Visualising inter-subject variability in fMRI using threshold-weighted overlap maps

Seeing more than the Tip of the Iceberg: Approaches to Subthreshold Effects in Functional Magnetic Resonance Imaging of the Brain

Many functional magnetic resonance imaging (fMRI) studies and presurgical mapping applications rely on mass-univariate inference with subsequent multiple comparison correction. Statistical results are frequently visualized as thresholded statistical maps. This approach has inherent limitations including the risk of drawing overly-selective conclusions based only on selective results passing such thresholds. This article gives an overview of both established and newly emerging scientific approaches to supplement such conventional analyses by incorporating information about subthreshold effects with the aim to improve interpretation of findings or leverage a wider array of information. Topics covered include neuroimaging data visualization, p-value histogram analysis and the related Higher Criticism approach for detecting rare and weak effects. Further examples from multivariate analyses and dedicated Bayesian approaches are provided.

Electrophysiological representations of multivariate human emotion experience

ABSTRACTDespite the fact that human daily emotions are co-occurring by nature, most neuroscience studies have primarily adopted a univariate approach to identify the neural representation of emotion (emotion experience within a single emotion category) without adequate consideration of the co-occurrence of different emotions (emotion experience across different emotion categories simultaneously). To investigate the neural representations of multivariate emotion experience, this study employed the inter-situation representational similarity analysis (RSA) method. Researchers used an EEG dataset of 78 participants who watched 28 video clips and rated their experience on eight emotion categories. The EEG-based electrophysiological representation was extracted as the power spectral density (PSD) feature per channel in the five frequency bands. The inter-situation RSA method revealed significant correlations between the multivariate emotion experience ratings and PSD features in the Alpha and Beta bands, primarily over the frontal and parietal-occipital brain regions. The study found the identified EEG representations to be reliable with sufficient situations and participants. Moreover, through a series of ablation analyses, the inter-situation RSA further demonstrated the stability and specificity of the EEG representations for multivariate emotion experience. These findings highlight the importance of adopting a multivariate perspective for a comprehensive understanding of the neural representation of human emotion experience.

Unraveling individual differences in learning potential: A dynamic framework for the case of reading development

Children show an enormous capacity to learn during development, but with large individual differences in the time course and trajectory of learning and the achieved skill level. Recent progress in developmental sciences has shown the contribution of a multitude of factors including genetic variation, brain plasticity, socio-cultural context and learning experiences to individual development. These factors interact in a complex manner, producing children's idiosyncratic and heterogeneous learning paths. Despite an increasing recognition of these intricate dynamics, current research on the development of culturally acquired skills such as reading still has a typical focus on snapshots of children's performance at discrete points in time. Here we argue that this 'static' approach is often insufficient and limits advancements in the prediction and mechanistic understanding of individual differences in learning capacity. We present a dynamic framework which highlights the importance of capturing short-term trajectories during learning across multiple stages and processes as a proxy for long-term development on the example of reading. This framework will help explain relevant variability in children's learning paths and outcomes and fosters new perspectives and approaches to study how children develop and learn.

Regional, circuit and network heterogeneity of brain abnormalities in psychiatric disorders

The substantial individual heterogeneity that characterizes people with mental illness is often ignored by classical case-control research, which relies on group mean comparisons. Here we present a comprehensive, multiscale characterization of the heterogeneity of gray matter volume (GMV) differences in 1,294 cases diagnosed with one of six conditions (attention-deficit/hyperactivity disorder, autism spectrum disorder, bipolar disorder, depression, obsessive-compulsive disorder and schizophrenia) and 1,465 matched controls. Normative models indicated that person-specific deviations from population expectations for regional GMV were highly heterogeneous, affecting the same area in <7% of people with the same diagnosis. However, these deviations were embedded within common functional circuits and networks in up to 56% of cases. The salience-ventral attention system was implicated transdiagnostically, with other systems selectively involved in depression, bipolar disorder, schizophrenia and attention-deficit/hyperactivity disorder. Phenotypic differences between cases assigned the same diagnosis may thus arise from the heterogeneous localization of specific regional deviations, whereas phenotypic similarities may be attributable to the dysfunction of common functional circuits and networks.

Interindividual variability and consistency of language mapping paradigms for presurgical use

Most functional MRI studies of language processing have focussed on group-level inference, but for clinical use, the aim is to predict outcomes at an individual patient level. This requires being able to identify atypical activation and understand how differences relate to language outcomes. A language mapping paradigm that selectively activates left hemisphere language regions in healthy individuals allows atypical activation in a patient to be more easily identified. We investigated the interindividual variability and consistency of language activation in 12 healthy participants using three tasks-verb generation, responsive naming, and sentence comprehension-for future presurgical use. Responsive naming produced the most consistent left-lateralised activation across participants in frontal and temporal regions that postsurgical voxel-based lesion-symptom mapping studies suggest are most critical for language outcomes. Studies with a long-term clinical aim of predicting language outcomes in neurosurgical patients and stroke patients should first establish paradigm validity at an individual level in healthy participants.

