Nuisance Regression of High-Frequency Functional Magnetic Resonance Imaging Data: Denoising Can Be Noisy

Unraveling the mesoscale functional connectivity of the human primary visual cortex using high-resolution functional MRI

Despite their importance in shaping visual perception, functional connectivity between ocular dominance columns (ODCs), the building blocks of neuronal processing within the human primary visual cortex (V1), remains largely unknown. Using high-resolution functional MRI (fMRI), we localized ODCs and assessed their resting-state functional connectivity (rs-FC) in 11 human adults (3 females). Consistent with anatomical studies in animals, we found stronger rs-FC in the middle compared to deep and superficial cortical depths and selectively stronger rs-FC between ODCs with alike compared to unalike ocular polarity. Beyond what was known from animal models, and consistent with human perceptual biases, we found stronger intra- and interhemispheric rs-FC in peripheral (compared to central) and in dorsal (compared to ventral) V1 subregions. Lastly, we found a significant correlation between rs-FC patterns and ODC maps, supporting the hypothesis that ODC maps can be predicted from rs-FC patterns within V1. These results highlight the heterogeneity in functional connectivity between ODCs across cortical depths and V1 subfields, underscoring their likely association with human perceptual biases.

Significance Statement

Our findings provide evidence for selective mesoscale rs-FC between ODCs, aligning with prior anatomical findings in animals.Beyond what is known from animal studies, we demonstrate that the mesoscale rs-FC pattern varies across V1 subregions, aligning with the expected heterogeneity in global visual processing across visual subfields.We provide evidence for the predictability of ODC maps from the rs-FC pattern, establishing one of the first steps toward leveraging rs-FC for segmentation of the visual cortex at mesoscale levels.

Brainwide hemodynamics predict neural rhythms across sleep and wakefulness in humans

The brain exhibits rich oscillatory dynamics that play critical roles in vigilance and cognition, such as the neural rhythms that define sleep. These rhythms continuously fluctuate, signaling major changes in vigilance, but the brainwide dynamics underlying these oscillations are unknown. Using simultaneous EEG and fast fMRI in humans drifting between sleep and wakefulness, we developed a machine learning approach to investigate which brainwide fMRI networks predict alpha (8-12 Hz) and delta (1-4 Hz) fluctuations. We predicted moment-to-moment EEG power variations from fMRI activity in held-out subjects, and found that information about alpha rhythms was highly separable in two networks linked to arousal and visual systems. Conversely, delta rhythms were diffusely represented on a large spatial scale across the cortex. These results identify the large-scale network patterns that underlie alpha and delta rhythms, and establish a novel framework for investigating multimodal, brainwide dynamics.

Functional connectivity differences of the olfactory network in Parkinson's Disease, mild cognitive impairment and cognitively normal individuals: A resting-state fMRI study

Olfactory dysfunction is an early sign of such neurodegenerative diseases as Parkinson's (PD) and Alzheimer's (AD), and is often present in Mild Cognitive Impairment (MCI), a precursor of AD. Understanding neuro-temporal relationships, i.e., functional connectivity, between olfactory eloquent structures in such disorders, could shed light on their basic pathophysiology. To this end, we employed region-based analyses using resting-state functional magnetic resonance imaging (rs-fMRI) obtained from cognitively normal (CN), MCI, and PD patients with cognitive impairment (PD-CogImp). Using machine learning (linear and ensemble learning), we determined whether the identified functional patterns could classify abnormal function from normal function. Olfaction, as measured by objective testing, was found to be most strongly associated with diagnostic status, emphasizing the fundamental association of this primary sensory system with these conditions. Consistently lower functional connectivity was observed in the PD-CogImp cohort compared to the CN cohort among all identified brain regions. Differences were also found between PD-CogImp and MCI at the level of the orbitofrontal and cingulate cortices. MCI and CN subjects had different functional connectivity between the posterior orbitofrontal cortex and thalamus. Regardless of study group, males showed significantly higher connectivity than females in connections involving the orbitofrontal cortex. The logistic regression model trained using the top discriminatory features revealed that caudate was the most involved olfaction-related brain structure (accuracy = 0.88, Area under the Receiver operator characteristic curve of 0.90). In aggregate, our study demonstrates that resting functional connectivity among olfactory eloquent structures has potential value in better understanding the pathophysiology of several neurodegenerative diseases.

