Sex classification from functional brain connectivity: Generalization to multiple datasets.

Machine learning (ML) approaches are increasingly being applied to neuroimaging data. Studies in neuroscience typically have to rely on a limited set of training data which may impair the generalizability of ML models. However, it is still unclear which kind of training sample is best suited to optimize generalization performance. In the present study, we systematically investigated the generalization performance of sex classification models trained on the parcelwise connectivity profile of either single samples or compound samples of two different sizes. Generalization performance was quantified in terms of mean across-sample classification accuracy and spatial consistency of accurately classifying parcels. Our results indicate that the generalization performance of parcelwise classifiers (pwCs) trained on single dataset samples is dependent on the specific test samples. Certain datasets seem to "match" in the sense that classifiers trained on a sample from one dataset achieved a high accuracy when tested on the respected other one and vice versa. The pwCs trained on the compound samples demonstrated overall highest generalization performance for all test samples, including one derived from a dataset not included in building the training samples. Thus, our results indicate that both a large sample size and a heterogeneous data composition of a training sample have a central role in achieving generalizable results.

A topography-based predictive framework for naturalistic viewing fMRI.

Functional magnetic resonance imaging (fMRI) during naturalistic viewing (NV) provides exciting opportunities for studying brain functions in more ecologically valid settings. Understanding individual differences in brain functions during NV and their behavioural relevance has recently become an important goal. However, methods specifically designed for this purpose remain limited. Here, we propose a topography-based predictive framework (TOPF) to fill this methodological gap. TOPF identifies individual-specific evoked activity topographies in a data-driven manner and examines their behavioural relevance using a machine learning-based predictive framework. We validate TOPF on both NV and task-based fMRI data from multiple conditions. Our results show that TOPF effectively and stably captures individual differences in evoked brain activity and successfully predicts phenotypes across cognition, emotion and personality on unseen subjects from their activity topographies. Moreover, TOPF compares favourably with functional connectivity-based approaches in prediction performance, with the identified predictive brain regions being neurobiologically interpretable. Crucially, we highlight the importance of examining individual evoked brain activity topographies in advancing our understanding of the brain-behaviour relationship. We believe that the TOPF approach provides a simple but powerful tool for understanding brain-behaviour relationships on an individual level with a strong potential for clinical applications.

Naturalistic viewing increases individual identifiability based on connectivity within functional brain networks.

Naturalistic viewing (NV) is currently considered a promising paradigm for studying individual differences in functional brain organization. While whole brain functional connectivity (FC) under NV has been relatively well characterized, so far little work has been done on a network level. Here, we extend current knowledge by characterizing the influence of NV on FC in fourteen meta-analytically derived brain networks considering three different movie stimuli in comparison to resting-state (RS). We show that NV increases identifiability of individuals over RS based on functional connectivity in certain, but not all networks. Furthermore, movie stimuli including a narrative appear more distinct from RS. In addition, we assess individual variability in network FC by comparing within- and between-subject similarity during NV and RS. We show that NV can evoke individually distinct NFC patterns by increasing inter-subject variability while retaining within-subject similarity. Crucially, our results highlight that this effect is not observable across all networks, but rather dependent on the network-stimulus combination. Our results confirm that NV can improve the detection of individual differences over RS and underline the importance of selecting the appropriate combination of movie and cognitive network for the research question at hand.

Imaging evolution of the primate brain: the next frontier?

Evolution, as we currently understand it, strikes a delicate balance between animals' ancestral history and adaptations to their current niche. Similarities between species are generally considered inherited from a common ancestor whereas observed differences are considered as more recent evolution. Hence comparing species can provide insights into the evolutionary history. Comparative neuroimaging has recently emerged as a novel subdiscipline, which uses magnetic resonance imaging (MRI) to identify similarities and differences in brain structure and function across species. Whereas invasive histological and molecular techniques are superior in spatial resolution, they are laborious, post-mortem, and oftentimes limited to specific species. Neuroimaging, by comparison, has the advantages of being applicable across species and allows for fast, whole-brain, repeatable, and multi-modal measurements of the structure and function in living brains and post-mortem tissue. In this review, we summarise the current state of the art in comparative anatomy and function of the brain and gather together the main scientific questions to be explored in the future of the fascinating new field of brain evolution derived from comparative neuroimaging.

