OSCR

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

Code ↔ Paper

5 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 5 matches
  1. [1] § Methods › Simulation 2 ↔ analysis scripts/SEREEG_simulations_P300.m, lines 94–137 · score 0.88 · peak latency, peak width, location, midbrain, mixed, scalp
  2. [2] § Methods › Simulation 1 ↔ analysis scripts/permutation_testing.m, lines 45–103 · score 0.63 · cosmo montecarlo, sign flip permutation, alpha, score, simulations, resampled
  3. [3] § Methods › Simulation 1 ↔ analysis scripts/SEREEG_simulations_P300.m, lines 2–49 · score 0.59 · libSVM, cosmoMVPA, multivariate, LDA, Bayes, simulate
  4. [4] § Methods › Simulation 2 ↔ analysis scripts/permutation_testing.m, lines 45–103 · score 0.59 · cosmo montecarlo, sign flip permutation, alpha, score, simulation
  5. [5] § Results › Simulation 2: SEREEGA › Influence of Simulation Type ↔ analysis scripts/SEREEG_simulations_P300.m, lines 2–49 · score 0.54 · SEREEGA toolbox, P300, 13 %, ERP, amplitude, SVM

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

MATLAB · 204 lines · 8.4 KB · no license · 3 matches

  2. % Optimising analysis choices for multivariate decoding: creating
  3. % pseudotrials using trial averaging and resampling.
  4. % Catriona Scrivener,Tijl Grootswagers, Alexandra Woolgar
  5. % 2025
  7. % Simulating EEG data using SEREEGA toolbox and EEGLab
  8. % Generate a P300-like ERP responses from a deep central source
  10. % SEREEGA toolbox paper:
  11. % Krol, L. R., Pawlitzki, J., Lotte, F., Gramann, K., & Zander, T. O. (2018).
  12. % SEREEGA: Simulating event-related EEG activity. Journal of neuroscience methods, 309, 13-24.
  14. % add path to SEREEGA folder and subfolders
  15. addpath(genpath('.../toolboxes/SEREEGA-master'))
  17. % add path to EEGlab
  18. addpath('...eeglab2021.1')
  20. % set path to cosmoMVPA
  21. addpath(genpath('.../toolboxes/CoSMoMVPA-master'))
  22. % libSVM
  23. addpath(genpath('.../toolboxes/libsvm-3.36'))
  25. % define results directory
  26. results_path = pwd;
  28. % Check which classifier to use
  29. if strcmp(class_type, 'svm')
  30. my_classifier = @cosmo_classify_libsvm;
  31. elseif strcmp(class_type, 'lda')
  32. my_classifier = @cosmo_classify_lda;
  33. elseif strcmp(class_type, 'naive_bayes')
  34. my_classifier = @cosmo_classify_naive_bayes;
  35. end
  37. % Define simulation parameters
  39. amp1 = 11; % P300 amplitude condition 1
  40. amp2 = 10.5; % P300 amplitude condition 2
  41. snr = 1/3; % signal to noise ratio
  42. nreps = 90; % here used as the number of trials per condition
  43. max_count = 30; % count = how many trials per pseudotrial
  44. max_resamp = 15; % resampling = how many times each trial can be used
  45. max_ptps = 30; % how many 'participants' to simulate
  46. nseeds = 50; % number of random trial allocations per ptp+param set
  47. nchunks = 3; % how many chunks
  48. nperm = 100; % how many permutation
  49. class_type = {'svm'}; % which classifier
  51. % create results variable
  52. x = NaN(max_count,max_resamp,nseeds,max_ptps);
  53. perm = NaN(nperm,max_count,max_resamp,nseeds,max_ptps);
  54. pvals = NaN(max_count,max_resamp,nseeds,max_ptps);
  55. copy_max_resamp = max_resamp;
  57. % Simulate a predefined 64-electrode montage using the ICBM-NY pre-generated lead field
  58. % with 64 channels from the 10-20 system
  59. % Assumes you have the NY Head leadfield in MATLAB format in the
  60. % path. This is available at http://www.parralab.org/nyhead/sa_nyhead.mat
  61. % In the ICBM-NY lead field, default orientations are included,
  62. % which orient the dipole perpendicular to the cortical surface.
  63. lf = lf_generate_fromnyhead('montage', 'S64');
  65. % iterate over different ptps/seeds
  66. for ptp = 1:max_ptps
  68. % generate epoch eeg data
  69. epochs = struct();
  70. epochs.n = nreps; % the number of epochs to simulate per condition
  71. epochs.srate = 250; % their sampling rate in Hz
  72. epochs.length = 600; % length of the epochs (with no baseline)
  74. % create white noise signals
  75. noise_white = struct( ...
  76. 'type', 'noise', ...
  77. 'color', 'white', ...
  78. 'amplitude', 5);
  79. noise_white = utl_check_class(noise_white);
  81. % generate 25 random noise sources
  82. % only use white noise
  83. sources = lf_get_source_spaced(lf, 10, 25);
  84. comps = utl_create_component(sources, noise_white, lf);
  85. [comps1, comps2] = deal(comps);
  87. % create separate noise dataset that we will
  88. % later add to the signal in varying amounts
  89. % to adjust the snr
  90. noiseact = struct('type', 'noise', 'color', 'white', 'amplitude', 1);
  91. noisecomps = utl_create_component(sources, noiseact, lf);
  92. noisedata = generate_scalpdata(noisecomps, lf, epochs,'showProgress',0,'useParallelPool',0);
  94. % generate a P300 ERP
  95. p3 = [];