Acquired reading impairment following brain injury

This large-scale patient study investigated the rate, unique signatures associated with acquired reading impairments, its neurocognitive correlates, and long-term outcome in 731 acute stroke patients using the sentence and non-word reading subtests of Birmingham Cognitive Screen (BCoS). The objectives for the study were to explore the (i) potentially different error patterns among adult patients, (ii) associative relationship between the different subclasses of reading impairment and performance in other cognitive domains, and (iii) recovery rates in patients nine months post-lesion compared with their initial performance. The study revealed distinctive reading impairment profiles in patients with left hemisphere (LH) and right hemisphere (RH) lesions. Some interesting associations between reading disorder and other cognitive functions were observed. Nine months post-lesion, both groups showed some recovery in reading performance compared with their baseline performance, but the rate of improvement was higher for the LH group. The study reveals unique reading profiles and impairment patterns among left and right hemisphere lesions. The findings of the study provide a deeper understanding of reading deficits that will inform clinical practice, planning of rehabilitative interventions of brain injured patients, and the scientific community.

Dynamic contrast enhanced MRI for characterization of blood-brain-barrier dysfunction after traumatic brain injury

BACKGROUND AND PURPOSE

Dysfunction of the blood-brain-barrier (BBB) is a recognized pathological consequence of traumatic brain injury (TBI) which may play an important role in chronic TBI pathophysiology. We hypothesized that BBB disruption can be detected with dynamic contrast-enhanced (DCE) MRI not only in association with focal traumatic lesions but also in normal-appearing brain tissue of TBI patients, reflecting microscopic microvascular injury. We further hypothesized that BBB integrity would improve but not completely normalize months after TBI.

MATERIALS AND METHODS

DCE MRI was performed in 40 adult patients a median of 23 days after hospitalized TBI and in 21 healthy controls. DCE data was analyzed using Patlak and linear models, and derived metrics of BBB leakage including the volume transfer constant (Ktrans) and the normalized permeability index (NPI) were compared between groups. BBB metrics were compared with focal lesion distribution as well as with contemporaneous measures of symptomatology and cognitive function in TBI patients. Finally, BBB metrics were examined longitudinally among 18 TBI patients who returned for a second MRI a median of 204 days postinjury.

RESULTS

TBI patients exhibited higher mean Ktrans (p = 0.0028) and proportion of suprathreshold NPI voxels (p = 0.001) relative to controls. Tissue-based analysis confirmed greatest TBI-related BBB disruption in association with focal lesions, however elevated Ktrans was also observed in perilesional (p = 0.011) and nonlesional (p = 0.044) regions. BBB disruption showed inverse correlation with quality of life (rho = -0.51, corrected p = 0.016). Among the subset of TBI patients who underwent a second MRI several months after the initial evaluation, metrics of BBB disruption did not differ significantly at the group level, though variable longitudinal changes were observed at the individual subject level.

CONCLUSIONS

This pilot investigation suggests that TBI-related BBB disruption is detectable in the early post-injury period in association with focal and diffuse brain injury.

The optimized combination of aCompCor and ICA-AROMA to reduce motion and physiologic noise in task fMRI data

One of the main challenges in fMRI processing is filtering the task BOLD signals from the noise. Independent component analysis with automatic removal of motion artifacts (ICA-AROMA) reduces motion artifacts by identifying ICA noise components based on their location at the brain edges and cerebrospinal fluid (CSF), high frequency content and correlation with motion regressors. In anatomical component correction (aCompCor), physiological noise regressors extracted from CSF were regressed out from the fMRI time series. In this study, we compared three methods to combine aCompCor and ICA-AROMA denoising in one denoising step. In the first analysis, we regressed the temporal signals of the ICA components identified as noise by ICA-AROMA together with the noise signals determined by aCompCor from the fMRI signals. For the second and third analyses, the correlation between the temporal signals of the ICA components and the aCompCor noise signals was used as an additional criterion to identify the noise components. In the second analysis, the temporal signals of the ICA components classified as noise were regressed from the fMRI signals. In the third analysis, the noise components were removed. To compare the denoising strategies, we examined the fractional amplitude of low-frequency fluctuations (fALFF) and the overlap between the contrast maps. Our results revealed that including the aCompCor noise signals as regressors in ICA-AROMA resulted in more correctly identified noise components, higher fALFF values, and larger activation maps. Moreover, combining the temporal signals of the noise components identified by ICA-AROMA with the aCompCor signals in a noise regression matrix resulted in deactivations. These results suggest that using the correlation between the ICA component temporal signals and the aCompCor signals as noise identification criteria in ICA-AROMA is the best approach for combining both denoising methods.