Sex Classification by Resting State Brain Connectivity.

A large amount of brain imaging research has focused on group studies delineating differences between males and females with respect to both cognitive performance as well as structural and functional brain organization. To supplement existing findings, the present study employed a machine learning approach to assess how accurately participants' sex can be classified based on spatially specific resting state (RS) brain connectivity, using 2 samples from the Human Connectome Project (n1 = 434, n2 = 310) and 1 fully independent sample from the 1000BRAINS study (n = 941). The classifier, which was trained on 1 sample and tested on the other 2, was able to reliably classify sex, both within sample and across independent samples, differing both with respect to imaging parameters and sample characteristics. Brain regions displaying highest sex classification accuracies were mainly located along the cingulate cortex, medial and lateral frontal cortex, temporoparietal regions, insula, and precuneus. These areas were stable across samples and match well with previously described sex differences in functional brain organization. While our data show a clear link between sex and regionally specific brain connectivity, they do not support a clear-cut dimorphism in functional brain organization that is driven by sex alone.

Semantic and attentional networks in bilingual processing: fMRI connectivity signatures of translation directionality.

Comparisons between backward and forward translation (BT, FT) have long illuminated the organization of bilingual memory, with neuroscientific evidence indicating that FT would involve greater linguistic and attentional demands. However, no study has directly assessed the functional interaction between relevant mechanisms. Against this background, we conducted the first fMRI investigation of functional connectivity (FC) differences between BT and FT. In addition to yielding lower behavioral outcomes, FT was characterized by increased FC between a core semantic hub (the left anterior temporal lobe, ATL) and key nodes of attentional and vigilance networks (left inferior frontal, left orbitofrontal, and bilateral parietal clusters). Instead, distinct FC patterns for BT emerged only between the left ATL and the right thalamus, a region implicated in automatic relaying of sensory information to cortical regions. Therefore, FT seems to involve enhanced coupling between semantic and attentional mechanisms, suggesting that asymmetries in cross-language processing reflect dynamic interactions between linguistic and domain-general systems.

Sex differences and menstrual cycle effects in cognitive and sensory resting state networks.

It has not yet been established if resting state (RS) connectivity reflects stable characteristics of the brain, or if it is modulated by the psychological and/or physiological state of the participant. Based on research demonstrating sex hormonal effects in task-related brain activity, the present study aimed to investigate corresponding differences in RS networks. RS functional Magnetic Resonance Imaging (RS fMRI) was conducted in women during three different menstrual cycle phases, while men underwent three repeated RS fMRI testing sessions. Independent component analysis was used to identify the default mode network (DMN) and an auditory RS network. For the DMN, RS connectivity was stable across testing sessions in men, but varied across the menstrual cycle in women. For the auditory network (AN), retest reliable sex difference was found. Although RS activity in the DMN has been interpreted as trait characteristic of functional brain organization, these findings suggest that RS activity in networks involving frontal areas might be less stable than in sensory-based networks and can dynamically fluctuate. This also implies that some of the previously reported effects of sex hormones on task-related activity might to some extent be mediated by cycle-related fluctuations in RS activity, especially when frontal areas are involved.

Sex hormones affect language lateralisation but not cognitive control in normally cycling women.

