Optimizing Performance of Transformer-based Models for Fetal Brain MR Image Segmentation.

"Just Accepted" papers have undergone full peer review and have been accepted for publication in Radiology: Artificial Intelligence. This article will undergo copyediting, layout, and proof review before it is published in its final version. Please note that during production of the final copyedited article, errors may be discovered which could affect the content. Purpose To test transformer-based models' performance when manipulating pretraining weights, dataset size, input size and comparing the best-model with reference standard and state-of-the-art models for a resting-state functional (rs-fMRI) fetal brain extraction task. Materials and Methods An internal retrospective dataset (fetuses = 172; images = 519; collected from 2018-2022) was used to investigate influence of dataset size, pretraining approaches and image input size on Swin-UNETR and UNETR models. The internal and an external (fetuses = 131; images = 561) datasets were used to cross-validate and to assess generalization capability of the best model against state-of-the-art models on different scanner types and number of gestational weeks (GW). The Dice similarity coefficient (DSC) and the Balanced average Hausdorff distance (BAHD) were used as segmentation performance metrics. GEE multifactorial models were used to assess significant model and interaction effects of interest. Results Swin-UNETR was not affected by pretraining approach and dataset size and performed best with the mean dataset image size, with a mean DSC of 0.92 and BAHD of 0.097. The Swin-UNETR was not affected by scanner type. Generalization results on the internal dataset showed that Swin-UNETR had lower performances compared with reference standard models and comparable performances on the external dataset. Cross-validation on internal and external test sets demonstrated better and comparable performance of Swin-UNETR versus convolutional neural network architectures during the late-fetal period (GWs > 25) but lower performance during the midfetal period (GWs ≤ 25). Conclusion Swin-UNTER showed flexibility in dealing with smaller datasets, regardless of pretraining approaches. For fetal brain extraction of rs-fMRI, Swin-UNTER showed comparable performance with reference standard models during the late-fetal period and lower performance during the early GW period. ©RSNA, 2024.

The maternal-fetal neurodevelopmental groundings of preterm birth risk.

Background: Altered neurodevelopment is a major clinical sequela of Preterm Birth (PTB) being currently unexplored in-utero. Aims: To study the link between fetal brain functional (FbF) connectivity and preterm birth, using resting-state functional magnetic resonance imaging (rs-fMRI). Study design: Prospective single-centre cohort study. Subjects: A sample of 31 singleton pregnancies at 28-34 weeks assigned to a low PTB risk (LR) (n = 19) or high PTB risk (HR) (n = 12) group based on a) the Maternal Frailty Inventory (MaFra) for PTB risk; b) a case-specific PTB risk gradient. Methods: Fetal brain rs-fMRI was performed on 1.5T MRI scanner. First, directed causal relations representing fetal brain functional connectivity measurements were estimated using the Greedy Equivalence Search (GES) algorithm. HR vs. LR group differences were then tested with a novel ad-hoc developed Monte Carlo permutation test. Second, a MaFra-only random forest (RF) was compared against a MaFra-Neuro RF, trained by including also the most important fetal brain functional connections. Third, correlation and regression analyses were performed between MaFra-Neuro class probabilities and i) the GA at birth; ii) PTB risk gradient, iii) perinatal clinical conditions and iv) PTB below 37 weeks. Results: First, fewer fetal brain functional connections were evident in the HR group. Second, the MaFra-Neuro RF improved PTB risk prediction. Third, MaFra-Neuro class probabilities showed a significant association with: i) GA at birth; ii) PTB risk gradient, iii) perinatal clinical conditions and iv) PTB below 37 weeks. Conclusion: Fetal brain functional connectivity is a novel promising predictor of PTB, linked to maternal risk profiles, ahead of birth, and clinical markers of neurodevelopmental risk, at birth, thus potentially "connecting" different PTB phenotypes.

Maternal anxiety-driven modulation of fetal limbic connectivity designs a backbone linking neonatal brain functional topology to socio-emotional development in early childhood.