Mapping of Language-and-Memory Networks in Patients With Temporal Lobe Epilepsy by Using the GE2REC Protocol

Preoperative mapping of language and declarative memory functions in temporal lobe epilepsy (TLE) patients is essential since they frequently encounter deterioration of these functions and show variable degrees of cerebral reorganization. Due to growing evidence on language and declarative memory interdependence at a neural and neuropsychological level, we propose the GE2REC protocol for interactive language-and-memory network (LMN) mapping. GE2REC consists of three inter-related tasks, sentence generation with implicit encoding (GE) and two recollection (2REC) memory tasks: recognition and recall. This protocol has previously been validated in healthy participants, and in this study, we showed that it also maps the LMN in the left TLE (N = 18). Compared to healthy controls (N = 19), left TLE (LTLE) showed widespread inter- and intra-hemispheric reorganization of the LMN through reduced activity of regions engaged in the integration and the coordination of this meta-network. We also illustrated how this protocol could be implemented in clinical practice individually by presenting two case studies of LTLE patients who underwent efficient surgery and became seizure-free but showed different cognitive outcomes. This protocol can be advantageous for clinical practice because it (a) is short and easy to perform; (b) allows brain mapping of essential cognitive functions, even at an individual level; (c) engages language-and-memory interaction allowing to evaluate the integrative processes within the LMN; (d) provides a more comprehensive assessment by including both verbal and visual modalities, as well as various language and memory processes. Based on the available postsurgical data, we presented preliminary results obtained with this protocol in LTLE patients that could potentially inform the clinical practice. This implies the necessity to further validate the potential of GE2REC for neurosurgical planning, along with two directions, guiding resection and describing LMN neuroplasticity at an individual level.

A Data-Based Approach for Selecting Pre- and Intra-Operative Language Mapping Tasks

Background: Pre- and intra-operative language mapping in neurosurgery patients frequently involves an object naming task. The choice of the optimal object naming paradigm remains challenging due to lack of normative data and standardization in mapping practices. The aim of this study was to identify object naming paradigms that robustly and consistently activate classical language regions and could therefore be used to improve the sensitivity of language mapping in brain tumor and epilepsy patients. Methods: Functional magnetic resonance imaging (fMRI) data from two independent groups of healthy controls (total = 79) were used to generate threshold-weighted voxel-based consistency maps. This novel approach allowed us to compare inter-subject consistency of activation for naming single objects in the visual and auditory modality and naming two objects in a phrase or a sentence. Results: We found that the consistency of activation in language regions was greater for naming two objects per picture than one object per picture, even when controlling for the number of names produced in 5 s. Conclusion: More consistent activation in language areas for naming two objects compared to one object suggests that two-object naming tasks may be more suitable for delimiting language eloquent regions with pre- and intra-operative language testing. More broadly, we propose that the functional specificity of brain mapping paradigms for a whole range of different linguistic and non-linguistic functions could be enhanced by referring to databased models of inter-subject consistency and variability in typical and atypical brain responses.

How to capture developmental brain dynamics: gaps and solutions

Capturing developmental and learning-induced brain dynamics is extremely challenging as changes occur interactively across multiple levels and emerging functions. Different levels include the (social) environment, cognitive and behavioral levels, structural and functional brain changes, and genetics, while functions include domains such as math, reading, and executive function. Here, we report the insights that emerged from the workshop “Capturing Developmental Brain Dynamics”, organized to bring together multidisciplinary approaches to integrate data on development and learning across different levels, functions, and time points. During the workshop, current main gaps in our knowledge and tools were identified including the need for: (1) common frameworks, (2) longitudinal, large-scale, multisite studies using representative participant samples, (3) understanding interindividual variability, (4) explicit distinction of understanding versus predicting, and (5) reproducible research. After illustrating interactions across levels and functions during development, we discuss the identified gaps and provide solutions to advance the capturing of developmental brain dynamics.

Individual Differences in Brain Responses: New Opportunities for Tailoring Health Communication Campaigns

Prevention neuroscience investigates the brain basis of attitude and behavior change. Over the years, an increasingly structurally and functionally resolved "persuasion network" has emerged. However, current studies have only identified a small handful of neural structures that are commonly recruited during persuasive message processing, and the extent to which these (and other) structures are sensitive to numerous individual difference factors remains largely unknown. In this project we apply a multi-dimensional similarity-based individual differences analysis to explore which individual factors-including characteristics of messages and target audiences-drive patterns of brain activity to be more or less similar across individuals encountering the same anti-drug public service announcements (PSAs). We demonstrate that several ensembles of brain regions show response patterns that are driven by a variety of unique factors. These results are discussed in terms of their implications for neural models of persuasion, prevention neuroscience and message tailoring, and methodological implications for future research.

