OSCR

Cortical knowledge structures guide word concept learning.

Code ↔ Paper

The paper beside its authors' code: matches between them have not been computed for this paper yet.

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

R · 228 lines · 7.7 KB · no license

  1. #### get p(h) ####
  2. # hc = hclust(dist(cor(NeuralPattern)))
  3. Bayes_getnode = function(hc,leafnodes = T){
  4. nodes_list = list()
  6. # normalize height
  7. hc$height = hc$height/(2*max(hc$height))
  10. # get leaf nodes
  11. leaf_node = list()
  12. for(i in 1:length(hc$labels)){
  13. temp = list(labs = c(hc$labels[i]),
  14. height = 0,
  15. parent_node = c(which(hc$merge[,1] == -i),which(hc$merge[,2] == -i)),
  16. domain = hc$labels[i] %>% substr(.,1,4))
  17. leaf_node[[i]] = temp
  18. }
  20. # get internal node
  21. internal_node = list()
  22. for(i in 1:nrow(hc$merge)){
  23. values = hc$merge[i,]
  24. if(sum(values < 0) == 2){
  25. labs = hc$labels[-values] %>% sort()
  26. height = hc$height[i]
  27. parent_node = ifelse(i == nrow(hc$merge), NA,
  28. c(which(hc$merge[,1] == i),which(hc$merge[,2] == i)))
  29. domains = labs %>% substr(.,1,4) %>% unique()
  30. internal_node[[i]] = list(labs = labs, height = height, parent_node = parent_node,
  31. domain = ifelse(length(domains) == 1, domains,'Mixed'))
  32. }
  33. if(sum(values < 0) == 1){
  34. newlab = hc$labels[-(values[which(values < 0)])]
  35. oldnode = values[which(values > 0)]
  36. labs = c(newlab, internal_node[[oldnode]]$labs) %>% sort()
  37. height = hc$height[i]
  38. parent_node = ifelse(i == nrow(hc$merge), NA,
  39. c(which(hc$merge[,1] == i),which(hc$merge[,2] == i)))
  40. domains = labs %>% substr(.,1,4) %>% unique()
  41. internal_node[[i]] = list(labs = labs, height = height, parent_node = parent_node,
  42. domain = ifelse(length(domains) == 1, domains,'Mixed'))
  43. }
  44. if(sum(values < 0) == 0){
  45. labs = c(internal_node[[values[1]]]$labs,
  46. internal_node[[values[2]]]$labs) %>% sort()
  47. height = hc$height[i]
  48. parent_node = ifelse(i == nrow(hc$merge), NA,
  49. c(which(hc$merge[,1] == i),which(hc$merge[,2] == i)))
  50. domains = labs %>% substr(.,1,4) %>% unique()
  51. internal_node[[i]] = list(labs = labs, height = height, parent_node = parent_node,
  52. domain = ifelse(length(domains) == 1, domains,'Mixed'))
  53. }
  54. }
  56. # get height diff
  57. leaf_height_diff = numeric(length = length(leaf_node))
  58. for(i in 1:length(leaf_node)){
  59. leaf_height_diff[i] = internal_node[[leaf_node[[i]]$parent_node]]$height
  60. }
  61. internal_height_diff = numeric(length = length(internal_node))
  62. for(i in 1:length(internal_node)){
  63. if(!is.na(internal_node[[i]]$parent_node)){
  64. height_self = internal_node[[i]]$height
  65. parent_node = internal_node[[i]]$parent_node
  66. height_parent = internal_node[[parent_node]]$height
  67. internal_height_diff[i] = height_parent - height_self
  69. internal_node_labs = internal_node[[i]]$labs %>% substr(.,1,nchar(.)-5)
  70. if(length(unique(internal_node_labs)) == 1){
  71. if(length(internal_node_labs) == 5){
  72. internal_height_diff[i] = internal_height_diff[i] *10
  73. }
  74. }
  76. }else{
  77. internal_height_diff[i] = 0.5
  78. }
  79. }
  81. # combine height diff into list
  82. for(i in 1:length(leaf_node)){
  83. leaf_node[[i]][['height_diff']] = leaf_height_diff[i]
  84. }
  85. for(i in 1:length(internal_node)){
  86. internal_node[[i]][['height_diff']] = internal_height_diff[i]
  87. }
  89. # combine leaf and internal nodes
  90. if(leafnodes){
  91. nodes_list = c(leaf_node,internal_node)
  92. }else{
  93. nodes_list = c(internal_node)
  94. }
  96. return(nodes_list)
  97. }
  99. #### get p(X|h) ####
  100. Bayes_getlikelihood = function(nodes_list, exp, sigma = .05){
  101. for(i in 1:length(nodes_list)){
  102. occurrence = sum(exp %in% nodes_list[[i]]$labs)
  104. if(occurrence == 0){
  105. nodes_list[[i]][['height_reverse']] = 0
  106. }else{
  107. node_height = nodes_list[[i]]$height+sigma