Web-based processing of physiological noise in fMRI: addition of the PhysIO toolbox to CBRAIN

Neuroimaging research requires sophisticated tools for analyzing complex data, but efficiently leveraging these tools can be a major challenge, especially on large datasets. CBRAIN is a web-based platform designed to simplify the use and accessibility of neuroimaging research tools for large-scale, collaborative studies. In this paper, we describe how CBRAIN's unique features and infrastructure were leveraged to integrate TAPAS PhysIO, an open-source MATLAB toolbox for physiological noise modeling in fMRI data. This case study highlights three key elements of CBRAIN's infrastructure that enable streamlined, multimodal tool integration: a user-friendly GUI, a Brain Imaging Data Structure (BIDS) data-entry schema, and convenient in-browser visualization of results. By incorporating PhysIO into CBRAIN, we achieved significant improvements in the speed, ease of use, and scalability of physiological preprocessing. Researchers now have access to a uniform and intuitive interface for analyzing data, which facilitates remote and collaborative evaluation of results. With these improvements, CBRAIN aims to become an essential open-science tool for integrative neuroimaging research, supporting FAIR principles and enabling efficient workflows for complex analysis pipelines.

Functional ultrasound detects frequency-specific acute and delayed S-ketamine effects in the healthy mouse brain

Introduction

S-ketamine has received great interest due to both its antidepressant effects and its potential to induce psychosis when administered subchronically. However, no studies have investigated both its acute and delayed effects using in vivo small-animal imaging. Recently, functional ultrasound (fUS) has emerged as a powerful alternative to functional magnetic resonance imaging (fMRI), outperforming it in sensitivity and in spatiotemporal resolution. In this study, we employed fUS to thoroughly characterize acute and delayed S-ketamine effects on functional connectivity (FC) within the same cohort at slow frequency bands ranging from 0.01 to 1.25 Hz, previously reported to exhibit FC.

Methods

We acquired fUS in a total of 16 healthy C57/Bl6 mice split in two cohorts (n = 8 received saline, n = 8 S-ketamine). One day after the first scans, performed at rest, the mice received the first dose of S-ketamine during the second measurement, followed by four further doses administered every 2 days. First, we assessed FC reproducibility and reliability at baseline in six frequency bands. Then, we investigated the acute and delayed effects at day 1 after the first dose and at day 9, 1 day after the last dose, for all bands, resulting in a total of four fUS measurements for every mouse.

Results

We found reproducible (r > 0.9) and reliable (r > 0.9) group-average readouts in all frequency bands, only the 0.01-0.27 Hz band performing slightly worse. Acutely, S-ketamine induced strong FC increases in five of the six bands, peaking in the 0.073-0.2 Hz band. These increases comprised both cortical and subcortical brain areas, yet were of a transient nature, FC almost returning to baseline levels towards the end of the scan. Intriguingly, we observed robust corticostriatal FC decreases in the fastest band acquired (0.75 Hz-1.25  Hz). These changes persisted to a weaker extent after 1 day and at this timepoint they were accompanied by decreases in the other five bands as well. After 9 days, the decreases in the 0.75-1.25 Hz band were maintained, however no changes between cohorts could be detected in any other bands.

Discussion

In summary, the study reports that acute and delayed ketamine effects in mice are not only dissimilar but have different directionalities in most frequency bands. The complementary readouts of the employed frequency bands recommend the use of fUS for frequency-specific investigation of pharmacological effects on FC.

Neuroimaging brain growth charts: A road to mental health

Mental disorders are common health concerns and contribute to a heavy global burden on our modern society. It is challenging to identify and treat them timely. Neuroimaging evidence suggests the incidence of various psychiatric and behavioral disorders is closely related to the atypical development of brain structure and function. The identification and understanding of atypical brain development provide chances for clinicians to detect mental disorders earlier, perhaps even prior to onset, and treat them more precisely. An invaluable and necessary method in identifying and monitoring atypical brain development are growth charts of typically developing individuals in the population. The brain growth charts can offer a series of standard references on typical neurodevelopment, representing an important resource for the scientific and medical communities. In the present paper, we review the relationship between mental disorders and atypical brain development from a perspective of normative brain development by surveying the recent progress in the development of brain growth charts, including four aspects on growth chart utility: 1) cohorts, 2) measures, 3) mechanisms, and 4) clinical translations. In doing so, we seek to clarify the challenges and opportunities in charting brain growth, and to promote the application of brain growth charts in clinical practice.