Patterns and predictors of language representation and the influence of epilepsy surgery on language reorganization in children and young adults with focal lesional epilepsy

Mapping brain functions is crucial for neurosurgical planning in patients with drug-resistant seizures. However, presurgical language mapping using either functional or structural networks can be challenging, especially in children. In fact, most of the evidence on this topic derives from cross-sectional or retrospective studies in adults submitted to anterior temporal lobectomy. In this prospective study, we used fMRI and DTI to explore patterns of language representation, their predictors and impact on cognitive performances in 29 children and young adults (mean age at surgery: 14.6 ± 4.5 years) with focal lesional epilepsy. In 20 of them, we also assessed the influence of epilepsy surgery on language lateralization. All patients were consecutively enrolled at a single epilepsy surgery center between 2009 and 2015 and assessed with preoperative structural and functional 3T brain MRI during three language tasks: Word Generation (WG), Rhyme Generation (RG) and a comprehension task. We also acquired DTI data on arcuate fasciculus in 24 patients. We first assessed patterns of language representation (relationship of activations with the epileptogenic lesion and Laterality Index (LI)) and then hypothesized a causal model to test whether selected clinical variables would influence the patterns of language representation and the ensuing impact of the latter on cognitive performances. Twenty out of 29 patients also underwent postoperative language fMRI. We analyzed possible changes of fMRI and DTI LIs and their clinical predictors. Preoperatively, we found atypical language lateralization in four patients during WG task, in one patient during RG task and in seven patients during the comprehension task. Diffuse interictal EEG abnormalities predicted a more atypical language representation on fMRI (p = 0.012), which in turn correlated with lower attention (p = 0.036) and IQ/GDQ scores (p = 0.014). Postoperative language reorganization implied shifting towards atypical language representation. Abnormal postoperative EEG (p = 0.003) and surgical failures (p = 0.015) were associated with more atypical language lateralization, in turn correlating with worsened fluency. Neither preoperative asymmetry nor postoperative DTI LI changes in the arcuate fasciculus were observed. Focal lesional epilepsy associated with diffuse EEG abnormalities may favor atypical language lateralization and worse cognitive performances, which are potentially reversible after successful surgery.

Brainglance: Visualizing Group Level MRI Data at One Glance

The vast majority of studies using functional magnetic resonance imaging (fMRI) are analyzed on the group level. Standard group-level analyses, however, come with severe drawbacks: First, they assume functional homogeneity within the group, building on the idea that we use our brains in similar ways. Second, group-level analyses require spatial warping and substantial smoothing to accommodate for anatomical variability across subjects. Such procedures massively distort the underlying fMRI data, which hampers the spatial specificity. Taken together, group statistics capture the effective overlap, rendering the modeling of individual deviations impossible – a major source of false positivity and negativity. The alternative analysis approach is to leave the data in the native subject space, but this makes comparison across individuals difficult. Here, we propose a new framework for visualizing group-level information, better preserving the information of individual subjects. Our proposal is to limit the use of invasive data procedures such as spatial smoothing and warping and rather extract regional information from the individuals. This information is then visualized for all subjects and brain areas at one glance – hence we term the method brainglance. Additionally, our method incorporates a means for clustering individuals to further identify common traits. We showcase our method on two publicly available data sets and discuss our findings.

Inflexibility in Autism Spectrum Disorder: Need for certainty and atypical emotion processing share the blame

Although ASD (Autism Spectrum Disorder) diagnosis requires the co-occurrence of socio-emotional deficits and inflexible behaviors, the interaction between these two domains remains unexplored. We used an emotional Wisconsin Card Sorting Test adapted to fMRI to explore this question. ASD and control participants matched a central card (a face) with one of four surrounding cards according to one of three rules: frame color, facial identity or expression. Feedback informed participants on whether to change or maintain the current sorting rule. For each rule, we modeled feedback onsets to change, switch (confirming the newly found rule) and maintenance events. "Bias error", which measures participants' willingness to switch, was larger in ASD participants for the emotional sorting rule. Brain activity to change events showed no group differences. In response to switch events significantly larger activity was observed for ASD participants in bilateral Inferior Parietal Sulci. Inflexibility in ASD appears characterized by the unwillingness to switch toward processing socio-emotional information, rather than a major disruption in cognitive flexibility. However, a larger activity to switch events in ASD highlights the need for a higher level of certainty before setting into a stable processing stage, which may be particularly detrimental in the highly changeable socio-emotional environment.

