain networks. When applied to the study of brain development, most studies consist of relatively small samples that are not always representative of the general population. Descriptions of these networks in the general population offer important insight for clinical studies examining, for instance, psychopathology or neurological conditions. Thus our goal was to characterize resting-state networks in a large sample of children using independent component analysis (ICA). The study further aimed to describe the robustness of these networks by examining which networks occur frequently after repeated ICA. Resting-state networks were obtained from a sample of 536 6-to-10 year old children. Distributions of networks were built from repeated subsampling and group ICA analyses, and meta-ICA was used to construct a representative set of components. Within- and between-network properties were tested for age-related developmental associations using spatio-temporal regression. After repeated ICA, many networks were present over 95% of the time suggesting the components are highly reproducible. Some networks were less robust, and were observed less than 70% of the time. Age-related associations were also observed in a selection of networks, including the default-mode network, offering further evidence of development in these networks at an early age. ICA-derived resting-state networks appear to be robust, although some networks should further scrutinized if subjected to group-level statistical analyses, such as spatiotemporal regression. The final set of ICA-derived networks and an age-appropriate T₁ -weighted template are made available to the neuroimaging community, https://www.nitrc.org/projects/genr. Hum Brain Mapp 37:4286-4300, 2016. © 2016 Wiley Periodicals, Inc.

Intrinsic Connectivity Provides the Baseline Framework for Variability in Motor Performance: A Multivariate Fusion Analysis of Low- and High-Frequency Resting-State Oscillations and Antisaccade Performance

Intrinsic brain activity provides the functional framework for the brain's full repertoire of behavioral responses; that is, a common mechanism underlies intrinsic and extrinsic neural activity, with extrinsic activity building upon the underlying baseline intrinsic activity. The generation of a motor movement in response to sensory stimulation is one of the most fundamental functions of the central nervous system. Since saccadic eye movements are among our most stereotyped motor responses, we hypothesized that individual variability in the ability to inhibit a prepotent saccade and make a voluntary antisaccade would be related to individual variability in intrinsic connectivity. Twenty-three individuals completed the antisaccade task and resting-state functional magnetic resonance imaging (fMRI). A multivariate analysis of covariance identified relationships between fMRI oscillations (0.01-0.2 Hz) of resting-state networks determined using high-dimensional independent component analysis and antisaccade performance (latency, error rate). Significant multivariate relationships between antisaccade latency and directional error rate were obtained in independent components across the entire brain. Some of the relationships were obtained in components that overlapped substantially with the task; however, many were obtained in components that showed little overlap with the task. The current results demonstrate that even in the absence of a task, spectral power in regions showing little overlap with task activity predicts an individual's performance on a saccade task.

Test-retest reliability of graph metrics of resting state MRI functional brain networks: A review

The employment of graph theory to analyze spontaneous fluctuations in resting state BOLD fMRI data has become a dominant theme in brain imaging studies and neuroscience. Analysis of resting state functional brain networks based on graph theory has proven to be a powerful tool to quantitatively characterize functional architecture of the brain and it has provided a new platform to explore the overall structure of local and global functional connectivity in the brain. Due to its increased use and possible expansion to clinical use, it is essential that the reliability of such a technique is very strongly assessed. In this review, we explore the outcome of recent studies in network reliability which apply graph theory to analyze connectome resting state networks. Therefore, we investigate which preprocessing steps may affect reproducibility the most. In order to investigate network reliability, we compared the test-retest (TRT) reliability of functional data of published neuroimaging studies with different preprocessing steps. In particular we tested influence of global signal regression, correlation metric choice, binary versus weighted link definition, frequency band selection and length of time-series. Statistical analysis shows that only frequency band selection and length of time-series seem to affect TRT reliability. Our results highlight the importance of the choice of the preprocessing steps to achieve more reproducible measurements.