  108. p_X_h = (1/node_height)^(sum(exp %in% nodes_list[[i]]$labs))
  109. nodes_list[[i]][['height_reverse']] = p_X_h
  110. }
  112. }
  113. return(nodes_list)
  114. }
  116. #### get p(h|X) ####
  117. # Mixedconsider = T
  118. Bayes_getpostprob = function(nodes_list, target,Mixedconsider = T){
  119. # get domain, position of domain, and numver of hypothesis
  120. domain = target %>% substr(.,1,4)
  121. if(isTRUE(Mixedconsider)){
  122. domain_id = 1:length(nodes_list)
  123. num_hypothesis = length(domain_id)
  125. # calculate p(X)
  126. p_X = matrix(data = 0, nrow = num_hypothesis,ncol = 3)
  127. for(i in 1:length(domain_id)){
  128. node_i = domain_id[i]
  129. p_X[i,1] = nodes_list[[node_i]]$height_diff
  130. p_X[i,2] = nodes_list[[node_i]]$height_reverse
  131. }
  132. p_X[,3] = p_X[,1]*p_X[,2]
  133. p_X = sum(p_X[,3],na.rm=T)
  135. # calculate p(h|X)
  136. p_h_X = numeric(length = length(nodes_list))
  137. for(i in 1:length(domain_id)){
  138. node_i = domain_id[i]
  139. p_h = nodes_list[[node_i]]$height_diff
  140. p_X_h = nodes_list[[node_i]]$height_reverse
  141. p_h_X[node_i] = p_X_h * p_h / p_X
  142. }
  144. # combine p(h|X) into node_list
  145. for(i in 1:length(p_h_X)){
  146. nodes_list[[i]][['Post_Prob']] = p_h_X[i]
  147. }
  148. }else{
  149. domain_id = c(which((nodes_list %>% unlist() %>% .[names(.) == 'domain']) == domain))
  150. num_hypothesis = length(domain_id)
  152. # calculate p(X)
  153. p_X = matrix(data = 0, nrow = num_hypothesis,ncol = 3)
  154. for(i in 1:length(domain_id)){
  155. node_i = domain_id[i]
  156. p_X[i,1] = nodes_list[[node_i]]$height_diff
  157. p_X[i,2] = nodes_list[[node_i]]$height_reverse
  158. }
  159. p_X[,3] = p_X[,1]*p_X[,2]
  160. p_X = sum(p_X[,3],na.rm=T)
  162. # calculate p(h|X)
  163. p_h_X = numeric(length = length(nodes_list))
  164. for(i in 1:length(domain_id)){
  165. node_i = domain_id[i]
  166. p_h = nodes_list[[node_i]]$height_diff
  167. p_X_h = nodes_list[[node_i]]$height_reverse
  168. p_h_X[node_i] = p_X_h * p_h / p_X
  169. }
  171. # combine p(h|X) into node_list
  172. for(i in 1:length(p_h_X)){
  173. nodes_list[[i]][['Post_Prob']] = p_h_X[i]
  174. }
  175. }
  177. return(nodes_list)
  178. }
  180. #### get p(y_belongsto_C|X) ####
  181. Bayes_getpredprob = function(nodes_list, exp, target,Mixedconsider = T){
  183. # get which hypothesis did the exps and target both belong
  184. domain = target %>% substr(.,1,4)
  185. if(isTRUE(Mixedconsider)){
  186. domain_id = 1:length(nodes_list)
  188. BothBelong = matrix(data = 0, nrow = length(exp), ncol = length(domain_id))
  189. for(i in 1:length(domain_id)){
  190. node_i = domain_id[i]
  191. for(e in 1:length(exp)){
  192. if(sum(c(exp[e], target) %in% nodes_list[[node_i]]$labs) == length(c(exp[e], target))){
  193. BothBelong[e,i] = 1
  194. }
  195. }
  197. }
  198. BothBelong = as.numeric(BothBelong %>% apply(X = ., MARGIN = 2, FUN = sum) > 0)
  199. # get all post probability
  200. PostProb = nodes_list[domain_id] %>% unlist() %>% .[names(.) == 'Post_Prob'] %>% as.numeric()
  201. # calculate predicted choose probability
  202. pred_choose_prob = sum(PostProb*BothBelong, na.rm = T)
  204. }else{
  205. domain_id = c(which((nodes_list %>% unlist() %>% .[names(.) == 'domain']) == domain))
  207. BothBelong = matrix(data = 0, nrow = length(exp), ncol = length(domain_id))
  208. for(i in 1:length(domain_id)){
  209. node_i = domain_id[i]
  210. for(e in 1:length(exp)){
  211. if(sum(c(exp[e], target) %in% nodes_list[[node_i]]$labs) == length(c(exp[e], target))){
  212. BothBelong[e,i] = 1
  213. }
  214. }
  216. }
  217. BothBelong = as.numeric(BothBelong %>% apply(X = ., MARGIN = 2, FUN = sum) > 0)
  218. # get all post probability
  219. PostProb = nodes_list[domain_id] %>% unlist() %>% .[names(.) == 'Post_Prob'] %>% as.numeric()
  221. # calculate predicted choose probability
  222. pred_choose_prob = sum(PostProb*BothBelong, na.rm = T)
  224. }
  226. return(pred_choose_prob)
  228. }

