How to remove or control confounds in predictive models, with applications to brain biomarkers
Open accessThe DOI matches its Crossref record; 70 of the 70 references it cites were found in a registry or an index.
Abstract
Background: With increasing data sizes and more easily available computational methods, neurosciences rely more and more on predictive modeling with machine learning, e.g., to extract disease biomarkers. Yet, a successful prediction may capture a confounding effect correlated with the outcome instead of brain features specific to the outcome of interest. For instance, because patients tend to move more in the scanner than controls, imaging biomarkers of a disease condition may mostly reflect head motion, leading to inefficient use of resources and wrong interpretation of the biomarkers. Results: Here we study how to adapt statistical methods that control for confounds to predictive modeling settings. We review how to train predictors that are not driven by such spurious effects. We also show how to measure the unbiased predictive accuracy of these biomarkers, based on a confounded dataset. For this purpose, cross-validation must be modified to account for the nuisance effect. To guide understanding and practical recommendations, we apply various strategies to assess predictive models in the presence of confounds on simulated data and population brain imaging settings. Theoretical and empirical studies show that deconfounding should not be applied to the train and test data jointly: modeling the effect of confounds, on the training data only, should instead be decoupled from removing confounds. Conclusions: Cross-validation that isolates nuisance effects gives an additional piece of information: confound-free prediction accuracy.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures












Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source doesn’t support the claim as written.
In the citing paper
“For example, regressing the covariates/confounds out of the whole dataset in a non-cross-validated manner has been shown to negatively affect prediction performance40.”
From Data leakage inflates prediction performance in connectome-based machine learning models (2024) · cites this paper as [40]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“However, the pre-fusion exclusion of biased subpopulation creates an artificial construct that fails to capture the true population-level associations, while diminishing available sample size and data utility in neuroimaging research, thereby severely limiting the generalizability of findings, particularly in small-sample studies.”
From Confound Controlled Multimodal Neuroimaging Data Fusion and Its Application to Developmental Disorders (2025) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Our results corroborate previous works, suggesting that covariate regression must be performed within the cross-validation loop to avoid false deflation of effect sizes38–40.”
From Data leakage inflates prediction performance in connectome-based machine learning models (2024) · cites this paper as [40]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Cross-validated and non-cross-validated covariate regression have been compared in several previous studies38,40, and we build upon those previous works here by evaluating them in three phenotypes and four datasets through the view of leakage.”
From Data leakage inflates prediction performance in connectome-based machine learning models (2024) · cites this paper as [40]
The cited source says what the claim attributes to it.
In the citing paper
“In addition, we performed cross-validated covariate regression, where several covariates were regressed from the functional connectivity data within the cross-validation scheme38,40.”
From Data leakage inflates prediction performance in connectome-based machine learning models (2024) · cites this paper as [40]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Sudlow, C., Gallacher, J., Allen, N., Beral, V., Burton, P., Danesh, J., Downey, P., Elliott, P., Green, J., Landray, M., et al. (2015) · PLOS Medicine · doi:10.1371/journal.pmed.1001779 · Cited by 12,728 · cited 1×
“(cid:2) The UK Biobank (UKBB) project [51] is a prospective epidemiological study to elucidate the development of diseases of the UK population over the years.”
Statistical parametric maps in functional imaging: A general linear approach
Friston, K. J., Holmes, A. P., Worsley, K. J., Poline, J. ‐., Frith, C. D., & Frackowiak, R. S. J. (1994) · Human Brain Mapping · doi:10.1002/hbm.460020402 · Cited by 9,688 · cited 2×
“In the context of the GLM,an alternative implementation relies on removing the effect of variables that are correlated [25].”
Mostly Harmless Econometrics: An Empiricist's Companion
Angrist, J. D., & Pischke, J.-S. (2009) · doi:10.1515/9781400829828 · Cited by 9,467
Power, J. D., Barnes, K. A., Snyder, A. Z., Schlaggar, B. L., & Petersen, S. E. (2012) · NeuroImage · doi:10.1016/j.neuroimage.2011.10.018 · Cited by 7,429 · cited 2×
“For instance, brain imaging reflects age quite accurately and actually carries information about age-related diseases [8,10,11], yet [12] showed that participants’ in-scanner motion varies with age and it creates systematic differences in recorded brain imaging signals.”