Investigating mechanisms of fast BOLD responses: The effects of stimulus intensity and of spatial heterogeneity of hemodynamics

Recent studies have demonstrated that fast fMRI can track neural activity well above the temporal limit predicted by the canonical hemodynamic response model. While these findings are promising, the biophysical mechanisms underlying these fast fMRI phenomena remain underexplored. In this study, we discuss two aspects of the hemodynamic response, complementary to several existing hypotheses, that can accommodate faster fMRI dynamics beyond those predicted by the canonical model. First, we demonstrate, using both visual and somatosensory paradigms, that the timing and shape of hemodynamic response functions (HRFs) vary across graded levels of stimulus intensity-with lower-intensity stimulation eliciting faster and narrower HRFs. Second, we show that as the spatial resolution of fMRI increases, voxel-wise HRFs begin to deviate from the canonical model, with a considerable portion of voxels exhibiting faster temporal dynamics than predicted by the canonical HRF. Collectively, both stimulus/task intensity and image resolution can affect the sensitivity of fMRI to fast brain activity, which may partly explain recent observations of fast fMRI signals. It is further noteworthy that, while the present investigations focus on fast neural responses, our findings suggest that a revised hemodynamic model may benefit the many fMRI studies using paradigms with wide ranges of contrast levels (e.g., resting or naturalistic conditions) or with modern, high-resolution MR acquisitions.

Statistical power or more precise insights into neuro-temporal dynamics? Assessing the benefits of rapid temporal sampling in fMRI

Functional magnetic resonance imaging (fMRI), a non-invasive and widely used human neuroimaging method, is most known for its spatial precision. However, there is a growing interest in its temporal sensitivity. This is despite the temporal blurring of neuronal events by the blood oxygen level dependent (BOLD) signal, the peak of which lags neuronal firing by 4-6 seconds. Given this, the goal of this review is to answer a seemingly simple question - "What are the benefits of increased temporal sampling for fMRI?". To answer this, we have combined fMRI data collected at multiple temporal scales, from 323 to 1000 milliseconds, with a review of both historical and contemporary temporal literature. After a brief discussion of technological developments that have rekindled interest in temporal research, we next consider the potential statistical and methodological benefits. Most importantly, we explore how fast fMRI can uncover previously unobserved neuro-temporal dynamics - effects that are entirely missed when sampling at conventional 1 to 2 second rates. With the intrinsic link between space and time in fMRI, this temporal renaissance also delivers improvements in spatial precision. Far from producing only statistical gains, the array of benefits suggest that the continued temporal work is worth the effort.

Energy-Period Profiles of Brain Networks in Group fMRI Resting-State Data: A Comparison of Empirical Mode Decomposition With the Short-Time Fourier Transform and the Discrete Wavelet Transform

Traditionally, functional networks in resting-state data were investigated with linear Fourier and wavelet-related methods to characterize their frequency content by relying on pre-specified frequency bands. In this study, Empirical Mode Decomposition (EMD), an adaptive time-frequency method, is used to investigate the naturally occurring frequency bands of resting-state data obtained by Group Independent Component Analysis. Specifically, energy-period profiles of Intrinsic Mode Functions (IMFs) obtained by EMD are created and compared for different resting-state networks. These profiles have a characteristic distribution for many resting-state networks and are related to the frequency content of each network. A comparison with the linear Short-Time Fourier Transform (STFT) and the Maximal Overlap Discrete Wavelet Transform (MODWT) shows that EMD provides a more frequency-adaptive representation of different types of resting-state networks. Clustering of resting-state networks based on the energy-period profiles leads to clusters of resting-state networks that have a monotone relationship with frequency and energy. This relationship is strongest with EMD, intermediate with MODWT, and weakest with STFT. The identification of these relationships suggests that EMD has significant advantages in characterizing brain networks compared to STFT and MODWT. In a clinical application to early Parkinson's disease (PD) vs. normal controls (NC), energy and period content were studied for several common resting-state networks. Compared to STFT and MODWT, EMD showed the largest differences in energy and period between PD and NC subjects. Using a support vector machine, EMD achieved the highest prediction accuracy in classifying NC and PD subjects among STFT, MODWT, and EMD.

Modular and state-relevant functional network connectivity in high-frequency eyes open vs eyes closed resting fMRI data

BACKGROUND

Resting-state fMRI (rs-fMRI) is employed to assess "functional connections" of signal between brain regions. However, multiple rs-fMRI paradigms and data-filtering strategies have been used, highlighting the need to explore BOLD signal across the spectrum. Rs-fMRI data is typically filtered at frequencies ranging between 0.008∼0.2 Hz to mitigate nuisance signal (e.g. cardiac, respiratory) and maximize neuronal BOLD signal. However, some argue neuronal BOLD signal may be parsed at higher frequencies.