Sensorimotor, language, and working memory representation within the human cerebellum

The cerebellum is involved in a wide range of behaviours. A key organisational principle from animal studies is that somatotopically corresponding sensory input and motor output reside in the same cerebellar cortical areas. However, compelling evidence for a similar arrangement in humans and whether it extends to cognitive functions is lacking. To address this, we applied cerebellar optimised whole‐brain functional MRI in 20 healthy subjects. To assess spatial overlap within the sensorimotor and cognitive domains, we recorded activity to a sensory stimulus (vibrotactile) and a motor task; the Sternberg verbal working memory (VWM) task; and a verb generation paradigm. Consistent with animal data, sensory and motor activity overlapped with a somatotopic arrangement in ipsilateral areas of the anterior and posterior cerebellum. During the maintenance phase of the Sternberg task, a positive linear relationship between VWM load and activity was observed in right Lobule VI, extending into Crus I bilaterally. Articulatory movement gave rise to bilateral activity in medial Lobule VI. A conjunction of two independent language tasks localised activity during verb generation in right Lobule VI‐Crus I, which overlapped with activity during VWM. These results demonstrate spatial compartmentalisation of sensorimotor and cognitive function in the human cerebellum, with each area involved in more than one aspect of a given behaviour, consistent with an integrative function. Sensorimotor localisation was uniform across individuals, but the representation of cognitive tasks was more variable, highlighting the importance of individual scans for mapping higher order functions within the cerebellum.

Reconstruction of behavior-relevant individual brain activity: an individualized fMRI study

Different patterns of brain activity are observed in various subjects across a wide functional domain. However, these individual differences, which are often neglected through the group average, are not yet completely understood. Based on the fundamental assumption that human behavior is rooted in the underlying brain function, we speculated that the individual differences in brain activity are reflected in the individual differences in behavior. Adopting 98 behavioral measures and assessing the brain activity induced at seven task functional magnetic resonance imaging states, we demonstrated that the individual differences in brain activity can be used to predict behavioral measures of individual subjects with high accuracy using the partial least square regression model. In addition, we revealed that behavior-relevant individual differences in brain activity transferred between different task states and can be used to reconstruct individual brain activity. Reconstructed individual brain activity retained certain individual differences which were lost in the group average and could serve as an individual functional localizer. Therefore, our results suggest that the individual differences in brain activity contain behavior-relevant information and should be included in group averaging. Moreover, reconstructed individual brain activity shows a potential use in precise and personalized medicine.

Mirror Neuron System Activation in Children With Unilateral Cerebral Palsy During Observation of Actions Performed by a Pathological Model

Background. Recent evidence suggested that Action Observation Therapy (AOT), based on observation of actions followed by immediate reproduction, could be a useful rehabilitative strategy for promoting functional recovery of children affected by unilateral cerebral palsy (UCP). AOT most likely exploits properties of the parieto-premotor mirror neuron system (MNS). This is more intensely activated when participants observe actions belonging to their own motor repertoire. Objective. The aim of the present study was to investigate the issue of whether MNS of UCP children is better activated by actions performed by a paretic hand rather than a healthy one. Methods. Using functional magnetic resonance imaging, we assessed brain activation in a homogeneous group of 10 right UCP children compared with that of 10 right-handed typically developing (TD) children, during observation of grasping actions performed by a healthy or a paretic hand. Results. The results revealed a significant activation within the MNS in both UCP and TD children, more lateralized to the left hemisphere in the TD group. Most important, region of interest (ROI) analysis on parietal and premotor regions showed that, in UCP, the MNS was more strongly activated by observation of actions performed by the paretic hand, a motor model more similar to the observer’s motor repertoire. Conclusions. This study shows that children affected by spastic UCP exhibit enhanced activation of the MNS during observation of goal-directed actions performed by a pathological model with respect to a healthy one.

Spatial distribution bias in subject-specific abnormalities analyses

The neuroimaging community has seen a renewed interest in algorithms that provide a location-independent summary of subject-specific abnormalities (SSA) to assess individual lesion load. More recently, these methods have been extended to assess whether multiple individuals within the same cohort exhibit extrema in the same spatial location (e.g., voxel or region of interest). However, the statistical validity of this approach has not been rigorously established. The current study evaluated the potential for a spatial bias in the distribution of SSA using several common z-transformation algorithms (leave-one-out [LOO]; independent sample [IDS]; Enhanced Z-Score Microstructural Assessment of Pathology [EZ-MAP]; distribution-corrected z-scores [DisCo-Z]) using both simulated data and DTI data from 50 healthy controls. Results indicated that methods which z-transformed data based on statistical moments from a reference group (LOO, DisCo-Z) led to bias in the spatial location of extrema for the comparison group. In contrast, methods that z-transformed data using an independent third group (EZ-MAP, IDS) resulted in no spatial bias. Importantly, none of the methods exhibited bias when results were summed across all individual elements. The spatial bias is primarily driven by sampling error, in which differences in the mean and standard deviation of the untransformed data have a higher probability of producing extrema in the same spatial location for the comparison but not reference group. In conclusion, evaluating SSA overlap within cohorts should be either be avoided in deference to established group-wise comparisons or performed only when data is available from an independent third group.