This article is part of a Special Issue "Estradiol and Cognition". Natural fluctuations of sex hormones during the menstrual cycle have been shown to modulate language lateralisation. Using the dichotic listening (DL) paradigm, a well-established measurement of language lateralisation, several studies revealed that the left hemispheric language dominance was stronger when levels of estradiol were high. A recent study (Hjelmervik et al., 2012) showed, however, that high levels of follicular estradiol increased lateralisation only in a condition that required participants to cognitively control (top-down) the stimulus-driven (bottom-up) response. This finding suggested that sex hormones modulate lateralisation only if cognitive control demands are high. The present study investigated language lateralisation in 73 normally cycling women under three attention conditions that differed in cognitive control demands. Saliva estradiol and progesterone levels were determined by luminescence immunoassays. Women were allocated to a high or low estradiol group. The results showed a reduced language lateralisation when estradiol and progesterone levels were high. The effect was independent of the attention condition indicating that estradiol marginally affected cognitive control. The findings might suggest that high levels of estradiol especially reduce the stimulus-driven (bottom-up) aspect of lateralisation rather than top-down cognitive control.

Inter- and intra-subject similarity in network functional connectivity across a full narrative movie.

Naturalistic paradigms, such as watching movies during functional magnetic resonance imaging (fMRI), are thought to prompt the emotional and cognitive processes typically elicited in real life situations. Therefore, naturalistic viewing (NV) holds great potential for studying individual differences. However, in how far NV elicits similarity within and between subjects on a network level, particularly depending on emotions portrayed in movies, is currently unknown. We used the studyforrest dataset to investigate the inter- and intra-subject similarity in network functional connectivity (NFC) of 14 meta-analytically defined networks across a full narrative, audio-visual movie split into 8 consecutive movie segments. We characterized the movie segments by valence and arousal portrayed within the sequences, before utilizing a linear mixed model to analyze which factors explain inter- and intra-subject similarity. Our results showed that the model best explaining inter-subject similarity comprised network, movie segment, valence and a movie segment by valence interaction. Intra-subject similarity was influenced significantly by the same factors and an additional three-way interaction between movie segment, valence and arousal. Overall, inter- and intra-subject similarity in NFC were sensitive to the ongoing narrative and emotions in the movie. Lowest similarity both within and between subjects was seen in the emotional regulation network and networks associated with long-term memory processing, which might be explained by specific features and content of the movie. We conclude that detailed characterization of movie features is crucial for NV research.

Sex classification from functional brain connectivity: Generalization to multiple datasets Generalizability of sex classifiers.

Machine learning (ML) approaches are increasingly being applied to neuroimaging data. Studies in neuroscience typically have to rely on a limited set of training data which may impair the generalizability of ML models. However, it is still unclear which kind of training sample is best suited to optimize generalization performance. In the present study, we systematically investigated the generalization performance of sex classification models trained on the parcelwise connectivity profile of either single samples or a compound sample containing data from four different datasets. Generalization performance was quantified in terms of mean across-sample classification accuracy and spatial consistency of accurately classifying parcels. Our results indicate that generalization performance of pwCs trained on single dataset samples is dependent on the specific test samples. Certain datasets seem to "match" in the sense that classifiers trained on a sample from one dataset achieved a high accuracy when tested on the respected other one and vice versa. The pwC trained on the compound sample demonstrated overall highest generalization performance for all test samples, including one derived from a dataset not included in building the training samples. Thus, our results indicate that a big and heterogenous training sample comprising data of multiple datasets is best suited to achieve generalizable results.

SpEx: a German-language dataset of speech and executive function performance.

This work presents data from 148 German native speakers (20-55 years of age), who completed several speaking tasks, ranging from formal tests such as word production tests to more ecologically valid spontaneous tasks that were designed to mimic natural speech. This speech data is supplemented by performance measures on several standardised, computer-based executive functioning (EF) tests covering domains of working-memory, cognitive flexibility, inhibition, and attention. The speech and EF data are further complemented by a rich collection of demographic data that documents education level, family status, and physical and psychological well-being. Additionally, the dataset includes information of the participants' hormone levels (cortisol, progesterone, oestradiol, and testosterone) at the time of testing. This dataset is thus a carefully curated, expansive collection of data that spans over different EF domains and includes both formal speaking tests as well as spontaneous speaking tasks, supplemented by valuable phenotypical information. This will thus provide the unique opportunity to perform a variety of analyses in the context of speech, EF, and inter-individual differences, and to our knowledge is the first of its kind in the German language. We refer to this dataset as SpEx since it combines speech and executive functioning data. Researchers interested in conducting exploratory or hypothesis-driven analyses in the field of individual differences in language and executive functioning, are encouraged to request access to this resource. Applicants will then be provided with an encrypted version of the data which can be downloaded.