Challenges in the reproducibility of clinical studies with resting state fMRI: An example in early Parkinson's disease

Resting state fMRI (rfMRI) is gaining in popularity, being easy to acquire and with promising clinical applications. However, rfMRI studies, especially those involving clinical groups, still lack reproducibility, largely due to the different analysis settings. This is particularly important for the development of imaging biomarkers. The aim of this work was to evaluate the reproducibility of our recent study regarding the functional connectivity of the basal ganglia network in early Parkinson's disease (PD) (Szewczyk-Krolikowski et al., 2014). In particular, we systematically analysed the influence of two rfMRI analysis steps on the results: the individual cleaning (artefact removal) of fMRI data and the choice of the set of independent components (template) used for dual regression.

Our experience suggests that the use of a cleaning approach based on single-subject independent component analysis, which removes non neural-related sources of inter-individual variability, can help to increase the reproducibility of clinical findings. A template generated using an independent set of healthy controls is recommended for studies where the aim is to detect differences from a “healthy” brain, rather than an “average” template, derived from an equal number of patients and controls. While, exploratory analyses (e.g. testing multiple resting state networks) should be used to formulate new hypotheses, careful validation is necessary before promising findings can be translated into useful biomarkers.

•

Reproducibility of clinical findings is crucial for imaging biomarker development.

•

We addressed the impact on reproducibility of different analysis settings in rfMRI.

•

ICA-based cleaning of rfMRI data increases reproducibility.

•

The effect of the template choice for dual regression is evaluated.

Disruptions in Resting State Functional Connectivity and Cerebral Blood Flow in Mild Traumatic Brain Injury Patients

Mild traumatic brain injury (mTBI) is often occult to conventional imaging techniques. However, there is growing evidence that mTBI patients who lack evidence of structural intracranial injury may develop post-concussive syndrome (PCS). We investigated longitudinal alterations in resting state functional connectivity (rs-FC) in brain networks in a population of 28 patients compared to 28 matched control participants. Rs-FC and cerebral blood flow (CBF) within the nodes of the Default Mode Network (DMN) and Task Positive Network (TPN) were assessed at three time points including acute, sub-acute, and chronic stages following mTBI. Participants received the Automated Neuropsychological Assessment Metrics (ANAM) to assess cognitive performance. Main findings indicate that despite normalized cognitive performance, chronic mTBI patients demonstrate increased rs-FC between the DMN and regions associated with the salience network (SN) and TPN compared to the control populations, as well as reduced strength of rs-FC within the DMN at the acute stage of injury. In addition, chronic mTBI patients demonstrate an imbalance in the ratio of CBF between nodes of the DMN and TPN. Furthermore, preliminary exploratory analysis suggests that compared to those without chronic PCS, patients with chronic PCS reveal an imbalance in the ratio of CBF between the DMN nodes and TPN nodes across multiple stages of recovery. Findings suggest that the altered network perfusion with the associated changes in rs-FC may be a possible predictor of which mTBI patients will develop chronic PCS.

Investigation of Multiple Frequency Ranges Using Discrete Wavelet Decomposition of Resting-State Functional Connectivity in Mild Traumatic Brain Injury Patients

The aim of this study was to investigate if discrete wavelet decomposition provides additional insight into resting-state processes through the analysis of functional connectivity within specific frequency ranges within the default mode network (DMN) that may be affected by mild traumatic brain injury (mTBI). Participants included 32 mTBI patients (15 with postconcussive syndrome [PCS+] and 17 without [PCS-]). mTBI patients received resting-state functional magnetic resonance imaging (rs-fMRI) at acute (within 10 days of injury) and chronic (6 months postinjury) time points and were compared with 31 controls (healthy control [HC]). The wavelet decomposition divides the time series into multiple frequency ranges based on four scaling factors (SF1: 0.125-0.250 Hz, SF2: 0.060-0.125 Hz, SF3: 0.030-0.060 Hz, SF4: 0.015-0.030 Hz). Within each SF, wavelet connectivity matrices for nodes of the DMN were created for each group (HC, PCS+, PCS-), and bivariate measures of strength and diversity were calculated. The results demonstrate reduced strength of connectivity in PCS+ patients compared with PCS- patients within SF1 during both the acute and chronic stages of injury, as well as recovery of connectivity within SF1 across the two time points. Furthermore, the PCS- group demonstrated greater network strength compared with controls at both time points, suggesting a potential compensatory or protective mechanism in these patients. These findings stress the importance of investigating resting-state connectivity within multiple frequency ranges; however, many of our findings are within SF1, which may overlap with frequencies associated with cardiac and respiratory activities.