  96. p3.signal{1} = struct();
  97. p3.signal{1}.peakLatency = 350; % in ms, starting at the start of the epoch
  98. p3.signal{1}.peakWidth = 350; % in ms
  99. p3.signal{1}.peakAmplitude = 3; % in microvolt - we will update
  100. p3.signal{1} = utl_check_class(p3.signal{1}, 'type', 'erp');
  102. % pick a source in the midbrain
  103. p3.source = lf_get_source_nearest(lf, [0 -40 -25]);
  104. p3.orientation = utl_get_orientation_pseudotangential(p3.source,lf);
  105. p3.orientationDv = [0 0 0];
  106. p3.orientation = [0.03,-0.16,1];
  108. % add a broad, weak P300, with condition difference
  109. p3.signal{1}.peakAmplitude = amp1;
  110. comps1(end+1) = p3;
  111. p3.signal{1}.peakAmplitude = amp2;
  112. comps2(end+1) = p3;
  114. % project ERPs to the scalp for each condition
  115. data1 = generate_scalpdata(comps1, lf, epochs,'useParallelPool',0,'showProgress',0);
  116. data2 = generate_scalpdata(comps2, lf, epochs,'useParallelPool',0,'showProgress',0);
  118. % create eeglab dataset with added noise
  119. % first mix the ERP data with noise with an SNR of 1/3, or -6 dB
  120. data1_snr = utl_mix_data(data1,noisedata,snr);
  121. data2_snr = utl_mix_data(data2,noisedata,snr);
  122. EEG1_snr = utl_create_eeglabdataset(data1_snr, epochs, lf, 'marker', 'event1');
  123. EEG2_snr = utl_create_eeglabdataset(data2_snr, epochs, lf, 'marker', 'event2');
  125. %__________________________________________________________________________
  126. % run decoding
  127. % cosmo wants the data as trial*channel*time
  128. % reformat and combine data
  129. allData = [permute(EEG1_snr.data,[3,1,2]);permute(EEG2_snr.data,[3,1,2])];
  130. % take channel locations and timevector from one dataset
  131. chanlocs = EEG1_snr.chanlocs; times = EEG1_snr.times;
  133. % flatten the data over chan and time
  134. % the data is organised such that it does time 1 over all channels, then
  135. % time 2 over all channels etc.
  136. ds = cosmo_flatten(allData,{'chan','time'},...
  137. {chanlocs,times});
  139. % add targets to dataset
  140. ds.sa.targets = [ones([size(EEG1_snr.data,3),1]);...
  141. repmat(2,[size(EEG1_snr.data,3),1])];
  143. % first set chunks so that all samples are assumed to be independent
  144. ds.sa.chunks = [(1:size(EEG1_snr.data,3))';(1:size(EEG2_snr.data,3))'];
  146. % re-assign chunks pseudo-randomly into blocks
  147. ds.sa.chunks = cosmo_chunkize(ds,nchunks);
  149. % iterate over count values - number of samples to select for each average
  150. for count = 1:max_count
  151. disp(['c',num2str(count)])
  153. % count 1 resamp 1 is always the same
  154. if count == 1; max_resamp =1; else; max_resamp = copy_max_resamp;end
  156. % iterate over resampling values - maximum number of times each sample in ds is used for averaging
  157. for resampling = 1:max_resamp
  158. disp(['r',num2str(resampling)])
  160. % use seed for pseudo-random sampling
  161. for seed = 1:nseeds
  163. % averages samples using count and resampling values
  164. % use a different seed each time
  165. ds_avg_rs = cosmo_average_samples(ds,'count',count,'resamplings',resampling,'seed',seed);
  167. % run classifier and save accuracy
  168. ma={};
  169. ma.classifier = my_classifier;
  170. ma.max_feature_count = 10000;
  171. ma.partitions = cosmo_nfold_partitioner(ds_avg_rs);
  172. res_rs = cosmo_crossvalidation_measure(ds_avg_rs,ma);
  173. x(count,resampling,seed,ptp) = res_rs.samples;
  175. % to use permutation testing
  176. if seed == 1
  177. acc0 = zeros(nperm,1);
  178. ds0 = ds_avg_rs; ma0 = ma; % copy of datasets
  179. for k = 1:nperm
  180. ds0.sa.targets = cosmo_randomize_targets(ds0);
  181. [~, acc0(k)] = cosmo_crossvalidate(ds0,ma0.classifier,ma0.partitions,ma0);
  182. end
  184. p = sum(res_rs.samples<acc0)/nperm;
  185. perm(:,count,resampling,seed,ptp) = acc0;
  186. pvals(count,resampling,seed,ptp) = p;
  187. end
  189. end % seed
  190. end % resampling
  191. end % count
  192. end % ptp
  194. % X = count x resampling x seeds x ptp (ptps)
  195. % mean over resampling seeds
  196. Y_res(:,:,1) = mean(mean(x,3),4);
  197. Y_res(:,:,2) = std(mean(x,3),0,4);
  199. save(fullfile(results_path,['Y_res_nreps' num2str(nreps) '_count' num2str(max_count) '_resamp' num2str(max_resamp) '_ptps' num2str(max_ptps) '_class' class_type '.mat']),'Y_res')
  200. save(fullfile(results_path,['x_nreps' num2str(nreps) '_count' num2str(max_count) '_resamp' num2str(max_resamp) '_ptps' num2str(max_ptps) '_class' class_type '.mat']),'x')
  201. save(fullfile(results_path,['perm_nreps' num2str(nreps) '_count' num2str(max_count) '_resamp' num2str(max_resamp) '_ptps' num2str(max_ptps) '_class' class_type '.mat']),'perm')
  202. save(fullfile(results_path,['pvals_nreps' num2str(nreps) '_count' num2str(max_count) '_resamp' num2str(max_resamp) '_ptps' num2str(max_ptps) '_class' class_type '.mat']),'pvals')

