How to establish robust brain–behavior relationships without thousands of individuals

Nature Neuroscience, Journal Year: 2022, Volume and Issue: 25(7), P. 835 - 837

Published: June 16, 2022

Language: Английский

---

Evidence for embracing normative modeling

eLife, Journal Year: 2023, Volume and Issue: 12

Published: March 13, 2023

In this work, we expand the normative model repository introduced in Rutherford et al., 2022a to include models charting lifespan trajectories of structural surface area and brain functional connectivity, measured using two unique resting-state network atlases (Yeo-17 Smith-10), an updated online platform for transferring these new data sources. We showcase value with a head-to-head comparison between features output by modeling raw several benchmarking tasks: mass univariate group difference testing (schizophrenia versus control), classification regression (predicting general cognitive ability). Across all benchmarks, show advantage features, strongest statistically significant results demonstrated tasks. intend accessible resources facilitate wider adoption across neuroimaging community.

Language: Английский

---

Task fMRI paradigms may capture more behaviorally relevant information than resting-state functional connectivity

NeuroImage, Journal Year: 2023, Volume and Issue: 270, P. 119946 - 119946

Published: Feb. 17, 2023

Characterizing the optimal fMRI paradigms for detecting behaviorally relevant functional connectivity (FC) patterns is a critical step to furthering our knowledge of neural basis behavior. Previous studies suggested that FC derived from task paradigms, which we refer as task-based FC, are better correlated with individual differences in behavior than resting-state but consistency and generalizability this advantage across conditions was not fully explored. Using data three tasks Adolescent Brain Cognitive Development Study ® (ABCD), tested whether observed improvement behavioral prediction power can be attributed changes brain activity induced by design. We decomposed time course each into model fit (the fitted condition regressors single-subject general linear model) residuals, calculated their respective compared performance these estimates original FC. The residual at predicting measure cognitive ability or two measures on tasks. superior content-specific insofar it only probed similar constructs predicted interest. To surprise, parameters, beta regressors, were equally if more predictive all measures. These results showed afforded largely driven associated Together previous studies, findings highlighted importance design eliciting meaningful activation patterns.

Language: Английский

---

Individual differences in computational psychiatry: A review of current challenges

Neuroscience & Biobehavioral Reviews, Journal Year: 2023, Volume and Issue: 148, P. 105137 - 105137

Published: March 20, 2023

Bringing precision to the understanding and treatment of mental disorders requires instruments for studying clinically relevant individual differences. One promising approach is development computational assays: integrating models with cognitive tasks infer latent patient-specific disease processes in brain computations. While recent years have seen many methodological advancements modelling cross-sectional patient studies, much less attention has been paid basic psychometric properties (reliability construct validity) measures provided by assays. In this review, we assess extent issue examining emerging empirical evidence. We find that suffer from poor properties, which poses a risk invalidating previous findings undermining ongoing research efforts using assays study (and even group) provide recommendations how address these problems and, crucially, embed them within broader perspective on key developments are needed translating clinical practice.

Language: Английский

---

The Burden of Reliability: How Measurement Noise Limits Brain-Behaviour Predictions

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: Feb. 10, 2023

Abstract Major efforts in human neuroimaging strive to understand individual differences and find biomarkers for clinical applications by predicting behavioural phenotypes from brain imaging data. An essential prerequisite identifying generalizable replicable brain-behaviour prediction models is sufficient measurement reliability. However, the selection of targets predominantly guided scientific interest or data availability rather than reliability considerations. Here we demonstrate impact low phenotypic on out-of-sample performance. Using simulated empirical Human Connectome Projects, found that levels common across many can markedly limit ability link behaviour. Next, using 5000 subjects UK Biobank, show only highly reliable fully benefit increasing sample sizes hundreds thousands participants. Overall, our findings highlight importance brain–behaviour associations differences.

Language: Английский

---

Systematic evaluation of fMRI data-processing pipelines for consistent functional connectomics

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: June 4, 2024

Abstract Functional interactions between brain regions can be viewed as a network, enabling neuroscientists to investigate function through network science. Here, we systematically evaluate 768 data-processing pipelines for reconstruction from resting-state functional MRI, evaluating the effect of parcellation, connectivity definition, and global signal regression. Our criteria seek that minimise motion confounds spurious test-retest discrepancies topology, while being sensitive both inter-subject differences experimental effects interest. We reveal vast systematic variability across pipelines’ suitability connectomics. Inappropriate choice pipeline produce results are not only misleading, but so, with majority failing at least one criterion. However, set optimal consistently satisfy all different datasets, spanning minutes, weeks, months. provide full breakdown each pipeline’s performance inform future best practices in

Language: Английский

---

Machine learning prediction of cognition from functional connectivity: Are feature weights reliable?