A link between maternal anxiety during pregnancy and adverse socio-emotional outcomes in childhood has been consistently sustained on the very early neurodevelopmental alteration of structural pathways between fetal limbic and cortical brain regions. In this study, we provide follow-up evidence for a feed-forward model linking (i) maternal anxiety, (ii) fetal functional neurodevelopment, (iii) neonatal functional network organization with (iv) socio-emotional neurobehavioral development in early childhood. Namely, we investigate a sample of 16 mother-fetus dyads and show how a maternal state-trait anxiety profile with pregnancy-specific worries can significantly influence functional synchronization patterns between regions of the fetal limbic system (i.e., hippocampus and amygdala) and the neocortex, as assessed through resting-state functional magnetic resonance imaging. Generalization of the findings was supported by leave-one-out cross-validation. We further show how this maternal-fetal cross-talk propagates to functional network topology in the neonate, specifically targeting connector hubs, and further maps onto socio-emotional profiles, assessed through Bayley-III socio-emotional scale in early childhood (i.e., in the 12-24 months range). Based on this evidence, we put forward the hypothesis of a "Maternal-Fetal-Neonatal Anxiety Backbone", through which neurobiological changes driven by maternal anxiety could trigger a divergence in the establishment of a cognitive-emotional development blueprint, in terms of the nascent functional homeostasis between bottom-up limbic and top-down higher-order neuronal circuitry.

RS-FetMRI: a MATLAB-SPM Based Tool for Pre-processing Fetal Resting-State fMRI Data.

Resting-state functional magnetic resonance imaging (rs-fMRI) most recently has proved to open a measureless window on functional neurodevelopment in utero. Fetal brain activation and connectivity maps can be heavily influenced by 1) fetal-specific motion effects on the time-series and 2) the accuracy of time-series spatial normalization to a standardized gestational-week (GW) specific fetal template space.Due to the absence of a standardized and generalizable image processing protocol, the objective of the present work was to implement a validated fetal rs-fMRI preprocessing pipeline (RS-FetMRI) divided into 6 inter-dependent preprocessing modules (i.e., M1 to M6) and designed to work entirely as an extension for Statistical Parametric Mapping (SPM).RS-FetMRI pipeline output analyses on rs-fMRI time-series sampled from a cohort of fetuses acquired on both 1.5 T and 3 T MRI scanning systems showed increased efficacy of estimation of the degree of movement coupled with an efficient motion censoring procedure, resulting in increased number of motion-uncorrupted volumes and temporal continuity in fetal rs-fMRI time-series data. Moreover, a "structural-free" SPM-based spatial normalization procedure granted a high degree of spatial overlap with high reproducibility and a significant improvement in whole-brain and parcellation-specific Temporal Signal-to-Noise Ratio (TSNR) mirrored by functional connectivity analysis.To our knowledge, the RS-FetMRI pipeline is the first semi-automatic and easy-to-use standardized fetal rs-fMRI preprocessing pipeline completely integrated in MATLAB-SPM able to remove entry barriers for new research groups into the field of fetal rs-fMRI, for both research or clinical purposes, and ultimately to make future fetal brain connectivity investigations more suitable for comparison and cross-validation.

In search of different categories of abstract concepts: a fMRI adaptation study.

Concrete conceptual knowledge is supported by a distributed neural network representing different semantic features according to the neuroanatomy of sensory and motor systems. If and how this framework applies to abstract knowledge is currently debated. Here we investigated the specific brain correlates of different abstract categories. After a systematic a priori selection of brain regions involved in semantic cognition, i.e. responsible of, respectively, semantic representations and cognitive control, we used a fMRI-adaptation paradigm with a passive reading task, in order to modulate the neural response to abstract (emotions, cognitions, attitudes, human actions) and concrete (biological entities, artefacts) categories. Different portions of the left anterior temporal lobe responded selectively to abstract and concrete concepts. Emotions and attitudes adapted the left middle temporal gyrus, whereas concrete items adapted the left fusiform gyrus. Our results suggest that, similarly to concrete concepts, some categories of abstract knowledge have specific brain correlates corresponding to the prevalent semantic dimensions involved in their representation.

A hierarchical procedure to select intrauterine and extrauterine factors for methodological validation of preterm birth risk estimation.