SEREEG_simulations_P300.m, no license · at the source

Overview

Authors: Catriona L. Scrivener1,2, Tijl Grootswagers3,4, Alexandra Woolgar1,5
ORCID iDs: Catriona L. Scrivener, Tijl Grootswagers, Alexandra Woolgar
  1. MRC Cognition and Brain Sciences Unit University of Cambridge Cambridge UK
  2. School of Philosophy, Psychology and Language Sciences University of Edinburgh Edinburgh UK
  3. The MARCS Institute for Brain, Behaviour and Development Western Sydney University Sydney New South Wales Australia
  4. School of Computer, Data and Mathematical Sciences Western Sydney University Sydney New South Wales Australia
  5. Department of Psychology University of Cambridge Cambridge UK
Institutions: MRC Cognition and Brain Sciences Unit (United Kingdom); University of Cambridge (United Kingdom); University of Edinburgh (United Kingdom); Western Sydney University (Australia)
Journal: The European journal of neuroscience, volume 64, issue 2, article e70601
Dates: received 18 March 2025; accepted 13 June 2026; published online 19 July 2026; in print July 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1111/ejn.70601 · PMID 42473262 · PMCID PMC13382189 · OpenAlex W4387460001
Open access: hybrid, a free copy (OpenAlex)
Preprint: doi:10.1101/2023.10.04.560678
Status: code verified
Categories: human (organism), clinical / translational (subfield)
Methods: Connectivity, Statistics, Machine learning, Evoked potentials
Keywords: decoding, multivariate pattern analysis, pseudotrials
MeSH: Brain*, Image Processing, Computer-Assisted*, Classification Algorithms, Humans, Magnetic Resonance Imaging, Multivariate Analysis (* major topic)
Topic: Neural Networks and Applications (Artificial Intelligence, Computer Science), according to OpenAlex
Funding: Australian Research Council (DE230100380, DP170101840); RCUK | Biotechnology and Biological Sciences Research Council (BBSRC) (BB/V003887/1); Medical Research Council (MC_UU_00030/15)
Citations: cited by 1 paper (Europe PMC); 21 references in the paper

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.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above, with 5 matches between paragraphs and lines of code.