fMRI prediction of naming change after adult temporal lobe epilepsy surgery: Activation matters

OBJECTIVE

We aimed to predict language deficits after epilepsy surgery. In addition to evaluating surgical factors examined previously, we determined the impact of the extent of functional magnetic resonance imaging (fMRI) activation that was resected on naming ability.

METHOD

Thirty-five adults (mean age 37.5 ± 10.9 years, 13 male) with temporal lobe epilepsy completed a preoperative fMRI auditory description decision task, which reliably activates frontal and temporal language networks. Patients underwent temporal lobe resections (20 left resection). The Boston Naming Test (BNT) was used to determine language functioning before and after surgery. Language dominance was determined for Broca and Wernicke area (WA) by calculating a laterality index following statistical parametric mapping processing. We used an innovative method to generate anatomic resection masks automatically from pre- and postoperative MRI tissue map comparison. This mask provided the following: (a) resection volume; (b) overlap between resection and preoperative activation; and (c) overlap between resection and WA. We examined postoperative language change predictors using stepwise linear regression. Predictors included parameters described above as well as age at seizure onset (ASO), preoperative BNT score, and resection side and its relationship to language dominance.

RESULTS

Seven of 35 adults had significant naming decline (6 dominant-side resections). The final regression model predicted 38% of the naming score change variance (adjusted r2  = 0.28, P = 0.012). The percentage of top 10% fMRI activation resected (P = 0.017) was the most significant contributor. Other factors in the model included WA LI, ASO, volume of WA resected, and WA LI absolute value (extent of laterality).

SIGNIFICANCE

Resection of fMRI activation during a word-definition decision task is an important factor for postoperative change in naming ability, along with other previously reported predictors. Currently, many centers establish language dominance using fMRI. Our results suggest that the amount of the top 10% of language fMRI activation in the intended resection area provides additional predictive power and should be considered when planning surgical resection.

Cognitive neuroscience using wearable magnetometer arrays: Non-invasive assessment of language function

Recent work has demonstrated that Optically Pumped Magnetometers (OPMs) can be utilised to create a wearable Magnetoencephalography (MEG) system that is motion robust. In this study, we use this system to map eloquent cortex using a clinically validated language lateralisation paradigm (covert verb generation: 120 trials, ∼10 min total duration) in healthy adults (n = 3). We show that it is possible to lateralise and localise language function on a case by case basis using this system. Specifically, we show that at a sensor and source level we can reliably detect a lateralising beta band (15-30 Hz) desynchronization in all subjects. This is the first study of human cognition using OPMs and not only highlights this technology's utility as tool for (developmental) cognitive neuroscience but also its potential to contribute to surgical planning via mapping of eloquent cortex, especially in young children.

Whole-Brain Network Connectivity Underlying the Human Speech Articulation as Emerged Integrating Direct Electric Stimulation, Resting State fMRI and Tractography

Production of fluent speech in humans is based on a precise and coordinated articulation of sounds. A speech articulation network (SAN) has been observed in multiple brain studies typically using either neuroimaging or direct electrical stimulation (DES), thus giving limited knowledge about the whole brain structural and functional organization of this network. In this study, seven right-handed patients underwent awake surgery resection of low-grade gliomas (4) and cavernous angiomas. We combined pre-surgical resting state fMRI (rs-fMRI) and diffusion MRI together with speech arrest sites obtained intra-operatively with DES to address the following goals: (i) determine the cortical areas contributing to the intrinsic functional SAN using the speech arrest sites as functional seeds for rs-fMRI; (ii) evaluate the relative contribution of gray matter terminations from the two major language dorsal stream bundles, the superior longitudinal fasciculus (SLF III) and the arcuate fasciculus (AF); and (iii) evaluate the possible pre-surgical prediction of SAN with rs-fMRI. In all these right-handed patients the intrinsic functional SAN included frontal, inferior parietal, temporal, and insular regions symmetrically and bilaterally distributed across the two hemispheres regardless of the side (four right) of speech arrest evocation. The SLF III provided a much higher density of terminations in the cortical regions of SAN in respect to AF. Pre-surgical rs-fMRI data demonstrated moderate ability to predict the SAN. The set of functional and structural data provided in this multimodal study characterized, at a whole-brain level, a distributed and bi-hemispherical network subserving speech articulation.

