Skip to content
DraconiticResearch
Search Sources API
194,230 studies31 sourcesUpdated 56 min ago
Peer-reviewedCC BYEuropean Journal of Neuroscience

Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling

  • Catriona L. Scrivener1,2
  • Tijl Grootswagers3
  • Alexandra Woolgar2
  1. 1Department of Psychology, School of Philosophy, Psychology and Language Sciences, University of Edinburgh, Edinburgh, UK
  2. 2MRC Cognition and Brain Sciences Unit, University of Cambridge, Cambridge, UK
  3. 3The MARCS Institute for Brain, Behaviour and Development, Western Sydney University, NSW, Australia
Open at sourceDOICite

Abstract

Multivariate pattern analysis (MVPA) is a popular technique that can distinguish between condition-specific patterns of activation. Applied to neuroimaging data, MVPA decoding for inference uses above chance decoding to identify statistically robust condition-specific information in neuroimaging data, which may be missed by univariate methods. However, several analysis choices influence decoding results, and the combined effects of these choices have not been fully evaluated. In particular, an increasingly popular approach is to average data from several trials together before training an MVPA classifier, but the decision about how much averaging to do is arbitrary and the effect of varying this parameter has not been documented. Here, we systematically assessed the influence of trial averaging and resampling on decoding accuracy and subsequent statistical outcome on data simulated using two different toolboxes (CoSMoMVPA and SEREEGA). Although the optimal parameters varied with the classifier and cross-validation approach used, we found that modest trial averaging using up to 5%-10% of the total number of trials per condition improved decoding accuracy and associated t-statistics. In addition, a small amount of resampling could improve t-statistics and classification performance, but was not always necessary. We provide code to allow researchers to optimise these analysis choices for the parameters of their data.

Related

  1. Evaluating Clustering and Imputation Approaches for Multivariate Longitudinal Clinical Data Using a Real-World eCDF-Copula Resampling Framework
    H. Cai, K. Xu, S. Osmundson +5
    bioRxivPreprintCC BY-NC-NDFull textBioinformatics & computational biologybioRxiv
  2. ssGSEA3.0: Feature-Level Bootstrap Resampling for Robust Assessment of Statistical Significance
    N. M. Clark, D. R. Mani
    bioRxivPreprintCC BYFull textBioinformatics & computational biologybioRxiv
  3. Optimising digital volume correlation across materials: a practical framework for accuracy and spatial resolution
    A. L. Parmenter, A. Sharma, B. K. Bay +1
    bioRxivPreprintCC BY-NC-NDFull textSystems & synthetic biologybioRxiv
  4. Enhancing dance learning by optimising instructional perspective: A preregistered study on the effect of dancers’ viewing perspective, cognitive load and expertise
    Elisabetta Versace, Manuela Angioi, Kristin Hadfield +9
    PLOS OnePeer-reviewedCC BY 4.0NeurosciencePLOS
  5. Optimising bat-friendly curtailment algorithms for wind turbines
    Kévin Barré, Kseniia Kravchenko, Thierry Chambert +8
    Journal of Environmental ManagementPeer-reviewedCC BY-NC-SA 4.0Full textEcologyEnvironmental science & pollutionHAL

Cite

BibTeXRIS
BibTeX
@article{scrivener2026optimising,
  title = {{Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling}},
  author = {Catriona L. Scrivener and Tijl Grootswagers and Alexandra Woolgar},
  journal = {European Journal of Neuroscience},
  year = {2026},
  month = jul,
  doi = {10.1111/ejn.70601},
  url = {https://doi.org/10.1111/ejn.70601},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling
AU  - Catriona L. Scrivener
AU  - Tijl Grootswagers
AU  - Alexandra Woolgar
PY  - 2026
DA  - 2026/07/01
JO  - European Journal of Neuroscience
DO  - 10.1111/ejn.70601
UR  - https://doi.org/10.1111/ejn.70601
AB  - Multivariate pattern analysis (MVPA) is a popular technique that can distinguish between condition-specific patterns of activation. Applied to neuroimaging data, MVPA decoding for inference uses above chance decoding to identify statistically robust condition-specific information in neuroimaging data, which may be missed by univariate methods. However, several analysis choices influence decoding results, and the combined effects of these choices have not been fully evaluated. In particular, an increasingly popular approach is to average data from several trials together before training an MVPA classifier, but the decision about how much averaging to do is arbitrary and the effect of varying this parameter has not been documented. Here, we systematically assessed the influence of trial averaging and resampling on decoding accuracy and subsequent statistical outcome on data simulated using two different toolboxes (CoSMoMVPA and SEREEGA). Although the optimal parameters varied with the classifier and cross-validation approach used, we found that modest trial averaging using up to 5%-10% of the total number of trials per condition improved decoding accuracy and associated t-statistics. In addition, a small amount of resampling could improve t-statistics and classification performance, but was not always necessary. We provide code to allow researchers to optimise these analysis choices for the parameters of their data.
SN  - 0953-816X
SN  - 1460-9568
ER  -