OSF hjf75

License: none: the authors keep all their rights
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Languages: MATLAB (4)
Size: 9 files, 4 scripts
Software Heritage: not checked
Found in: “Data Availability Statement”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: CoSMoMVPA (3 files), Statistics and Machine Learning Toolbox (1 file)
Availability: 1 check, the latest on 27 September 2026: the link answers (HTTP 200)
  • 27 September 2026: the link answers (HTTP 200)
4 files
At the source: osf.io/hjf75/

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 4 scripts, each with its path and the digest of its content;
  • 5 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data Availability Statement

Analysis scripts can be found at https://osf.io/hjf75/.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 2, 28 September 2026

  • Publisher: n/a → Wiley
  • Authors: added Alexandra Woolgar (0000-0002-8453-7424); removed Alexandra Woolgar

Version 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 3 authors, 3 keywords, 6 MeSH terms, 3 funders, 20 references.

Cite

This paper

Scrivener, C. L., Grootswagers, T., & Woolgar, A. (2026). Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling. The European journal of neuroscience, 64(2), e70601. https://doi.org/10.1111/ejn.70601

BibTeX

@article{scrivener2026optimising,
author = {Scrivener, Catriona L. and Grootswagers, Tijl and Woolgar, Alexandra},
title = {{Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling}},
journal = {The European journal of neuroscience},
year = {2026},
month = jul,
volume = {64},
number = {2},
pages = {e70601},
publisher = {Wiley},
issn = {0953-816X},
doi = {10.1111/ejn.70601},
url = {https://doi.org/10.1111/ejn.70601},
pmid = {42473262},
pmcid = {PMC13382189}
}

RIS

TY - JOUR
AU - Scrivener, Catriona L.
AU - Grootswagers, Tijl
AU - Woolgar, Alexandra
TI - Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling
T2 - The European journal of neuroscience
J2 - Eur J Neurosci
PY - 2026
DA - 2026/07/01
VL - 64
IS - 2
SP - e70601
SN - 0953-816X
PB - Wiley
DO - 10.1111/ejn.70601
UR - https://doi.org/10.1111/ejn.70601
LA - en
ER -

CSL-JSON

{
"id": "10.1111/ejn.70601",
"type": "article-journal",
"title": "Optimising Analysis Choices for Multivariate Decoding: Creating Pseudotrials Using Trial Averaging and Resampling",
"container-title": "The European journal of neuroscience",
"author": [
{
"family": "Scrivener",
"given": "Catriona L."
},
{
"family": "Grootswagers",
"given": "Tijl"
},
{
"family": "Woolgar",
"given": "Alexandra"
}
],
"container-title-short": "Eur J Neurosci",
"volume": "64",
"issue": "2",
"page": "e70601",
"DOI": "10.1111/ejn.70601",
"PMID": "42473262",
"PMCID": "PMC13382189",
"ISSN": "0953-816X",
"publisher": "Wiley",
"URL": "https://doi.org/10.1111/ejn.70601",
"language": "en",
"issued": {
"date-parts": [
[
2026,
7,
1
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1167/jov.26.8.2 [code]
Disentangling objects' contextual associations from perceptual and conceptual attributes using time-resolved neural decoding.
Journal: Journal of vision
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 3 references
[2] doi:10.1162/imag.a.1332 [code]
Decision processes underlying effort avoidance and their relationship with metacognition.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 3 references
[3] doi:10.1371/journal.pone.0347992 [code]
Rotation-tolerant representations elucidate the time-course of high-level object processing.
Journal: PloS one
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 2 references
[4] doi:10.1167/jov.26.8.4 [code]
The neural processes of illusory occlusion in object recognition.
Journal: Journal of vision
In common: 1 reference, author Tijl Grootswagers
[5] doi:10.1038/s41467-026-73153-6 [code]
Latent neural architecture organising shared aesthetic evaluations of visual artworks.
Journal: Nature communications
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 1 reference
[6] doi:10.1007/s00429-026-03111-x [code]
Direction selectivity in naturalistic action observation: distributed representations across the action observation network.
Journal: Brain structure & function
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 1 reference
[7] doi:10.1038/s41598-026-46149-x [code]
Dynamic neural representations of scene beauty are relatively unaffected by stimulus timing and task.
Journal: Scientific reports
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 1 reference
[8] doi:10.1162/imag.a.1203 [code]
Motor cortical areas facilitate schema-mediated integration of new motor information into memory.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, 1 reference
[9] doi:10.1371/journal.pbio.3003779 [code]
Past and present goals are represented concurrently during visual search.
Journal: PLoS biology
In common: 3 references
[10] doi:10.1016/j.bbih.2026.101299 [code]
Multimodal approach to identify neuropsychophysiological subgroups in myalgic encephalomyelitis/chronic fatigue syndrome and their relevance for rehabilitation: protocol for a mechanistic cross-sectional and longitudinal study.
Journal: Brain, behavior, & immunity - health
In common: CoSMoMVPA, Statistics and Machine Learning Toolbox, clinical / translational

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.