NEW METHOD

To assess the contributions of rs-fMRI along the BOLD spectra on functional network connectivity (FNC) matrices, a spatially constrained independent component analysis (ICA) was performed at seven different frequency "bins", after which FNC values and FNC measures of matrix-randomness were assessed using linear mixed models.

RESULTS

Results show FNCs at higher-frequency bins display similar randomness to those from the typical frequency bins (0.01-0.15), while the largest values are in the 0.31-0.46 Hz bin. Eyes open (EO) vs eyes closed (EC) comparison found EC was less random than EO across most frequency bins. Further, FNC was greater in EC across auditory and cognitive control pairings while EO values were greater in somatomotor, visual, and default mode FNC.

COMPARISON WITH EXISTING METHODS

Effect sizes of frequency and resting-state paradigm vary from small to large, but the most notable results are specific to frequency ranges and resting-state paradigm with artifacts like motion displaying negligible effect sizes.

CONCLUSIONS

These suggest unique information may be derived from FNC matrices across frequencies and paradigms, but additional data is necessary prior to any definitive conclusions.

Integrated and segregated frequency architecture of the human brain network

The frequency of brain activity modulates the relationship between the brain and human behavior. Insufficient understanding of frequency-specific features may thus lead to inconsistent explanations of human behavior. However, to date, the frequency-specific features of the human brain functional network at the whole-brain level remain poorly understood. Here, we used resting-state fMRI data and graph-theory analyses to investigate the frequency-specific characteristics of fMRI signals in 12 frequency bands (frequency range 0.01-0.7 Hz) in 75 healthy participants. We found that brain regions with higher level and more complex functions had a more variable functional connectivity pattern but engaged less in higher frequency ranges. Moreover, brain regions that engaged in fewer frequency bands played more integrated roles (i.e., higher network participation coefficient and lower within-module degree) in the functional network, whereas regions that engaged in broader frequency ranges exhibited more segregated functions (i.e., lower network participation coefficient and higher within-module degree). Finally, behavioral analyses revealed that regional frequency variability was associated with a spectrum of behavioral functions from sensorimotor functions to complex cognitive and social functions. Taken together, our results showed that segregated functions are executed in wide frequency ranges, whereas integrated functions are executed mainly in lower frequency ranges. These frequency-specific features of brain networks provided crucial insights into the frequency mechanism of fMRI signals, suggesting that signals in higher frequency ranges should be considered for their relation to cognitive functions.

Frequency-specific task modulation of human brain functional networks: A fast fMRI study

How coherent neural oscillations are involved in task execution is a fundamental question in neuroscience. Although several electrophysiological studies have tackled this issue, the brain-wide task modulation of neural coherence remains uncharacterized. Here, with a fast fMRI technique, we studied shifts of brain-wide neural coherence across different task states in the ultraslow frequency range (0.01-0.7 Hz). First, we examined whether the shifts of the brain-wide neural coherence occur in a frequency-dependent manner. We quantified the shift of a region's average neural coherence by the inter-state variance of the mean coherence between the region and the rest of the brain. A clustering analysis based on the variance's spatial correlation between frequency components revealed four frequency bands (0.01-0.15 Hz, 0.15-0.37 Hz, 0.37-0.53 Hz, and 0.53-0.7 Hz) showing band-specific shifts of the brain-wide neural coherence. Next, we investigated the similarity of the inter-state variance's spectra between all pairs of regions. We found that regions showing similar spectra correspond to those forming functional modules of the brain network. Then, we investigated the relationship between identified frequency bands and modules' inter-state variances. We found that modules showing the highest variance are those made up of parieto-occipital regions at 0.01-0.15 Hz, while it is replaced with another consisting of frontal regions above 0.15 Hz. Furthermore, these modules showed specific shifting patterns of the mean coherence across states at 0.01-0.15 Hz and above 0.15 Hz, suggesting that identified frequency bands differentially contribute to neural interactions during task execution. Our results highlight that usage of the fast fMRI enables brain-wide investigation of neural coherence up to 0.7 Hz, which opens a promising track for assessment of the large-scale neural interactions in the ultraslow frequency range.