Sex differences in intrinsic functional cortical organization reflect differences in network topology rather than cortical morphometry.

Brain size robustly differs between sexes. However, the consequences of this anatomical dimorphism on sex differences in intrinsic brain function remain unclear. We investigated the extent to which sex differences in intrinsic cortical functional organization may be explained by differences in cortical morphometry, namely brain size, microstructure, and the geodesic distances of connectivity profiles. For this, we computed a low dimensional representation of functional cortical organization, the sensory-association axis, and identified widespread sex differences. Contrary to our expectations, observed sex differences in functional organization were not fundamentally associated with differences in brain size, microstructural organization, or geodesic distances, despite these morphometric properties being per se associated with functional organization and differing between sexes. Instead, functional sex differences in the sensory-association axis were associated with differences in functional connectivity profiles and network topology. Collectively, our findings suggest that sex differences in functional cortical organization extend beyond sex differences in cortical morphometry.

Accurate sex prediction of cisgender and transgender individuals without brain size bias.

The increasing use of machine learning approaches on neuroimaging data comes with the important concern of confounding variables which might lead to biased predictions and in turn spurious conclusions about the relationship between the features and the target. A prominent example is the brain size difference between women and men. This difference in total intracranial volume (TIV) can cause bias when employing machine learning approaches for the investigation of sex differences in brain morphology. A TIV-biased model will not capture qualitative sex differences in brain organization but rather learn to classify an individual's sex based on brain size differences, thus leading to spurious and misleading conclusions, for example when comparing brain morphology between cisgender- and transgender individuals. In this study, TIV bias in sex classification models applied to cis- and transgender individuals was systematically investigated by controlling for TIV either through featurewise confound removal or by matching the training samples for TIV. Our results provide strong evidence that models not biased by TIV can classify the sex of both cis- and transgender individuals with high accuracy, highlighting the importance of appropriate modeling to avoid bias in automated decision making.

Voice identity recognition failure in patients with schizophrenia.

Cognitive models propose that auditory verbal hallucinations arise through inner speech misidentification. However, such models cannot explain why the voices in hallucinations often have identities different from the hearer. This study investigated whether a general voice identity recognition difficulty might be present in schizophrenia and related to auditory verbal hallucinations. Twenty-five schizophrenia patients and 13 healthy controls were tested on recognition of famous voices. Signal detection theory was used to calculate perceptual sensitivity and response criterion measures. Schizophrenia patients obtained fewer hits and had lower perceptual sensitivity to detect famous voices than healthy controls did. There were no differences between groups in false alarm rate or response criterion. A symptom-based analysis demonstrated that especially those patients with auditory verbal hallucinations performed poorly in the task. The results indicate that patients with hallucinations are impaired at voice identity recognition because of decreased sensitivity, which may result in inner speech misidentification.

Is it left or is it right? A classification approach for investigating hemispheric differences in low and high dimensionality.