Mostly Harmless Econometrics: An Empiricist's Companion
Angrist, J. D., & Pischke, J.-S. (2008) · doi:10.2307/j.ctvcm4j72 · Cited by 5,051
Statistical Modeling: The Two Cultures (with comments and a rejoinder by the author)
Breiman, L. (2001) · Statistical Science · doi:10.1214/ss/1009213726 · Cited by 4,340 · cited 1×
“Often, a predictive analysis does not build on a generative model of the signal but on optimizing discrimination [18].”
Non-Uniform Random Variate Generation
Devroye, L. (1986) · Springer eBooks · doi:10.1007/978-1-4613-8643-8 · Cited by 4,041
Estimating Causal Effects from Large Data Sets Using Propensity Scores
Rubin, D. B. (1997) · Annals of Internal Medicine · doi:10.7326/0003-4819-127-8_part_2-199710151-00064 · Cited by 2,877 · cited 1×
“For in-sample analysis, propensity scores are a classic reweighting technique used to obtain reliable effect estimates from confounded datasets [68,69].”
Causality: models, reasoning, and inference
(2000) · Choice Reviews Online · doi:10.5860/choice.38-1494a · Cited by 2,873
Estimation of Average Treatment Effects Based on Propensity Scores
Becker, S. O., & Ichino, A. (2002) · The Stata Journal: Promoting communications on statistics and Stata · doi:10.1177/1536867x0200200403 · Cited by 2,754 · cited 1×
“For in-sample analysis, propensity scores are a classic reweighting technique used to obtain reliable effect estimates from confounded datasets [68,69].”
Cited by
Data leakage inflates prediction performance in connectome-based machine learning models
Rosenblatt, M., Tejavibulya, L., Jiang, R., Noble, S., & Scheinost, D. (2024) · Nature Communications · doi:10.1038/s41467-024-46150-w · Cited by 236
“In addition, we performed cross-validated covariate regression, where several covariates were regressed from the functional connectivity data within the cross-validation scheme 38,40 .”
InfluentialWang, R., Chaudhari, P., & Davatzikos, C. (2023) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2211613120 · Cited by 78
Biological markers and psychosocial factors predict chronic pain conditions
Fillingim, M., Tanguay-Sabourin, C., Parisien, M., Zare, A., Guglietti, G. V., Norman, J., Petre, B., Bortsov, A. V., Ware, M. A., Pérez, J., et al. (2025) · Nature Human Behaviour · doi:10.1038/s41562-025-02156-y · Cited by 69
Wang, H.-T., Meisler, S. L., Sharmarke, H., Clarke, N., Gensollen, N., Markiewicz, C. J., Paugam, F., Thirion, B., & Bellec, P. (2024) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1011942 · Cited by 56
“One such key challenge is the effective correction of non-neuronal sources of fluctuations (called confounds), known as denoising, which is important to reduce bias when studying the association between connectomes and behavioral measures of interest [5].”
Self-citationChopra, S., Dhamala, E., Lawhead, C., Ricard, J. A., Orchard, E. R., An, L., Chen, P., Wulan, N., Kumar, P., Rubenstein, A., et al. (2024) · Science Advances · doi:10.1126/sciadv.adn1862 · Cited by 43
“For each of the 100 train-test splits, the variables were first regressed out of the global cognition training data, and the resulting regression coefficients were used to residualize the global cognition test data ( 88 ), after which the entire prediction modeling procedure was repeated.”
Machine learning of brain-specific biomarkers from EEG
Bomatter, P., Paillard, J., Garcés, P., Hipp, J. F., & Engemann, D. A. (2024) · EBioMedicine · doi:10.1016/j.ebiom.2024.105259 · Cited by 44
Statistical quantification of confounding bias in machine learning models
Spisak, T. (2022) · GigaScience · doi:10.1093/gigascience/giac082 · Cited by 34
“In a number of cases, removing or controlling for a confounder can remove variance of interest and complicate model interpretations [24, 29, 30], rendering the choice of confound mitigation strategy as one of the most difficult compromises in predictive model development.”