Disentangling time series between brain tissues improves fMRI data quality using a time-dependent deep neural network

Functional MRI (fMRI) is a prominent imaging technique to probe brain function, however, a substantial proportion of noise from multiple sources influences the reliability and reproducibility of fMRI data analysis and limits its clinical applications. Extensive effort has been devoted to improving fMRI data quality, but in the last two decades, there is no consensus reached which technique is more effective. In this study, we developed a novel deep neural network for denoising fMRI data, named denoising neural network (DeNN). This deep neural network is 1) applicable without requiring externally recorded data to model noise; 2) spatially and temporally adaptive to the variability of noise in different brain regions at different time points; 3) automated to output denoised data without manual interference; 4) trained and applied on each subject separately and 5) insensitive to the repetition time (TR) of fMRI data. When we compared DeNN with a number of nuisance regression methods for denoising fMRI data from Alzheimer's Disease Neuroimaging Initiative (ADNI) database, only DeNN had connectivity for functionally uncorrelated regions close to zero and successfully identified unbiased correlations between the posterior cingulate cortex seed and multiple brain regions within the default mode network or task positive network. The whole brain functional connectivity maps computed with DeNN-denoised data are approximately three times as homogeneous as the functional connectivity maps computed with raw data. Furthermore, the improved homogeneity strengthens rather than weakens the statistical power of fMRI in detecting intrinsic functional differences between cognitively normal subjects and subjects with Alzheimer's disease.

Tools of the trade: estimating time-varying connectivity patterns from fMRI data

Given the dynamic nature of the brain, there has always been a motivation to move beyond 'static' functional connectivity, which characterizes functional interactions over an extended period of time. Progress in data acquisition and advances in analytical neuroimaging methods now allow us to assess the whole brain's dynamic functional connectivity (dFC) and its network-based analog, dynamic functional network connectivity at the macroscale (mm) using fMRI. This has resulted in the rapid growth of analytical approaches, some of which are very complex, requiring technical expertise that could daunt researchers and neuroscientists. Meanwhile, making real progress toward understanding the association between brain dynamism and brain disorders can only be achieved through research conducted by domain experts, such as neuroscientists and psychiatrists. This article aims to provide a gentle introduction to the application of dFC. We first explain what dFC is and the circumstances under which it can be used. Next, we review two major categories of analytical approaches to capture dFC. We discuss caveats and considerations in dFC analysis. Finally, we walk readers through an openly accessible toolbox to capture dFC properties and briefly review some of the dynamic metrics calculated using this toolbox.

Incorporating structured assumptions with probabilistic graphical models in fMRI data analysis

With the wide adoption of functional magnetic resonance imaging (fMRI) by cognitive neuroscience researchers, large volumes of brain imaging data have been accumulated in recent years. Aggregating these data to derive scientific insights often faces the challenge that fMRI data are high-dimensional, heterogeneous across people, and noisy. These challenges demand the development of computational tools that are tailored both for the neuroscience questions and for the properties of the data. We review a few recently developed algorithms in various domains of fMRI research: fMRI in naturalistic tasks, analyzing full-brain functional connectivity, pattern classification, inferring representational similarity and modeling structured residuals. These algorithms all tackle the challenges in fMRI similarly: they start by making clear statements of assumptions about neural data and existing domain knowledge, incorporate those assumptions and domain knowledge into probabilistic graphical models, and use those models to estimate properties of interest or latent structures in the data. Such approaches can avoid erroneous findings, reduce the impact of noise, better utilize known properties of the data, and better aggregate data across groups of subjects. With these successful cases, we advocate wider adoption of explicit model construction in cognitive neuroscience. Although we focus on fMRI, the principle illustrated here is generally applicable to brain data of other modalities.

Resting-state "physiological networks"