Hemispheric asymmetries, i.e., differences between the two halves of the brain, have extensively been studied with respect to both structure and function. Commonly employed pairwise comparisons between left and right are suitable for finding differences between the hemispheres, but they come with several caveats when assessing multiple asymmetries. What is more, they are not designed for identifying the characterizing features of each hemisphere. Here, we present a novel data-driven framework-based on machine learning-based classification-for identifying the characterizing features that underlie hemispheric differences. Using voxel-based morphometry data from two different samples (n = 226, n = 216), we separated the hemispheres along the midline and used two different pipelines: First, for investigating global differences, we embedded the hemispheres into a two-dimensional space and applied a classifier to assess if the hemispheres are distinguishable in their low-dimensional representation. Second, to investigate which voxels show systematic hemispheric differences, we employed two classification approaches promoting feature selection in high dimensions. The two hemispheres were accurately classifiable in both their low-dimensional (accuracies: dataset 1 = 0.838; dataset 2 = 0.850) and high-dimensional (accuracies: dataset 1 = 0.966; dataset 2 = 0.959) representations. In low dimensions, classification of the right hemisphere showed higher precision (dataset 1 = 0.862; dataset 2 = 0.894) compared to the left hemisphere (dataset 1 = 0.818; dataset 2 = 0.816). A feature selection algorithm in the high-dimensional analysis identified voxels that most contribute to accurate classification. In addition, the map of contributing voxels showed a better overlap with moderate to highly lateralized voxels, whereas conventional t test with threshold-free cluster enhancement best resembled the LQ map at lower thresholds. Both the low- and high-dimensional classifiers were capable of identifying the hemispheres in subsamples of the datasets, such as males, females, right-handed, or non-right-handed participants. Our study indicates that hemisphere classification is capable of identifying the hemisphere in their low- and high-dimensional representation as well as delineating brain asymmetries. The concept of hemisphere classifiability thus allows a change in perspective, from asking what differs between the hemispheres towards focusing on the features needed to identify the left and right hemispheres. Taking this perspective on hemispheric differences may contribute to our understanding of what makes each hemisphere special.

Confound-leakage: confound removal in machine learning leads to leakage.

BACKGROUND: Machine learning (ML) approaches are a crucial component of modern data analysis in many fields, including epidemiology and medicine. Nonlinear ML methods often achieve accurate predictions, for instance, in personalized medicine, as they are capable of modeling complex relationships between features and the target. Problematically, ML models and their predictions can be biased by confounding information present in the features. To remove this spurious signal, researchers often employ featurewise linear confound regression (CR). While this is considered a standard approach for dealing with confounding, possible pitfalls of using CR in ML pipelines are not fully understood. RESULTS: We provide new evidence that, contrary to general expectations, linear confound regression can increase the risk of confounding when combined with nonlinear ML approaches. Using a simple framework that uses the target as a confound, we show that information leaked via CR can increase null or moderate effects to near-perfect prediction. By shuffling the features, we provide evidence that this increase is indeed due to confound-leakage and not due to revealing of information. We then demonstrate the danger of confound-leakage in a real-world clinical application where the accuracy of predicting attention-deficit/hyperactivity disorder is overestimated using speech-derived features when using depression as a confound. CONCLUSIONS: Mishandling or even amplifying confounding effects when building ML models due to confound-leakage, as shown, can lead to untrustworthy, biased, and unfair predictions. Our expose of the confound-leakage pitfall and provided guidelines for dealing with it can help create more robust and trustworthy ML models.

How different types of conceptual relations modulate brain activation during semantic priming.

Semantic priming, a well-established technique to study conceptual representation, has thus far produced variable fMRI results, both regarding the type of priming effects and their correlation with brain activation. The aims of the current study were (a) to investigate two types of semantic relations--categorical versus associative--under controlled processing conditions and (b) to investigate whether categorical and associative relations between words are correlated with response enhancement or response suppression. We used fMRI to examine neural correlates of semantic priming as subjects performed a lexical decision task with a long SOA (800 msec). Four experimental conditions were compared: categorically related trials (couch-bed), associatively related trials (couch-pillow), unrelated trials (couch-bridge), and nonword trials (couch-sibor). We found similar behavioral priming effects for both categorically and associatively related pairs. However, the neural priming effects differed: Categorically related pairs resulted in a neural suppression effect in the right MFG, whereas associatively related pairs resulted in response enhancement in the left IFG. A direct contrast between them revealed activation for categorically related trials in the right insular lobe. We conclude that perceptual and functional similarity of categorically related words may lead to response suppression within right-lateralized frontal regions that represent more retrieval effort and the recruitment of a broader semantic field. Associatively related pairs that require a different processing of the related target compared to the prime may lead to the response enhancement within left inferior frontal regions. Nevertheless, the differences between associative and categorical relations might be parametrical rather than absolutely distinct as both relationships recruit similar regions to a different degree.