NeuroImage, Journal Year: 2021, Volume and Issue: 245, P. 118648 - 118648

Published: Oct. 20, 2021

Cognitive performance can be predicted from an individual's functional brain connectivity with modest accuracy using machine learning approaches. As yet, however, predictive models have arguably yielded limited insight into the neurobiological processes supporting cognition. To do so, feature selection and weight estimation need to reliable ensure that important connections circuits high utility reliably identified. We comprehensively investigate test-retest reliability for various of cognitive built resting-state networks in healthy young adults (n=400). Despite achieving prediction accuracies (r=0.2–0.4), we find is generally poor all (ICC< 0.3), significantly poorer than overt biological attributes such as sex (ICC≈0.5). Larger sample sizes (n=800), Haufe transformation, non-sparse selection/regularization smaller spaces marginally improve 0.4). elucidate a tradeoff between univariate statistics are more weights models. Finally, show measuring agreement cross-validation folds provides inflated estimates reliability. thus recommend estimated out-of-sample, if possible. argue rebalancing focus model may facilitate mechanistic understanding cognition

Language: Английский

---

Exploring MEG brain fingerprints: Evaluation, pitfalls, and interpretations

NeuroImage, Journal Year: 2021, Volume and Issue: 240, P. 118331 - 118331

Published: July 5, 2021

Individual characterization of subjects based on their functional connectome (FC), termed "FC fingerprinting", has become a highly sought-after goal in contemporary neuroscience research. Recent magnetic resonance imaging (fMRI) studies have demonstrated unique and accurate identification individuals as an accomplished task. However, FC fingerprinting magnetoencephalography (MEG) data is still widely unexplored. Here, we study resting-state MEG from the Human Connectome Project to assess its relationship with several factors including amplitude- phase-coupling connectivity measures, spatial leakage correction, frequency bands, behavioral significance. To this end, first employ two scoring methods, differential identifiability success rate, provide quantitative fingerprint scores for each measurement. Secondly, explore edgewise nodal patterns across different bands (delta, theta, alpha, beta, gamma). Finally, investigate cross-modality obtained fMRI recordings same subjects. We significance measures modalities using partial least square correlation analyses. Our results suggest that performance heavily dependent measure, band, method, correction. report higher performances central (alpha beta), visual, frontoparietal, dorsal-attention, default-mode networks. Furthermore, comparisons reveal certain degree concordance between data, especially visual system. multivariate analyses show connectomes strong significance, which however depends considered measure temporal scale. This comprehensive, albeit preliminary investigation test-retest offers relation methodological electrophysiological contributes understanding cross-modal relationships. hope will contribute setting grounds identification.

Language: Английский

---

Individualized event structure drives individual differences in whole-brain functional connectivity

NeuroImage, Journal Year: 2022, Volume and Issue: 252, P. 118993 - 118993

Published: Feb. 19, 2022

Resting-state functional connectivity is typically modeled as the correlation structure of whole-brain regional activity. It studied widely, both to gain insight into brain's intrinsic organization but also develop markers sensitive changes in an individual's cognitive, clinical, and developmental state. Despite this, origins drivers connectivity, especially at level densely sampled individuals, remain elusive. Here, we leverage novel methodology decompose its precise framewise contributions. Using two dense sampling datasets, investigate individualized focusing specifically on role brain network "events" - short-lived peaked patterns high-amplitude cofluctuations. a statistical test identify events empirical recordings. We show that cofluctuation expressed during are repeated across multiple scans same individual represent idiosyncratic variants template group level. Lastly, propose simple model based event cofluctuations, demonstrating group-averaged cofluctuations suboptimal for explaining participant-specific connectivity. Our work complements recent studies implicating brief instants primary static, extends those studies, individualized, positing dynamic basis

Language: Английский

---

What is the best brain state to predict autistic traits?

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 17, 2025

Abstract Autism is a heterogeneous condition, and functional magnetic resonance imaging-based studies have advanced understanding of neurobiological correlates autistic features. Nevertheless, little work has focused on the optimal brain states to reveal brain-phenotype relationships. In addition, there need better understand relevance attentional abilities in mediating Using connectome-based predictive modelling, we interrogate three datasets determine scanning conditions that can boost prediction clinically relevant phenotypes assess generalizability. dataset one, sample youth with autism neurotypical participants, find sustained attention task (the gradual onset continuous performance task) results high traits compared free-viewing social resting-state condition. two, observe network model generated from generalizes predict measures adults. three, show same one further responsiveness data Brain Imaging Data Exchange. sum, our suggest an in-scanner challenge help delineate robust markers support continued investigation under which psychiatric conditions.

Language: Английский

---