BACKGROUND: Etiopathogenesis of preterm birth (PTB) is multifactorial, with a universe of risk factors interplaying between the mother and the environment. It is of utmost importance to identify the most informative factors in order to estimate the degree of PTB risk and trace an individualized profile. The aims of the present study were: 1) to identify all acknowledged risk factors for PTB and to select the most informative ones for defining an accurate model of risk prediction; 2) to verify predictive accuracy of the model and 3) to identify group profiles according to the degree of PTB risk based on the most informative factors. METHODS: The Maternal Frailty Inventory (MaFra) was created based on a systematic review of the literature including 174 identified intrauterine (IU) and extrauterine (EU) factors. A sample of 111 pregnant women previously categorized in low or high risk for PTB below 37 weeks, according to ACOG guidelines, underwent the MaFra Inventory. First, univariate logistic regression enabled p-value ordering and the Akaike Information Criterion (AIC) selected the model including the most informative MaFra factors. Second, random forest classifier verified the overall predictive accuracy of the model. Third, fuzzy c-means clustering assigned group membership based on the most informative MaFra factors. RESULTS: The most informative and parsimonious model selected through AIC included Placenta Previa, Pregnancy Induced Hypertension, Antibiotics, Cervix Length, Physical Exercise, Fetal Growth, Maternal Anxiety, Preeclampsia, Antihypertensives. The random forest classifier including only the most informative IU and EU factors achieved an overall accuracy of 81.08% and an AUC of 0.8122. The cluster analysis identified three groups of typical pregnant women, profiled on the basis of the most informative IU and EU risk factors from a lower to a higher degree of PTB risk, which paralleled time of birth delivery. CONCLUSIONS: This study establishes a generalized methodology for building-up an evidence-based holistic risk assessment for PTB to be used in clinical practice. Relevant and essential factors were selected and were able to provide an accurate estimation of degree of PTB risk based on the most informative constellation of IU and EU factors.

The effects of the functional interplay between the Default Mode and Executive Control Resting State Networks on cognitive outcome in preterm born infants at 6 months of age.

Preterm birth can affect cognitive functions, such as attention or more generally executive control mechanisms, with severity in impairments proportional to prematurity. The functional cross-talk between the Default Mode (DMN) and Executive Control (ECN) networks mirrors the integrity of cognitive processing and is directly related to brain development. In this study, a cohort of 20 preterm-born infants was investigated using rs-fMRI. First, we addressed biological maturity of the DMN per se and its interplay with the ECN in terms of patterns of increased functional connectivity. Second, we assessed the impact of the degree of prematurity on the DMN-ECN functional interplay development in relation to cognitive outcome at six months. Our results highlighted the emergence of DMN in preterm neonates, with connectivity strength and synchronization between the anterior DMN hub and frontal areas increasing as a function of biological maturity. Further, cognitive scores at 6 months were predicted by mPFC-ECN connectivity strength with degree of prematurity impacting on mPFC-ECN connectivity and triggering differential patterns of functional maturation of the ECN for very early/early and moderate/late preterm neonates. Our findings suggest that the prematurity window allows to observe precursors of functional plasticity that may underlie different developmental trajectories in preterm children.

Subcortico-Cortical Functional Connectivity in the Fetal Brain: A Cognitive Development Blueprint.

Recent evidence has shown that patterns of cortico-cortical functional synchronization are consistently traceable by the end of the third trimester of pregnancy. The involvement of subcortical structures in early functional and cognitive development has never been explicitly investigated, notwithstanding their pivotal role in different cognitive processes. We address this issue by exploring subcortico-cortical functional connectivity at rest in a group of normally developing fetuses between the 25th and 32nd weeks of gestation. Results show significant functional coupling between subcortical nuclei and cortical networks related to: (i) sensorimotor processing, (ii) decision making, and (iii) learning capabilities. This functional maturation framework unearths a Cognitive Development Blueprint, according to which grounding cognitive skills are planned to develop with higher ontogenetic priority. Specifically, our evidence suggests that a newborn already possesses the ability to: (i) perceive the world and interact with it, (ii) create salient representations for the selection of adaptive behaviors, and (iii) store, retrieve, and evaluate the outcomes of interactions, in order to gradually improve adaptation to the extrauterine environment.

The left inferior frontal gyrus: A neural crossroads between abstract and concrete knowledge.