InfluentialCè, M., Chiriac, M. D., Cozzi, A., Macrì, L., Rabaiotti, F. L., Irmici, G., Fazzini, D., Carrafiello, G., & Cellina, M. I. (2024) · Diagnostics · doi:10.3390/diagnostics14222473 · Cited by 35
“These include statistical methods such as multivariate analysis, propensity score matching, and others [97,98].”
Exploring electroencephalographic chronic pain biomarkers: a mega-analysis
Bott, F. S., Zebhauser, P. T., Hohn, V. D., Turgut, Ö., May, E. S., Tiemann, L., Ávila, C. G., Heitmann, H., Nickel, M. M., Day, M. A., et al. (2025) · EBioMedicine · doi:10.1016/j.ebiom.2025.105955 · Cited by 32
“34,37,38 In this study, ComBat was applied to mitigate site-related variability while preserving variance associated with the biologically meaningful factor of age, except when age was the dependent variable.”
Power and reproducibility in the external validation of brain-phenotype predictions
Rosenblatt, M., Tejavibulya, L., Sun, H., Camp, C., Khaitova, M., Adkinson, B. D., Jiang, R., Westwater, M. L., Noble, S., & Scheinost, D. (2024) · Nature Human Behaviour · doi:10.1038/s41562-024-01931-7 · Cited by 30
“Crucially, the covariate regression parameters and features obtained from the training set were applied to the test set to avoid data leakage 76,79,80 .”
Accurate sex prediction of cisgender and transgender individuals without brain size bias
Wiersch, L., Hamdan, S., Hoffstaedter, F., Votinov, M., Habel, U., Clemens, B., Derntl, B., Eickhoff, S. B., Patil, K. R., & Weis, S. (2023) · Scientific Reports · doi:10.1038/s41598-023-37508-z · Cited by 29
“Consequently, an accurate model performance for the ATM hold-out sample indicates a non-TIV-biased model behavior as the model classifies a person’s sex based on other features than TIV, providing a “confound-free accuracy” 31 .”
MethodInterpreting mental state decoding with deep learning models
Thomas, A. W., Ré, C., & Poldrack, R. A. (2022) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2022.07.003 · Cited by 29
Confound-leakage: confound removal in machine learning leads to leakage
Hamdan, S., Love, B. C., von Polier, G. G., Weis, S., Schwender, H. R., Eickhoff, S. B., & Patil, K. R. (2022) · GigaScience · doi:10.1093/gigascience/giad071 · Cited by 24
“Although a recent study has speculated on the pitfall of confound-leakage in ML workflows [16], it has not yet been systematically shown, analyzed nor explained.”
Novel multi-omics deconfounding variational autoencoders can obtain meaningful disease subtyping
Li, Z., Katz, S., Saccenti, E., Fardo, D. W., Claes, P. D., dos Santos, V. A. P. M., Steen, K. V., & Roshchupkin, G. V. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae512 · Cited by 23
“There are other methods working on the confounding issue for out-of-sample prediction based on, e.g. confound-isolating cross-validation [35] and feature selection in the embedding space [36].”
MethodPatterns of brain-wide associations reflect socioeconomics
Marek, S. A., Donohue, M. R., Karcher, N. R., Hoyniak, C. P., Chauvin, R. J., Meyer, A. C., Miller, J. M., Van, A. N., Wang, A., Baden, N. J., et al. (2026) · Science · doi:10.1126/science.aee6213 · Cited by 25
“Regressing confounds from both the training and test datasets can result in either inflated or deflated out- of-sample associations and is not recommended ( 61 ).”
InfluentialBrain morphology predicts individual sensitivity to pain: a multicenter machine learning approach
Kotikalapudi, R., Kincses, B., Zunhammer, M., Schlitt, F., Asan, L., Schmidt‐Wilcke, T., Kincses, Z. T., Bingel, U., & Spisák, T. (2023) · Pain · doi:10.1097/j.pain.0000000000002958 · Cited by 21
“11,17,31,82 ComBat was configured to mitigate site effects, while still preserving the biologically relevant effects of age and sex.”