Slow changes in systemic brain physiology can elicit large fluctuations in fMRI time series, which manifest as structured spatial patterns of temporal correlations between distant brain regions. Here, we investigated whether such "physiological networks"-sets of segregated brain regions that exhibit similar responses following slow changes in systemic physiology-resemble patterns associated with large-scale networks typically attributed to remotely synchronized neuronal activity. By analyzing a large group of subjects from the 3T Human Connectome Project (HCP) database, we demonstrate brain-wide and noticeably heterogenous dynamics tightly coupled to either respiratory variation or heart rate changes. We show, using synthesized data generated from physiological recordings across subjects, that these physiologically-coupled fluctuations alone can produce networks that strongly resemble previously reported resting-state networks, suggesting that, in some cases, the "physiological networks" seem to mimic the neuronal networks. Further, we show that such physiologically-relevant connectivity estimates appear to dominate the overall connectivity observations in multiple HCP subjects, and that this apparent "physiological connectivity" cannot be removed by the use of a single nuisance regressor for the entire brain (such as global signal regression) due to the clear regional heterogeneity of the physiologically-coupled responses. Our results challenge previous notions that physiological confounds are either localized to large veins or globally coherent across the cortex, therefore emphasizing the necessity to consider potential physiological contributions in fMRI-based functional connectivity studies. The rich spatiotemporal patterns carried by such "physiological" dynamics also suggest great potential for clinical biomarkers that are complementary to large-scale neuronal networks.

Current methods and new directions in resting state fMRI

Resting state functional connectivity magnetic resonance imaging (rsfcMRI) has become a key component of investigations of neurocognitive and psychiatric behaviors. Over the past two decades, several methods and paradigms have been adopted to utilize and interpret data from resting-state fluctuations in the brain. These findings have increased our understanding of changes in many disease states. As the amount of resting state data available for research increases with big datasets and data-sharing projects, it is important to review the established traditional analysis methods and recognize areas where research methodology can be adapted to better accommodate the scale and complexity of rsfcMRI analysis. In this paper, we review established methods of analysis as well as areas that have been receiving increasing attention such as dynamic rsfcMRI, independent vector analysis, multiband rsfcMRI and network of networks.

Model-based physiological noise removal in fast fMRI

Recent improvements in the speed and sensitivity of fMRI acquisition techniques suggest that fast fMRI can be used to detect and precisely localize sub-second neural dynamics. This enhanced temporal resolution has enormous potential for neuroscientists. However, physiological noise poses a major challenge for the analysis of fast fMRI data. Physiological noise scales with sensitivity, and its autocorrelation structure is altered in rapidly sampled data, suggesting that new approaches are needed for physiological noise removal in fast fMRI. Existing strategies either rely on external physiological recordings, which can be noisy or difficult to collect, or employ data-driven approaches which make assumptions that may not hold true in fast fMRI. We created a statistical model of harmonic regression with autoregressive noise (HRAN) to estimate and remove cardiac and respiratory noise from the fMRI signal directly. This technique exploits the fact that cardiac and respiratory noise signals are fully sampled (rather than aliasing) when imaging at fast rates, allowing us to track and model physiology over time without requiring external physiological measurements. We then created a joint model of neural hemodynamics, and physiological and autocorrelated noise to more accurately remove noise. We first verified that HRAN accurately estimates cardiac and respiratory dynamics and that our model demonstrates goodness-of-fit in fast fMRI data. In task-driven data, we then demonstrated that HRAN is able to remove physiological noise while leaving the neural signal intact, thereby increasing detection of task-driven voxels. Finally, we established that in both simulations and fast fMRI data HRAN is able to improve statistical inferences as compared with gold-standard physiological noise removal techniques. In conclusion, we created a tool that harnesses the novel information in fast fMRI to remove physiological noise, enabling broader use of the technology to study human brain function.

The Cerebral Cortex is Bisectionally Segregated into Two Fundamentally Different Functional Units of Gyri and Sulci

The human cerebral cortex is highly folded into diverse gyri and sulci. Accumulating evidences suggest that gyri and sulci exhibit anatomical, morphological, and connectional differences. Inspired by these evidences, we performed a series of experiments to explore the frequency-specific differences between gyral and sulcal neural activities from resting-state and task-based functional magnetic resonance imaging (fMRI) data. Specifically, we designed a convolutional neural network (CNN) based classifier, which can differentiate gyral and sulcal fMRI signals with reasonable accuracies. Further investigations of learned CNN models imply that sulcal fMRI signals are more diverse and more high frequency than gyral signals, suggesting that gyri and sulci truly play different functional roles. These differences are significantly associated with axonal fiber wiring and cortical thickness patterns, suggesting that these differences might be deeply rooted in their structural and cellular underpinnings. Further wavelet entropy analyses demonstrated the validity of CNN-based findings. In general, our collective observations support a new concept that the cerebral cortex is bisectionally segregated into 2 functionally different units of gyri and sulci.