Small sample sizes reduce the replicability of task-based fMRI studies

Despite a growing body of research suggesting that task-based functional magnetic resonance imaging (fMRI) studies often suffer from a lack of statistical power due to too-small samples, the proliferation of such underpowered studies continues unabated. Using large independent samples across eleven tasks, we demonstrate the impact of sample size on replicability, assessed at different levels of analysis relevant to fMRI researchers. We find that the degree of replicability for typical sample sizes is modest and that sample sizes much larger than typical (e.g., N = 100) produce results that fall well short of perfectly replicable. Thus, our results join the existing line of work advocating for larger sample sizes. Moreover, because we test sample sizes over a fairly large range and use intuitive metrics of replicability, our hope is that our results are more understandable and convincing to researchers who may have found previous results advocating for larger samples inaccessible.

Interpreting and Utilising Intersubject Variability in Brain Function

We consider between-subject variance in brain function as data rather than noise. We describe variability as a natural output of a noisy plastic system (the brain) where each subject embodies a particular parameterisation of that system. In this context, variability becomes an opportunity to: (i) better characterise typical versus atypical brain functions; (ii) reveal the different cognitive strategies and processing networks that can sustain similar tasks; and (iii) predict recovery capacity after brain damage by taking into account both damaged and spared processing pathways. This has many ramifications for understanding individual learning preferences and explaining the wide differences in human abilities and disabilities. Understanding variability boosts the translational potential of neuroimaging findings, in particular in clinical and educational neuroscience.

The increase in medial prefrontal glutamate/glutamine concentration during memory encoding is associated with better memory performance and stronger functional connectivity in the human medial prefrontal-thalamus-hippocampus network

The classical model of the declarative memory system describes the hippocampus and its interactions with representational brain areas in posterior neocortex as being essential for the formation of long‐term episodic memories. However, new evidence suggests an extension of this classical model by assigning the medial prefrontal cortex (mPFC) a specific, yet not fully defined role in episodic memory. In this study, we utilized 1H magnetic resonance spectroscopy (MRS) and psychophysiological interaction (PPI) analysis to lend further support for the idea of a mnemonic role of the mPFC in humans. By using MRS, we measured mPFC γ‐aminobutyric acid (GABA) and glutamate/glutamine (GLx) concentrations before and after volunteers memorized face–name association. We demonstrate that mPFC GLx but not GABA levels increased during the memory task, which appeared to be related to memory performance. Regarding functional connectivity, we used the subsequent memory paradigm and found that the GLx increase was associated with stronger mPFC connectivity to thalamus and hippocampus for associations subsequently recognized with high confidence as opposed to subsequently recognized with low confidence/forgotten. Taken together, we provide new evidence for an mPFC involvement in episodic memory by showing a memory‐related increase in mPFC excitatory neurotransmitter levels that was associated with better memory and stronger memory‐related functional connectivity in a medial prefrontal–thalamus–hippocampus network.

Using transcranial magnetic stimulation of the undamaged brain to identify lesion sites that predict language outcome after stroke

Transcranial magnetic stimulation focused on either the left anterior supramarginal gyrus or opercular part of the left inferior frontal gyrus has been reported to transiently impair the ability to perform phonological more than semantic tasks. Here we tested whether phonological processing abilities were also impaired following lesions to these regions in right-handed, English speaking adults, who were investigated at least 1 year after a left-hemisphere stroke. When our regions of interest were limited to 0.5 cm³ of grey matter centred around sites that had been identified with transcranial magnetic stimulation-based functional localization, phonological impairments were observed in 74% (40/54) of patients with damage to the regions and 21% (21/100) of patients sparing these regions. This classification accuracy was better than that observed when using regions of interest centred on activation sites in previous functional magnetic resonance imaging studies of phonological processing, or transcranial magnetic stimulation sites that did not use functional localization. New regions of interest were generated by redefining the borders of each of the transcranial magnetic stimulation sites to include areas that were consistently damaged in the patients with phonological impairments. This increased the incidence of phonological impairments in the presence of damage to 85% (46/54) and also reduced the incidence of phonological impairments in the absence of damage to 15% (15/100). The difference in phonological processing abilities between those with and without damage to these ‘transcranial magnetic stimulation-guided’ regions remained highly significant even after controlling for the effect of lesion size. The classification accuracy of the transcranial magnetic stimulation-guided regions was validated in a second sample of 108 patients and found to be better than that for (i) functional magnetic resonance imaging-guided regions; (ii) a region identified from an unguided lesion overlap map; and (iii) a region identified from voxel-based lesion-symptom mapping. Finally, consistent with prior findings from functional imaging and transcranial magnetic stimulation in healthy participants, we show how damage to our transcranial magnetic stimulation-guided regions affected performance on phonologically more than semantically demanding tasks. The observation that phonological processing abilities were impaired years after the stroke, suggests that other brain regions were not able to fully compensate for the contribution that the transcranial magnetic stimulation-guided regions make to language tasks. More generally, our novel transcranial magnetic stimulation-guided lesion-deficit mapping approach shows how non-invasive stimulation of the healthy brain can be used to guide the identification of regions where brain damage is likely to cause persistent behavioural effects.