Individual characteristics outperform resting-state fMRI for the prediction of behavioral phenotypes
Omidvarnia, A., Sasse, L., Larabi, D. I., Raimondo, F., Hoffstaedter, F., Kasper, J., Dukart, J., Petersen, M., Cheng, B., Thomalla, G., et al. (2024) · Communications Biology · doi:10.1038/s42003-024-06438-5 · Cited by 22
Advanced brain age in community-dwelling population with combined physical and cognitive impairments
Kuo, C.-Y., Lee, P., Peng, L., Lee, W., Wang, P., Chen, L., Chou, K., Chung, C., & Lin, C. (2023) · Neurobiology of Aging · doi:10.1016/j.neurobiolaging.2023.06.017 · Cited by 20
Ocular surface microbiome: Influences of physiological, environmental, and lifestyle factors
Rizzuto, V., Settino, M., Stroffolini, G., Covello, G., Vanags, J., Naccarato, M., Montanari, R., Rocha‐de‐Lossada, C., Mazzotta, C. G., Forestiero, A., et al. (2025) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2025.110046 · Cited by 17
Benchmarking explanation methods for mental state decoding with deep learning models
Thomas, A. W., Ré, C., & Poldrack, R. A. (2023) · NeuroImage · doi:10.1016/j.neuroimage.2023.120109 · Cited by 18
“Similar examples are imaginable in functional neuroimaging, as recently suggested by Chyzhyk et al. (2022) who hypothesize that biomarker models for specific disease conditions could learn to distinguish patients from controls by their generally increased head motion.”
BackgroundPavlovich, K., Pang, J. C. S., Holmes, A., Constable, T., & Fornito, A. (2025) · Imaging Neuroscience · doi:10.1162/imag.a.97 · Cited by 16
“The regression coefficients were calculated from the training set and then applied to the test set to avoid leakage between train and test sets in the prediction model (see below) ( Chyzhyk et al., 2022).”
Jahanshad, N., Lenzini, P., & Bijsterbosch, J. (2025) · Neuropsychopharmacology · doi:10.1038/s41386-024-01938-8 · Cited by 12
“…siblings/twins cannot be treated as independent subjects) [57] or confound processing or other group-level steps performed prior to machine learning (e.g., regressions performed across all subjects expose and potentially bias training data by using information from the full dataset) [58, 59].”
BackgroundOverview of Challenges in Brain-Based Predictive Modeling: Toward Meaningful Predictive Insights
Komeyer, V., Nieto, N., Eickhoff, S. B., Raimondo, F., & Patil, K. R. (2025) · Biological Psychiatry · doi:10.1016/j.biopsych.2025.09.003 · Cited by 12
Galindo‐Prieto, B., & Mudway, I. (2025) · Analytica Chimica Acta · doi:10.1016/j.aca.2025.344826 · Cited by 14
Murchan, P., Broin, P. Ó., Baird, A., Sheils, O. M., & Finn, S. P. (2024) · Journal of Pathology Informatics · doi:10.1016/j.jpi.2024.100396 · Cited by 12
Modeling Brain Representations of Words' Concreteness in Context Using GPT‐2 and Human Ratings
Bruera, A., Tao, Y., Anderson, A. J., Çokal, D., Haber, J., & Poesio, M. (2023) · Cognitive Science · doi:10.1111/cogs.13388 · Cited by 13
An externally validated resting-state brain connectivity signature of pain-related learning
Kincses, B., Forkmann, K., Schlitt, F., Pawlik, R. J., Schmidt, K., Timmann, D., Elsenbruch, S., Wiech, K., Bingel, U., & Spisák, T. (2024) · Communications Biology · doi:10.1038/s42003-024-06574-y · Cited by 11
Lee, S., Niu, R., Zhu, L., Kayser, A. S., & Hsu, M. (2024) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2412881121 · Cited by 10
Xu, L., Ma, H., Guan, Y., Liu, J., Huang, H., Zhang, Y., & Tian, L. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3304974 · Cited by 9
Zabihi, M., Zar, A. C., Grover, P., & Rosenthal, E. S. (2023) · Computing in cardiology · doi:10.22489/cinc.2023.142 · Cited by 9
“…from the initial training data formed by withholding a random 20% subset to produce 5 intersecting training subsets; 2) Confounder-isolating [18] CV (cluster analysis) utilizing k-means clustering to group training data based on potential confounding variables, namely hospital type and sex,…”
MethodLinking the exposome to the brain–behaviour phenotype
Genon, S., Ibáñez, A. M., Tahmasian, M., & Eickhoff, S. B. (2026) · Nature reviews. Neuroscience · doi:10.1038/s41583-026-01049-x · Cited by 8
“…multiple exposomic variables that requires different preprocessing steps (that is, imputation, standardization, deconfounding and so on) because great care has to be taken to implement preprocessing within the cross-validation scheme to avoid data leakage that would invalidate the results 83,84 .”