Time of acquisition and network stability in pediatric resting-state functional magnetic resonance imaging

Resting-state functional magnetic resonance imaging (rs-fMRI) has been shown to elucidate reliable patterns of brain networks in both children and adults. Studies in adults have shown that rs-fMRI acquisition times of ∼5 to 6 min provide adequate sampling to produce stable spatial maps of a number of different brain networks. However, it is unclear whether the acquisition time directly translates to studies of children. While there are many similarities between the brains of children and adults, many differences are also evident. Children have increased metabolism, differences in brain morphology and connectivity strengths, greater brain plasticity, and increased brain noise. Furthermore, there are differences in physiologic parameters, such as heart and respiratory rates, and compliance of the blood vessels. These developmental differences could translate into different acquisition times for rs-fMRI studies in pediatric populations. Longer scan times, however, increase the subject burden and the risk for greater movement, especially in children. Thus, the goal of this study was to assess the optimum acquisition time of rs-fMRI to extract stable brain networks in school-age children. We utilized fuzzy set theory in 84 six-to-eight year-old children and found that eight networks, including the default mode, salience, frontal, left frontoparietal, right frontoparietal, sensorimotor, auditory, and visual networks, all stabilized after ∼5½ min. The sensorimotor network showed the least stability, whereas the salience and auditory networks showed the greatest stability. A secondary analysis using dual regression confirmed these results. In conclusion, in young children with little head motion, rs-fMRI acquisition times of ∼5½ min can extract the full complement of brain networks.

Static and Dynamic Intrinsic Connectivity following Mild Traumatic Brain Injury

Mild traumatic brain injury (mTBI) is the most common neurological disorder and is typically characterized by temporally limited cognitive impairment and emotional symptoms. Previous examinations of intrinsic resting state networks in mTBI have primarily focused on abnormalities in static functional connectivity, and deficits in dynamic functional connectivity have yet to be explored in this population. Resting-state data was collected on 48 semi-acute (mean = 14 days post-injury) mTBI patients and 48 matched healthy controls. A high-dimensional independent component analysis (N = 100) was utilized to parcellate intrinsic connectivity networks (ICN), with a priori hypotheses focusing on the default-mode network (DMN) and sub-cortical structures. Dynamic connectivity was characterized using a sliding window approach over 126 temporal epochs, with standard deviation serving as the primary outcome measure. Finally, distribution-corrected z-scores (DisCo-Z) were calculated to investigate changes in connectivity in a spatially invariant manner on a per-subject basis. Following appropriate correction for multiple comparisons, no significant group differences were evident on measures of static or dynamic connectivity within a priori ICN. Reduced (HC > mTBI patients) static connectivity was observed in the DMN at uncorrected (p < 0.005) thresholds. Finally, a trend (p = 0.07) for decreased dynamic connectivity in patients across all ICN was observed during spatially invariant analyses (DisCo-Z). In the semi-acute phase of recovery, mTBI was not reliably associated with abnormalities in static or dynamic functional connectivity within the DMN or sub-cortical structures.