Bayesian Model functions.R, no license · at the source

Overview

Authors: Guangyao Zhang1, Xiaosha Wang2,3,4, Dingchen Zhang1, Siwen Xie5,6, Lusha Zhu2,3,4,7, Yanchao Bi2,3,4,5
ORCID iDs: Lusha Zhu, Yanchao Bi
  1. Faculty of Psychology, Beijing Normal University,Beijing, China
  2. School of Psychological and Cognitive Sciences, Peking University,Beijing, China
  3. IDG/McGovern Institute for Brain Research, Peking University,Beijing, China
  4. Key Laboratory of Machine Perception (Ministry of Education),Beijing, China
  5. Institute for Artificial Intelligence, Peking University,Beijing, China
  6. Yuanpei College, Peking University,Beijing, China
  7. Peking-Tsinghua Center for Life Sciences, Peking University,Beijing, China
Institutions: Beijing Normal University (China); Peking University (China)
Journal: Nature communications, volume 17, issue 1, article 6366
Dates: received 21 July 2025; accepted 21 April 2026; published online 13 May 2026
Type: Research article · Language: English
License: CC BY-NC-ND
Identifiers: DOI 10.1038/s41467-026-72868-w · PMID 42129217 · PMCID PMC13376883 · OpenAlex W7161005493
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: human (organism), cognitive (subfield)
Methods: Statistics, Machine learning, Connectivity, fMRI & imaging
Keywords: Learning and memory, Cognitive neuroscience
MeSH: Cerebral Cortex*, Concept Formation*, Learning*, Adult, Bayes Theorem, Brain Mapping, Female, Hippocampus, Humans, Large Language Models, Magnetic Resonance Imaging, Male, Models, Neurological, Young Adult (* major topic)
Topic: Memory and Neural Mechanisms (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Funding: National Natural Science Foundation of China (National Science Foundation of China) (32595491)
Citations: cited by 1 paper (Europe PMC); 90 references in the paper

Abstract

The abstract is not reproduced here: the paper's license (CC BY-NC-ND) does not allow it. Read it in the paper, at the publisher or on Europe PMC.

Repository

Its files are read in the Code ↔ Paper reader above.

OSF wrt9s

License: none: the authors keep all their rights
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Languages: R (5)
Size: 73 files, 5 scripts
Software Heritage: not checked
Found in: “Code availability”
Holds: README
Not found: license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: tidyverse (3 files), BayesFactor (1 file)
Availability: 1 check, the latest on 28 September 2026: the link answers (HTTP 200)
  • 28 September 2026: the link answers (HTTP 200)
6 files
At the source:

Code availability statement

The paper has a code availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:

  • it points to the authors' code: OSF wrt9s

Read it in the paper: doi.org/10.1038/s41467-026-72868-w.