Evidence from both neuropsychology and neuroimaging suggests that different types of information are necessary for representing and processing concrete and abstract word meanings. Both abstract and concrete concepts, however, conjointly rely on perceptual, verbal and contextual knowledge, with abstract concepts characterized by low values of imageability (IMG) (low sensory-motor grounding) and low context availability (CA) (more difficult to contextualize). Imaging studies supporting differences between abstract and concrete concepts show a greater recruitment of the left inferior frontal gyrus (LIFG) for abstract concepts, which has been attributed either to the representation of abstract-specific semantic knowledge or to the request for more executive control than in the case of concrete concepts. We conducted an fMRI study on 27 participants, using a lexical decision task involving both abstract and concrete words, whose IMG and CA values were explicitly modelled in separate parametric analyses. The LIFG was significantly more activated for abstract than for concrete words, and a conjunction analysis showed a common activation for words with low IMG or low CA only in the LIFG, in the same area reported for abstract words. A regional template map of brain activations was then traced for words with low IMG or low CA, and BOLD regional time-series were extracted and correlated with the specific LIFG neural activity elicited for abstract words. The regions associated to low IMG, which were functionally correlated with LIFG, were mainly in the left hemisphere, while those associated with low CA were in the right hemisphere. Finally, in order to reveal which LIFG-related network increased its connectivity with decreases of IMG or CA, we conducted generalized psychophysiological interaction analyses. The connectivity strength values extracted from each region connected with the LIFG were correlated with specific LIFG neural activity for abstract words, and a regression analysis was conducted to highlight which areas recruited by low IMG or low CA predicted the greater activation of the IFG for abstract concepts. Only the left middle temporal gyrus/angular gyrus, known to be involved in semantic processing, was a significant predictor of LIFG activity differentiating abstract from concrete words. The results show that the abstract conceptual processing requires the interplay of multiple brain regions, necessary for both the intrinsic and extrinsic properties of abstract knowledge. The LIFG can be thus identified as the neural crossroads between different types of information equally necessary for representing processing and differentiating abstract concepts from concrete ones.

Interference and conflict monitoring in individuals with amnestic mild cognitive impairment: A structural study of the anterior cingulate cortex.

Amnestic mild cognitive impairment (aMCI) is a clinical condition characterized by memory impairment in the absence of any other cognitive impairment and is commonly associated with high conversion to Alzheimer's disease. Recent evidence shows that executive functions and selective attention mechanisms could also be impaired in aMCI. In this study, we investigated performance differences (i.e., reaction times [RTs] and accuracy) between a group of aMCI participants and a group of age-matched healthy individuals on the attentional network task (ANT) focusing on situations with increased interference. In particular, we assessed the relationship between interference and conflict effects and grey matter volumes (GMVs) of the anterior cingulate cortex (ACC)/pre-supplementary motor area in the entire sample because of its crucial role in conflict monitoring. When compared with controls, aMCI participants were less accurate on the ANT, showing increased interference and conflict effects, but no differences in RTs. In addition, aMCI participants exhibited lower GMV in the ACC than controls. While better accuracy for interference and conflict effects was associated with an increase of GMV in the ACC for both groups, RTs from the interference effect were negatively correlated with GMV of the ACC only in aMCI participants. In other words, lower GMV values of the ACC were paralleled with significantly impaired performance in terms of interference resolution. In conclusion, our study suggests the presence of a selective impairment in interference and conflict monitoring in aMCI, which in turn is associated with decreased GMVs in the ACC.

Semantic interference and its control: A functional neuroimaging and connectivity study.

During picture naming, the ease with which humans generate words is dependent upon the context in which they are named. For instances, naming previously presented items results in facilitation. Instead, naming a picture semantically related to previous items displays persistent interference effects (i.e., cumulative semantic interference, CSI). The neural correlates of CSI are still unclear and it is a matter of debate whether semantic control, or cognitive control more in general, is necessary for the resolution of CSI. We carried out an event-related fMRI experiment to assess the neural underpinnings of the CSI effect and the involvement and nature of semantic control. Both left inferior frontal gyrus (LIFG) and the left caudate nucleus (LCN) showed a linear increase of BOLD response positively associated with the consecutive number of presentations of semantically related pictures independently of task-load. The generalized psychophysiological interaction analysis showed that LIFG demonstrated a quantitative neural connectivity difference with the left supramarginal and angular gyri for increases of task-load and with the fusiform gyri for linear CSI increases. Furthermore, seed-to-voxel functional connectivity showed that LIFG activity coupled with different regions involved in cognitive control and lexicosemantic processing when semantic interference was elicited to a minimum or maximum degree. Our results are consistent with the lexical-competitive nature of the CSI effect, and we provide novel evidence that semantic control lies upon a more general cognitive control network (i.e., LIFG and LCN) responsible for resolving interference between competing semantically related items through connectivity with different brain areas in order to guarantee the correct response. Hum Brain Mapp 37:4179-4196, 2016. © 2016 Wiley Periodicals, Inc.