Resting state BOLD functional connectivity at 3T: spin echo versus gradient echo EPI

Previous evidence showed that, due to refocusing of static dephasing effects around large vessels, spin-echo (SE) BOLD signals offer an increased linearity and promptness with respect to gradient-echo (GE) acquisition, even at low field. These characteristics suggest that, despite the reduced sensitivity, SE fMRI might also provide a potential benefit when investigating spontaneous fluctuations of brain activity. However, there are no reports on the application of spin-echo fMRI for connectivity studies at low field. In this study we compared resting state functional connectivity as measured with GE and SE EPI sequences at 3T. Main results showed that, within subject, the GE sensitivity is overall larger with respect to that of SE, but to a less extent than previously reported for activation studies. Noteworthy, the reduced sensitivity of SE was counterbalanced by a reduced inter-subject variability, resulting in comparable group statistical connectivity maps for the two sequences. Furthermore, the SE method performed better in the ventral portion of the default mode network, a region affected by signal dropout in standard GE acquisition. Future studies should clarify if these features of the SE BOLD signal can be beneficial to distinguish subtle variations of functional connectivity across different populations and/or treatments when vascular confounds or regions affected by signal dropout can be a critical issue.

Evaluating the reliability of different preprocessing steps to estimate graph theoretical measures in resting state fMRI data

With resting-state functional MRI (rs-fMRI) there are a variety of post-processing methods that can be used to quantify the human brain connectome. However, there is also a choice of which preprocessing steps will be used prior to calculating the functional connectivity of the brain. In this manuscript, we have tested seven different preprocessing schemes and assessed the reliability between and reproducibility within the various strategies by means of graph theoretical measures. Different preprocessing schemes were tested on a publicly available dataset, which includes rs-fMRI data of healthy controls. The brain was parcellated into 190 nodes and four graph theoretical (GT) measures were calculated; global efficiency (GEFF), characteristic path length (CPL), average clustering coefficient (ACC), and average local efficiency (ALE). Our findings indicate that results can significantly differ based on which preprocessing steps are selected. We also found dependence between motion and GT measurements in most preprocessing strategies. We conclude that by using censoring based on outliers within the functional time-series as a processing, results indicate an increase in reliability of GT measurements with a reduction of the dependency of head motion.

Modulation of resting state functional connectivity of the motor network by transcranial pulsed current stimulation

The effects of transcranial pulsed current stimulation (tPCS) on resting state functional connectivity (rs-FC) within the motor network were investigated. Eleven healthy participants received one magnetic resonance imaging (MRI) session with three resting state functional MRI (rs-fMRI) scans, one before stimulation (PRE-STIM) to collect baseline measures, one during stimulation (STIM), and one after 13 min of stimulation (POST-STIM). Rs-FC measures during the STIM and POST-STIM conditions were compared to the PRE-STIM baseline. Regions of interest for the rs-FC analysis were extracted from the significantly activated clusters obtained during a finger tapping motor paradigm and included the right primary motor cortex (R M1), left primary motor cortex (L M1), supplemental motor area (SMA), and cerebellum (Cer). The main findings were reduced rs-FC between the left M1 and surrounding motor cortex, and increased rs-FC between the left M1 and left thalamus during stimulation, but increased rs-FC between the Cer and right insula after stimulations. Bivariate measures of connectivity demonstrate reduced strength of connectivity for the whole network average (p=0.044) and reduced diversity of connectivity for the network average during stimulation (p=0.024). During the POST-STIM condition, the trend of reduced diversity for the network average was statistically weaker (p=0.071). In conclusion, while many of the findings are comparable to previous reports using simultaneous transcranial direct current stimulation (tDCS) and fMRI acquisition, we also demonstrate additional changes in connectivity patterns that are induced by tPCS.

The utility of data-driven feature selection: re: Chu et al. 2012

The recent Chu et al. (2012) manuscript discusses two key findings regarding feature selection (FS): (1) data driven FS was no better than using whole brain voxel data and (2) a priori biological knowledge was effective to guide FS. Use of FS is highly relevant in neuroimaging-based machine learning, as the number of attributes can greatly exceed the number of exemplars. We strongly endorse their demonstration of both of these findings, and we provide additional important practical and theoretical arguments as to why, in their case, the data-driven FS methods they implemented did not result in improved accuracy. Further, we emphasize that the data-driven FS methods they tested performed approximately as well as the all-voxel case. We discuss why a sparse model may be favored over a complex one with similar performance. We caution readers that the findings in the Chu et al. report should not be generalized to all data-driven FS methods.