Brasanac, J., & Chien, C. (2023) · Frontiers in Human Neuroscience · doi:10.3389/fnhum.2023.1151531 · Cited by 10
“, 2018), sex (Voskuhl et al., 2020; Chyzhyk et al., 2022), and often brain lesion load (Sinnecker et al.”
BackgroundBracher‐Smith, M., & Escott‐Price, V. (2025) · Molecular Psychiatry · doi:10.1038/s41380-025-03330-4 · Cited by 6
Age, Motion, Medical, and Psychiatric Associations With Incidental Findings in Brain MRI
Tobe, R. H., Tu, L., Roberts, M., Kiar, G., Breland, M. M., Tian, Y., Kang, M., Ross, R., Ryan, M. M., Valenza, E., et al. (2024) · JAMA Network Open · doi:10.1001/jamanetworkopen.2023.55901 · Cited by 6
Self-citationZabihi, M., Rubin, D. B., Ack, S. E., Gilmore, E. J., Junior, V. F. M., Zafar, S. F., Li, Q., Young, M. J., Edlow, B. L., Bodien, Y. G., et al. (2022) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.1101/2022.09.30.510334 · Cited by 6
Bruera, A., & Poesio, M. (2024) · PLoS ONE · doi:10.1371/journal.pone.0291099 · Cited by 5
Hearne, L. J., Yeo, B. T. T., Webb, L., Zalesky, A., Fitzgerald, P. B., Murphy, O. W., Tian, Y., Breakspear, M. J., Hall, C. V., Choi, S., et al. (2025) · Molecular Psychiatry · doi:10.1038/s41380-025-03416-z · Cited by 5
Babin, É., Vigneau, É., Antignac, J., Bizec, B. L., & Cano-Sancho, G. (2024) · Chemosphere · doi:10.1016/j.chemosphere.2024.142465 · Cited by 4
Komeyer, V., Eickhoff, S. B., Grefkes, C., Patil, K. R., & Raimondo, F. (2024) · medRxiv · doi:10.1101/2024.02.02.24302198 · Cited by 4
“Overall, generic conventional and non-contextual confound removal (too many or too little) can result in suboptimal models 19,22 .”
BackgroundMethodLiang, C., Silva, R. F., Adali, T., Jiang, R., Zhang, D., Qi, S., & Calhoun, V. D. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3597045 · Cited by 3
“Another method, confound-isolating cross-validation, involves subsampling the entire dataset to obtain an unbiased subpopulation (without the confounding e ff ects), followed by fusion analysis within this subpopulation to control for confounding e ff ects [13].”
Chen, S., Guo, X., Liu, X., Liu, S., & Ming, D. (2025) · Journal of Affective Disorders · doi:10.1016/j.jad.2025.120189 · Cited by 3
Martin, D., Houedry, P., Derbré, F., & Monbet, V. (2025) · Methods in Ecology and Evolution · doi:10.1111/2041-210x.70223 · Cited by 3
“The approach implemented here is closely related to out- of-sample deconfounding, providing empirical evidence that enterotype-related information may not be essential for predicting host health and acts as a confounding factor (Chyzhyk et al., 2022).”
Dai, P., He, Z., Luo, J., Huang, K., Hu, T., Chen, Q., Liao, S., & Yi, X. (2025) · Journal of Neuroscience Methods · doi:10.1016/j.jneumeth.2025.110406 · Cited by 3
Narasimhan, V., Warrier, S. P., John, J. J., Monisha Priya, T., Varadaraj, N., Thomas, G. G., & Veeraraghavan, B. (2026) · BMC Gastroenterology · doi:10.1186/s12876-025-04595-3 · Cited by 2
“While statistical literature is widely available for controlling confounding factors through regressing-out analyses, it is not applied to predictive ensemble models due to the non-linear assumptions that they rely on [40].”