Erroneous Resting-State fMRI Connectivity Maps Due to Prolonged Arterial Arrival Time and How to Fix Them

In resting-state functional MRI (rs-fMRI), functional networks are assessed utilizing the temporal correlation between spontaneous blood oxygen level-dependent signal fluctuations of spatially remote brain regions. Recently, several groups have shown that temporal shifts are present in rs-fMRI maps in patients with cerebrovascular disease due to spatial differences in arterial arrival times, and that this can be exploited to map arrival times in the brain. This suggests that rs-fMRI connectivity mapping may be similarly sensitive to such temporal shifts, and that standard rs-fMRI analysis methods may fail to identify functional connectivity networks. To investigate this, we studied the default mode network (DMN) in Moyamoya disease patients and compared it with normal healthy volunteers. Our results show that using standard independent component analysis (ICA) and seed-based approaches, arterial arrival delays lead to inaccurate incomplete characterization of functional connectivity within the DMN in Moyamoya disease patients. Furthermore, we propose two techniques to correct these errors, for seed-based and ICA methods, respectively. Using these methods, we demonstrate that it is possible to mitigate the deleterious effects of arterial arrival time on the assessment of functional connectivity of the DMN. As these corrections have not been applied to the vast majority of >200 prior rs-fMRI studies in patients with cerebrovascular disease, we suggest that they be interpreted with great caution. Correction methods should be applied in any rs-fMRI connectivity study of subjects expected to have abnormally delayed arterial arrival times.

Gastric stimulation drives fast BOLD responses of neural origin

Functional magnetic resonance imaging (fMRI) is commonly thought to be too slow to capture any neural dynamics faster than 0.1 Hz. However, recent findings demonstrate the feasibility of detecting fMRI activity at higher frequencies beyond 0.2 Hz. The origin, reliability, and generalizability of fast fMRI responses are still under debate and await confirmation through animal experiments with fMRI and invasive electrophysiology. Here, we acquired single-echo and multi-echo fMRI, as well as local field potentials, from anesthetized rat brains given gastric electrical stimulation modulated at 0.2, 0.4 and 0.8 Hz. Such gastric stimuli could drive widespread fMRI responses at corresponding frequencies from the somatosensory and cingulate cortices. Such fast fMRI responses were linearly dependent on echo times and thus indicative of blood oxygenation level dependent nature (BOLD). Local field potentials recorded during the same gastric stimuli revealed transient and phase-locked broadband neural responses, preceding the fMRI responses by as short as 0.5 s. Taken together, these results suggest that gastric stimulation can drive widespread and rapid fMRI responses of BOLD and neural origin, lending support to the feasibility of using fMRI to detect rapid changes in neural activity up to 0.8 Hz under visceral stimulation.

On the analysis of rapidly sampled fMRI data

Recent advances in parallel imaging and simultaneous multi-slice techniques have permitted whole-brain fMRI acquisitions at sub-second sampling intervals, without significantly sacrificing the spatial coverage and resolution. Apart from probing brain function at finer temporal scales, faster sampling rates may potentially lead to enhanced functional sensitivity, owing possibly to both cleaner neural representations (due to less aliased physiological noise) and additional statistical benefits (due to more degrees of freedom for a fixed scan duration). Accompanying these intriguing aspects of fast acquisitions, however, confusion has also arisen regarding (1) how to preprocess/analyze these fast fMRI data, and (2) what exactly is the extent of benefits with fast acquisitions, i.e., how fast is fast enough for a specific research aim? The first question is motivated by the altered spectral distribution and noise characteristics at short sampling intervals, while the second question seeks to reconcile the complicated trade-offs between the functional contrast-to-noise ratio and the effective degrees of freedom. Although there have been recent efforts to empirically approach different aspects of these two questions, in this work we discuss, from a theoretical perspective accompanied by some illustrative, proof-of-concept experimental in vivo human fMRI data, a few considerations that are rarely mentioned, yet are important for both preprocessing and optimizing statistical inferences for studies that employ acquisitions with sub-second sampling intervals. Several summary recommendations include concerns regarding advisability of relying on low-pass filtering to de-noise physiological contributions, employment of statistical models with sufficient complexity to account for the substantially increased serial correlation, and cautions regarding using rapid sampling to enhance functional sensitivity given that different analysis models may associate with distinct trade-offs between contrast-to-noise ratios and the effective degrees of freedom. As an example, we demonstrate that as TR shortens, the intrinsic differences in how noise is accommodated in general linear models and Pearson correlation analyses (assuming Gaussian distributed stochastic signals and noise) can result in quite different outcomes, either gaining or losing statistical power.