Language Control in Bilinguals: Monitoring and Response Selection.

Language control refers to the cognitive mechanism that allows bilinguals to correctly speak in one language avoiding interference from the nontarget language. Bilinguals achieve this feat by engaging brain areas closely related to cognitive control. However, 2 questions still await resolution: whether this network is differently engaged when controlling nonlinguistic representations, and whether this network is differently engaged when control is exerted upon a restricted set of lexical representations that were previously used (i.e., local control) as opposed to control of the entire language system (i.e., global control). In the present event-related functional magnetic resonance imaging study, we investigated these 2 questions by employing linguistic and nonlinguistic blocked switching tasks in the same bilingual participants. We first report that the left prefrontal cortex is driven similarly for control of linguistic and nonlinguistic representations, suggesting its domain-general role in the implementation of response selection. Second, we propose that language control in bilinguals is hierarchically organized with the dorsal anterior cingulate cortex/presupplementary motor area acting as the supervisory attentional system, recruited for increased monitoring demands such as local control in the second language. On the other hand, prefrontal, inferior parietal areas and the caudate would act as the response selection system, tailored for language selection for both local and global control.

Bilingualism provides a neural reserve for aging populations.

It has been postulated that bilingualism may act as a cognitive reserve and recent behavioral evidence shows that bilinguals are diagnosed with dementia about 4-5 years later compared to monolinguals. In the present study, we investigated the neural basis of these putative protective effects in a group of aging bilinguals as compared to a matched monolingual control group. For this purpose, participants completed the Erikson Flanker task and their performance was correlated to gray matter (GM) volume in order to investigate if cognitive performance predicts GM volume specifically in areas affected by aging. We performed an ex-Gaussian analysis on the resulting RTs and report that aging bilinguals performed better than aging monolinguals on the Flanker task. Bilingualism was overall associated with increased GM in the ACC. Likewise, aging induced effects upon performance correlated only for monolinguals to decreased gray matter in the DLPFC. Taken together, these neural regions might underlie the benefits of bilingualism and act as a neural reserve that protects against the cognitive decline that occurs during aging.

Computerized neuropsychological assessment in aging: testing efficacy and clinical ecology of different interfaces.

Digital technologies have opened new opportunities for psychological testing, allowing new computerized testing tools to be developed and/or paper and pencil testing tools to be translated to new computerized devices. The question that rises is whether these implementations may introduce some technology-specific effects to be considered in neuropsychological evaluations. Two core aspects have been investigated in this work: the efficacy of tests and the clinical ecology of their administration (the ability to measure real-world test performance), specifically (1) the testing efficacy of a computerized test when response to stimuli is measured using a touch-screen compared to a conventional mouse-control response device; (2) the testing efficacy of a computerized test with respect to different input modalities (visual versus verbal); and (3) the ecology of two computerized assessment modalities (touch-screen and mouse-control), including preference measurements of participants. Our results suggest that (1) touch-screen devices are suitable for administering experimental tasks requiring precise timings for detection, (2) intrinsic nature of neuropsychological tests should always be respected in terms of stimuli presentation when translated to new digitalized environment, and (3) touch-screen devices result in ecological instruments being proposed for the computerized administration of neuropsychological tests with a high level of preference from elderly people.

Bilingualism protects anterior temporal lobe integrity in aging.

Cerebral gray-matter volume (GMV) decreases in normal aging but the extent of the decrease may be experience-dependent. Bilingualism may be one protective factor and in this article we examine its potential protective effect on GMV in a region that shows strong age-related decreases-the left anterior temporal pole. This region is held to function as a conceptual hub and might be expected to be a target of plastic changes in bilingual speakers because of the requirement for these speakers to store and differentiate lexical concepts in 2 languages to guide speech production and comprehension processes. In a whole brain comparison of bilingual speakers (n = 23) and monolingual speakers (n = 23), regressing out confounding factors, we find more extensive age-related decreases in GMV in the monolingual brain and significantly increased GMV in left temporal pole for bilingual speakers. Consistent with a specific neuroprotective effect of bilingualism, region of interest analyses showed a significant positive correlation between naming performance in the second language and GMV in this region. The effect appears to be bilateral though because there was a nonsignificantly different effect of naming performance on GMV in the right temporal pole. Our data emphasize the vulnerability of the temporal pole to normal aging and the value of bilingualism as both a general and specific protective factor to GMV decreases in healthy aging.