The role of age in the relationship between brain structure and cognition: moderator or confound?
Griffin, B., Gohil, C., Woolrich, M. W., Smith, S. M., & Vidaurre, D. (2026) · Cerebral Cortex · doi:10.1093/cercor/bhag024 · Cited by 2
“By contrast, pooled training provides more consistent generalization across age groups, and may be preferable for heterogeneous cohorts or when the target age range is uncertain.”
Hu, J., Hou, Y., Peng, B., Liao, B., Xu, Z., Hou, G., & Dong, S. (2025) · Journal of Psychiatric Research · doi:10.1016/j.jpsychires.2025.01.001 · Cited by 1
Pavlovich, K., Pang, J. C. S., & Fornito, A. (2025) · Network Neuroscience · doi:10.1162/netn.a.534 · Cited by 1
“The regression coefficients were calculated from the training set and then applied to the test set to avoid leakage between train and test sets in the prediction model (see below) (Chyzhyk et al., 2022).”
Structural Causal Modeling with Double Machine Learning for Robust Policy Evaluation
Wismarini, T. D., Prasetyo, S. Y. J., Susanto, B., & Nataliani, Y. (2025) · IEEE Information Technology International Seminar (ITIS) · doi:10.1109/itis67966.2025.11308780 · Cited by 1
“Specifically, DML proceeds in three steps: First, Predicting the outcome ( conflict_presence ) from confounders (W) [31][32].”
Subject fingerprinting and task classification rely on distinct functional connectivity features
Aggarwal, H., Ponce, A. F., Shankar, S., Mzayek, Y., Pérez‐Millan, A., Pinho, A. L., Thual, A., & Thirion, B. (2026) · Brain Structure and Function · doi:10.1007/s00429-025-03068-3 · Cited by 1
“Effect of motion on classification performance To test whether the variable motion artifacts across subjects and runs contribute to the classification performance, we used frame-wise displacement (FD) as the feature for classification analyses (similar to Chyzhyk et al. ( 2022)).”
Self-citationLiu, Y., Li, W., Wang, Y., Chen, N., & Chen, M. (2026) · Big Data and Cognitive Computing · doi:10.3390/bdcc10020056 · Cited by 1
“Combining previous research [13,14], we argue that the hallucination issue in RAG tasks combined with long-chain reasoning models may stem from the influence of external knowledge noise and reasoning process noise.”
Zhao, Q., & Pohl, K. M. (2026) · Alcohol research · doi:10.35946/arcr.v46.1.03 · Cited by 1
“61,207,208 Also , evaluation of the models could be enhanced by going beyond reporting a single AUC and incorporating fairness metrics 209 (e.g., subgroup AUCs) and sensitivity analyses 210 (e.g., performance changes with and without confounders).”
Eliminating confounder-induced bias in the statistics of interventionPreprint
Loukas, O., & Chung, H. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2312.00225 · Cited by 1
“Operating on the combined contingency table (point 1 in the list) enables in a dataset sampled from q a thorough investigation of higher-order associations such as the predictivity [28] of profiles x and s on the outcome y .”
BackgroundDafflon, J., Moraczewski, D., Earl, E., Nielson, D. M., Loewinger, G. C., McClure, P., Thomas, A. G., & Pereira, F. L. F. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2405.00255 · Cited by 1
“- Hypothesis: In accordance with the literature (Chyzhyk et al., 2022) and the results we obtained on our exploratory analysis (Figure 4), failure to remove age and sex confounds will lead to a higher correlation coefficient for variables that are strongly correlated with age/sex.”
BackgroundResultInfluentialA causally informed framework for robust confounder control in biomedical machine learningPreprint
Komeyer, V., Herrmann, C., Eickhoff, S. B., Rathkopf, C. A., Raimondo, F., & Patil, K. R. (2024) · medRxiv · doi:10.1101/2024.09.20.24314055 · Cited by 1
“Reporting statistical associations appears appropriate when following the ubiquitous (but faulty – see Box 3 ) definition of confounders as any variable that correlates with the feature (predictor) and the target (outcome), but which’s variance is of no interest 34,35 .”