Beware detrending: Optimal preprocessing pipeline for low-frequency fluctuation analysis

Resting‐state functional magnetic resonance imaging (rs‐fMRI) offers the possibility to assess brain function independent of explicit tasks and individual performance. This absence of explicit stimuli in rs‐fMRI makes analyses more susceptible to nonneural signal fluctuations than task‐based fMRI. Data preprocessing is a critical procedure to minimise contamination by artefacts related to motion and physiology. We herein investigate the effects of different preprocessing strategies on the amplitude of low‐frequency fluctuations (ALFFs) and its fractional counterpart, fractional ALFF (fALFF). Sixteen artefact reduction schemes based on nuisance regression are applied to data from 82 subjects acquired at 1.5 T, 30 subjects at 3 T, and 23 subjects at 7 T, respectively. In addition, we examine test–retest variance and effects of bias correction. In total, 569 data sets are included in this study. Our results show that full artefact reduction reduced test–retest variance by up to 50%. Polynomial detrending of rs‐fMRI data has a positive effect on group‐level t‐values for ALFF but, importantly, a negative effect for fALFF. We show that the normalisation process intrinsic to fALFF calculation causes the observed reduction and introduce a novel measure for low‐frequency fluctuations denoted as high‐frequency ALFF (hfALFF). We demonstrate that hfALFF values are not affected by the negative detrending effects seen in fALFF data. Still, highest grey matter (GM) group‐level t‐values were obtained for fALFF data without detrending, even when compared to an exploratory detrending approach based on autocorrelation measures. From our results, we recommend the use of full nuisance regression including polynomial detrending in ALFF data, but to refrain from using polynomial detrending in fALFF data. Such optimised preprocessing increases GM group‐level t‐values by up to 60%.

A window-less approach for capturing time-varying connectivity in fMRI data reveals the presence of states with variable rates of change

Functional connectivity during the resting state has been shown to change over time (i.e., has a dynamic connectivity). However, resting-state fluctuations, in contrast to task-based experiments, are not initiated by an external stimulus. Consequently, a more complicated method needs to be designed to measure the dynamic connectivity. Previous approaches have been based on assumptions regarding the nature of the underlying dynamic connectivity to compensate for this knowledge gap. The most common assumption is what we refer to as locality assumption. Under a locality assumption, a single connectivity state can be estimated from data that are close in time. This assumption is so natural that it has been either explicitly or implicitly embedded in many current approaches to capture dynamic connectivity. However, an important drawback of methods using this assumption is they are unable to capture dynamic changes in connectivity beyond the embedded rate while, there has been no evidence that the rate of change in brain connectivity matches the rates enforced by this assumption. In this study, we propose an approach that enables us to capture functional connectivity with arbitrary rates of change, varying from very slow to the theoretically maximum possible rate of change, which is only imposed by the sampling rate of the imaging device. This method allows us to observe unique patterns of connectivity that were not observable with previous approaches. As we explain further, these patterns are also significantly correlated to the age and gender of study subjects, which suggests they are also neurobiologically related.

On the detection of high frequency correlations in resting state fMRI

Current studies of resting-state connectivity rely on coherent signal fluctuations at frequencies below 0.1 Hz, however, recent studies using high-speed fMRI have shown that fluctuations above 0.5 Hz may exist. This study replicates the feasibility of measuring high frequency (HF) correlations in six healthy controls and a patient with a brain tumor while analyzing non-physiological signal sources via simulation. Resting-state data were acquired using a high-speed multi-slab echo-volumar imaging pulse sequence with 136 ms temporal resolution. Bandpass frequency filtering in combination with sliding window seed-based connectivity analysis using running mean of the correlation maps was employed to map HF correlations up to 3.7 Hz. Computer simulations of Rician noise and the underlying point spread function were analyzed to estimate baseline spatial autocorrelation levels in four major networks (auditory, sensorimotor, visual, and default-mode). Using seed regions based on Brodmann areas, the auditory and default-mode networks were observed to have significant frequency band dependent HF correlations above baseline spatial autocorrelation levels. Correlations in the sensorimotor network were at trend level. The auditory network was still observed using a unilateral single voxel seed. In the patient, HF auditory correlations showed a spatial displacement near the tumor consistent with the displacement seen at low frequencies. In conclusion, our data suggest that HF connectivity in the human brain may be observable with high-speed fMRI, however, the detection sensitivity may depend on the network observed, data acquisition technique, and analysis method.