Dynamic changes in functional cerebral connectivity of spatial cognition during the menstrual cycle.

Functional cerebral asymmetries (FCAs) in women have been shown to vary with changing levels of sex hormones during the menstrual cycle. Previous studies have suggested that interhemispheric interaction forms a key component in generating FCAs and it has been shown behaviorally and by functional imaging that interhemispheric interaction changes during the menstrual cycle, at least for a left hemisphere dominant task. We used functional MRI and an analysis of functional connectivity to examine whether changes in right hemisphere advantage for a figure comparison task as found in behavioral studies, are based on comparable mechanisms like those identified for the verbal task. Women were examined three times during the menstrual cycle, during the menstrual, follicular and luteal phases. The behavioral data confirmed the right hemisphere advantage for the figure comparison task as well as changes of the right hemisphere advantage during the menstrual cycle. Imaging data showed cycle phase-related changes in lateralized brain activation within the task-dominant hemisphere and changes in connectivity between nonhomotopic areas of both hemispheres, suggesting that changes in functional brain organization in women during the menstrual cycle are not only restricted to hormone-related changes of interhemispheric inhibition between homotopic areas, as has been proposed earlier, but might additionally apply to changes of neuronal processes within the hemispheres which seem to be modulated by heterotopic functional connectivity between hemispheres.

Functional neuroanatomy of sustained memory encoding performance in healthy aging and in Alzheimer's disease.

The aim of our study was to examine brain networks involved with sustaining memory encoding performance in healthy aging and in Alzheimer's disease (AD). Since different brain regions are affected by degradation in these two conditions, it might be conceivable that different compensation mechanisms occur to keep up memory performance in aging and in AD. Using an event-related functional magnetic resonance imaging (FMRI) design and a correlation analysis, 8 patients suffering from AD and 29 elderly control subjects were scanned while they studied a list of words for a subsequent memory test. Individual performance was assessed on the basis of a subsequent recognition test, and brain regions were identified where functional activations during study correlated with memory performance. In both groups, successful memory encoding performance was significantly correlated with the activation of the right frontal cortex. Furthermore, in healthy controls, there was a significant correlation of memory performance and the activation of the left medial and lateral temporal lobe. In contrast, in AD patients, increasing memory performance goes along with increasing activation of the hippocampus and a bilateral brain network including the frontal and temporal cortices. Our data show that in healthy aging and in AD, common and distinct compensatory mechanisms are employed to keep up a certain level of memory performance. Both in healthy aging and in patients with AD, an increased level of monitoring and control processes mediated by the (right) frontal lobe seems to be necessary to maintain a certain level of memory performance. In addition, memory performance in healthy older subjects seems to rely on an increased effort in encoding item-specific semantic and contextual information in lateral areas of the (left) temporal lobe. In AD patients, on the other hand, the maintenance of memory performance is related to an increase of activation of the (left) hippocampus in conjunction with a bilateral network of cortical areas that might be involved with phonological and visual rehearsal of the incoming information.

Sex differences in the brain: More than just male or female.

Sex differences in the brain are widely studied, but results are often inconsistent and it is assumed that many negative findings are not even being reported. The lack of consistent findings might be based on the highly questionable assumption of a clear-cut sexual dimorphism in brain structure and function, that underlies commonly used group comparisons between males and females. Without having to rely on this assumption, state of the art statistical learning methods based on large neuroimaging data sets might offer the tools necessary to disentangle the complex pattern of sex-related variations in brain structure and organization.