BackgroundThe State of Motion: Demographic and Representational Consequences of Motion-Based Exclusion in FMRIPreprint
Sadil, P. S., Ansari, B., Casamento‐Moran, A., Choe, A. S., Choi, J., Farahani, F. V., Ismaila, L., Johnson, M. A., Kannan, A., Nebel, M. B., et al. (2026) · bioRxiv (Cold Spring Harbor Laboratory) · doi:10.64898/2026.07.23.739855 · Cited by 2
Ack, S. E., Dolmans, R. G. F., Foreman, B. P., Manley, G. T., Rosenthal, E. S., & Zabihi, M. (2024) · Critical Care Explorations · doi:10.1097/cce.0000000000001118
“Unlike conventional cross-validation, which randomly partitions data into training and testing sets, confound-isolating cross-validation uses various techniques to stratify the data into a series of strata or folds (35, 38).”
Xie, J. (2026) · Social Sciences & Humanities Open · doi:10.1016/j.ssaho.2026.102775
Neuroimaging genetics and developmental psychopathology: A systematic review
Cheek, C. L., Dobrynin, P. V., Khafizova, G. V., Koirala, N., Mahaffy, K. E., & Grigorenko, E. L. (2025) · Development and Psychopathology · doi:10.1017/s0954579425100400 · Cited by 1
Multi-task learning for joint diagnosis of CNVs and psychiatric conditions from rs-fMRI
Harvey, A. (2023) · Papyrus : Institutional Repository (Université de Montréal) · doi:10.71781/10494
Herclassificatie van individuen via multi-view data in het licht van confounders
Li, Z. (2025) · Lirias
Tatosh, S., Brito, S. A. D., Chester, S. C., Ogawa, T., Terao, M., Nakai, R., Abe, N., & Dugre, J. R. (2026) · medRxiv · doi:10.64898/2026.01.12.26343689
Eggshell Crack Classification Using a Hybrid Texture-Color Descriptor with Machine Learning Methods
Setiawan, A., Syofian, S., & Budiman, A. A. (2026) · Engineering Technology & Applied Science Research · doi:10.48084/etasr.15450
“Cross-validation is essential to generate an impartial model and facilitate the calculation of prediction errors and accuracies that more closely represent the true value [24].”
Sabac, D. S. A. (2024) · TSpace (University of Toronto)
Controlling for Confounding Bias in Linear Prediction Models for Binary Classification
Avalos, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-35387-0_29
Yan, Y., Chan, Y. H., Cheng, J., Chew, Q. H., Sim, K., & Rajapakse, J. C. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-38091-3_36
Sadoun, M. Y., Sharif, S., Liu, Y., Siddique, Z., & Banad, Y. M. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2610.06973
“The procedure adapts established confound-audit practice [10, 71, 73, 76] and is released as runnable scripts with committed reference outputs.”
Ruiz, J. G., Díaz, H., & Ariza‐Ruiz, D. (2026) · Algorithms · doi:10.3390/a19110907 · Cited by 1
Metadata
- Authors
- Chyzhyk, D., Varoquaux, G., Milham, M., & Thirion, B.
- Affiliations
- Institut national de recherche en sciences et technologies du numérique (FR)
- Parietal project-team , INRIA Saclay-île de France,
- Nathan Kline Institute for Psychiatric Research (US)
- Corresponding author
- Chyzhyk, Darya
- Venue
- GigaScience
- Publisher
- Oxford University Press (OUP)
- Published
- 2022
- Volume
- 11
- ISSN
- 2047-217X (ISSN-L)
- PMC
- PMC8917515
- OpenAlex
- W4220876827
- Semantic Scholar
- c705dc0b28422c58e3d1a615da3427b760cfc49d
- Cited by
- 65
- References
- 68
- Citations by year
- 2026: 12 · 2025: 16 · 2024: 21 · 2023: 10 · 2022: 6
- FWCI
- 9.82
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- GigaScience (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- HAL (submittedVersion, other-oa)
- Funders
- Child Mind Institute
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Epidemiology
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Keywords
- confounding, deconfounding, neuroimaging biomarkers, predictive modeling, head movement, cross-validation, prediction accuracy, machine learning
- MeSH terms
- Machine Learning, Brain, Humans, Biomarkers
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.