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;
  • 5 scripts, each with its path and the digest of its content;
  • no match between paragraphs and code yet;
  • 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

The paper has a data availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:

  • no repository, dataset or request procedure was recognized in it

Read it in the paper: doi.org/10.1038/s41467-026-72868-w.

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 1, 28 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 6 authors, 2 keywords, 14 MeSH terms, 1 funder, 81 references.

Cite

This paper

Zhang, G., Wang, X., Zhang, D., Xie, S., Zhu, L., & Bi, Y. (2026). Cortical knowledge structures guide word concept learning. Nature communications, 17(1), 6366. https://doi.org/10.1038/s41467-026-72868-w

BibTeX

@article{zhang2026cortical,
author = {Zhang, Guangyao and Wang, Xiaosha and Zhang, Dingchen and Xie, Siwen and Zhu, Lusha and Bi, Yanchao},
title = {{Cortical knowledge structures guide word concept learning}},
journal = {Nature communications},
year = {2026},
month = may,
volume = {17},
number = {1},
pages = {6366},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/s41467-026-72868-w},
url = {https://doi.org/10.1038/s41467-026-72868-w},
pmid = {42129217},
pmcid = {PMC13376883}
}

RIS

TY - JOUR
AU - Zhang, Guangyao
AU - Wang, Xiaosha
AU - Zhang, Dingchen
AU - Xie, Siwen
AU - Zhu, Lusha
AU - Bi, Yanchao
TI - Cortical knowledge structures guide word concept learning
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/05/13
VL - 17
IS - 1
SP - 6366
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/s41467-026-72868-w
UR - https://doi.org/10.1038/s41467-026-72868-w
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41467-026-72868-w",
"type": "article-journal",
"title": "Cortical knowledge structures guide word concept learning",
"container-title": "Nature communications",
"author": [
{
"family": "Zhang",
"given": "Guangyao"
},
{
"family": "Wang",
"given": "Xiaosha"
},
{
"family": "Zhang",
"given": "Dingchen"
},
{
"family": "Xie",
"given": "Siwen"
},
{
"family": "Zhu",
"given": "Lusha"
},
{
"family": "Bi",
"given": "Yanchao"
}
],
"container-title-short": "Nat Commun",
"volume": "17",
"issue": "1",
"page": "6366",
"DOI": "10.1038/s41467-026-72868-w",
"PMID": "42129217",
"PMCID": "PMC13376883",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41467-026-72868-w",
"language": "en",
"issued": {
"date-parts": [
[
2026,
5,
13
]
]
}
}

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:10.7554/elife.107933 [code]
Modality-agnostic decoding of vision and language from fMRI.
Journal: eLife
In common: cognitive, 5 references
[2] doi:10.3390/brainsci16070668
Animate Categories Show Higher Cross-Duration Representational Selectivity in Ventral Occipitotemporal Cortex Under Brief Visual Input.
Journal: Brain sciences
In common: 5 references
[3] doi:10.1162/nol.a.244 [code]
A Novel Approach to Map the Causal Impact of Brain Stimulation on Semantic Processing With Language Models.
Journal: Neurobiology of language (Cambridge, Mass.)
In common: 5 references
[4] doi:10.1016/j.cub.2026.05.068 [code]
An abstract relational map emerges in the human medial prefrontal cortex with consolidation.
Journal: Current biology : CB
In common: 4 references
[5] doi:10.1162/imag.a.1201 [code]
All spectral frequencies of neural activity reveal semantic representation in the human anterior ventral temporal cortex.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: tidyverse, 4 references
[6] doi:10.1523/eneuro.0254-25.2026 [code]
Spatiotemporal Dynamics in Prespeech Semantic Category Decoding: An Intracranial EEG Study.
Journal: eNeuro
In common: cognitive, 4 references
[7] doi:10.1038/s41467-025-65499-0
Towards decoding individual words from non-invasive brain recordings
Journal: n/a
In common: 4 references
[8] doi:10.1038/s41467-026-71270-w [code]
Spatiotemporal dynamics of the human cortical functional hierarchy across the lifespan.
Journal: Nature communications
In common: author Yanchao Bi
[9] doi:10.1016/j.isci.2026.117285 [code]
Working memory demands modulate memory brain state engagement.
Journal: iScience
In common: BayesFactor, tidyverse, cognitive, 1 reference
[10] doi:10.1038/s41597-026-07955-0 [code]
Human neuron activity during an 83-minute movie from 2,286 neurons and 29 patients.
Journal: Scientific data
In common: 3 references

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.