Inter-Subject Variability of Axonal Injury in Diffuse Traumatic Brain Injury

Traumatic brain injury (TBI) is a leading cause of cognitive morbidity worldwide for which reliable biomarkers are needed. Diffusion tensor imaging (DTI) is a promising biomarker of traumatic axonal injury (TAI); however, existing studies have been limited by a primary reliance on group-level analytic methods not well suited to account for inter-subject variability. In this study, 42 adults with TBI of at least moderate severity were examined 3 months following injury and compared with 35 healthy controls. DTI data were used for both traditional group-level comparison and subject-specific analysis using the distribution-corrected Z-score (DisCo-Z) approach. Inter-subject variation in TAI was assessed in a threshold-invariant manner using a threshold-weighted overlap map derived from subject-specific analysis. Receiver operator curve analysis was used to examine the ability of subject-specific DTI analysis to identify TBI subjects with significantly impaired processing speed in comparison with region of interest-based fractional anisotropy (FA) measurements and clinical characteristics. Traditional group-wise analysis demonstrated widespread reductions of white matter FA within the TBI group (voxel-wise p < 0.05, corrected), despite relatively low consistency of subject-level effects secondary to widespread variation in the spatial distribution of TAI. Subject-specific mapping of TAI with the DisCo-Z approach was the best predictor of impaired processing speed, achieving high classification accuracy (area under the curve [AUC] = 0.94). In moderate-to-severe TBI, there is substantial inter-subject variation in TAI, with extent strongly correlated to post-traumatic deficits in processing speed. Significant group-level effects do not necessarily represent consistent effects at the individual level. Better accounting for inter-subject variability in neurobiological manifestations of TBI may substantially improve the ability to detect and classify patterns of injury.

Unsupervised Classification During Time-Series Model Building

Researchers who collect multivariate time-series data across individuals must decide whether to model the dynamic processes at the individual level or at the group level. A recent innovation, group iterative multiple model estimation (GIMME), offers one solution to this dichotomy by identifying group-level time-series models in a data-driven manner while also reliably recovering individual-level patterns of dynamic effects. GIMME is unique in that it does not assume homogeneity in processes across individuals in terms of the patterns or weights of temporal effects. However, it can be difficult to make inferences from the nuances in varied individual-level patterns. The present article introduces an algorithm that arrives at subgroups of individuals that have similar dynamic models. Importantly, the researcher does not need to decide the number of subgroups. The final models contain reliable group-, subgroup-, and individual-level patterns that enable generalizable inferences, subgroups of individuals with shared model features, and individual-level patterns and estimates. We show that integrating community detection into the GIMME algorithm improves upon current standards in two important ways: (1) providing reliable classification and (2) increasing the reliability in the recovery of individual-level effects. We demonstrate this method on functional MRI from a sample of former American football players.

Ten problems and solutions when predicting individual outcome from lesion site after stroke

In this paper, we consider solutions to ten of the challenges faced when trying to predict an individual's functional outcome after stroke on the basis of lesion site. A primary goal is to find lesion-outcome associations that are consistently observed in large populations of stroke patients because consistent associations maximise confidence in future individualised predictions. To understand and control multiple sources of inter-patient variability, we need to systematically investigate each contributing factor and how each factor depends on other factors. This requires very large cohorts of patients, who differ from one another in typical and measurable ways, including lesion site, lesion size, functional outcome and time post stroke (weeks to decades). These multivariate investigations are complex, particularly when the contributions of different variables interact with one another. Machine learning algorithms can help to identify the most influential variables and indicate dependencies between different factors. Multivariate lesion analyses are needed to understand how the effect of damage to one brain region depends on damage or preservation in other brain regions. Such data-led investigations can reveal predictive relationships between lesion site and outcome. However, to understand and improve the predictions we need explanatory models of the neural networks and degenerate pathways that support functions of interest. This will entail integrating the results of lesion analyses with those from functional imaging (fMRI, MEG), transcranial magnetic stimulation (TMS) and diffusor tensor imaging (DTI) studies of healthy participants and patients.