OSCR

Predicting emotional valence in autism: a preregistered study in the Bayesian Brain framework.

Code ↔ Paper

13 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.

10 blocks of code here have a permanent identifier (oscr:b:…), listed with the matches below: citable for good, with a page of its own. Select a match and press Cite to copy it. What a block identifier is.

The 13 matches
  1. [1] § Method › Procedure ↔ EMBA_PAL-ADHD/PAL-ADHD_supps-ext.Rmd, lines 101 to 160 · score 0.96 · Lenovo Legion Pro, Lightning eye tracker, Cambridge RS, stable viewing distance, laptop stand, LiveTrack · block oscr:b:FNNM9192SM6N0-8 · Report a mismatch
  2. [2] § Method › Procedure ↔ EMBA_PAL-ADHD/PAL-ADHD_supps.Rmd, lines 100 to 151 · score 0.96 · Lenovo Legion Pro, Lightning eye tracker, Cambridge RS, stable viewing distance, laptop stand, LiveTrack · block oscr:b:70TNB4CCPXH4B-U · Report a mismatch
  3. [3] § Results › Volatility parameters and learning rate updates ↔ EMBA_PAL-ASD/PAL-ASD_results.Rmd, lines 137 to 272 · score 0.94 · Bayesian ANOVA revealed, interpreted cautiously, interaction log, suboptimal fit, anecdotal evidence, ranked learning rates · block oscr:b:SHVNK4F5SSXVF-X · Report a mismatch
  4. [4] § Results › Volatility parameters and learning rate updates ↔ EMBA_PAL-ASD/PAL-ASD_results.Rmd, lines 99 to 135 · score 0.91 · explorative Bernoulli model, preregistered cutoff, lending support, inside ROPE, comparable phasic volatility, phasic volatility predicting · block oscr:b:CBWG6CFMEYECW-E · Report a mismatch
  5. [5] § Results › Model evaluation ↔ EMBA_PAL-ASD/PAL-ASD_supps-rev2.Rmd, lines 117 to 197 · score 0.69 · behavioural patterns, simulated reaction, oHGF, best fitting, posterior predictive checks, correlated · block oscr:b:K1Z9QN7YN6J92-4 · Report a mismatch
  6. [6] § Method › Analysis with Bayesian linear (mixed) models ↔ EMBA_PAL-ASD/PAL-ASD_results.Rmd, lines 99 to 135 · score 0.65 · highest density interval, Bayes Factors, posterior probabilities, HDI, intercept, favour · block oscr:b:CBWG6CFMEYECW-E · Report a mismatch
  7. [7] § Method › Hierarchical gaussian filter ↔ EMBA_PAL-ASD/S4_brms-HGF_PAL-ASD.Rmd, lines 581 to 633 · score 0.65 · cue outcome uncertainty, stimulus surprise, Stimulus uncertainty, Phasic volatility, filter, HGF · block oscr:b:EN0A7J5FNKH82-1 · Report a mismatch
  8. [8] § Method › Hierarchical gaussian filter ↔ EMBA_PAL-ASD/PAL-ASD_supps-rev2.Rmd, lines 841 to 879 · score 0.65 · cue outcome uncertainty, stimulus surprise, Stimulus uncertainty, Phasic volatility, filter, HGF · block oscr:b:D524FQ2WHHBD0-B · Report a mismatch
  9. [9] § Method › Analysis with Bayesian linear (mixed) models ↔ EMBA_PAL-ADHD/PAL-ADHD_results.Rmd, lines 8 to 121 · score 0.64 · Bayes Factors, posterior probabilities, anecdotal, moderate, interval, sum · block oscr:b:38MKW4MHK8XE2-B · Report a mismatch
  10. [10] § Method › Probabilistic associative learning task: emotion recognition ↔ experiment/taskPAL.m, lines 268 to 350 · score 0.57 · fixation cross, blocks, duration, window, tone, emotional · block oscr:b:AFP0549E9G6Q3-4 · Report a mismatch
  11. [11] § Results › Model evaluation ↔ EMBA_PAL-ASD/S5_rev1_PAL_ASD.m, lines 129 to 145 · score 0.56 · RW model, Model identifiability, empirical priors, S5, simulated · block oscr:b:TR37N78BCE3RF-N · Report a mismatch
  12. [12] § Method › Sample ↔ EMBA_PAL-ADHD/PAL-ADHD_supps-ext.Rmd, lines 101 to 160 · score 0.54 · technical difficulties, smaller sample, recruited, suicidality, analysed, ADHD · block oscr:b:FNNM9192SM6N0-8 · Report a mismatch
  13. [13] § Method › Hierarchical gaussian filter ↔ EMBA_PAL-ASD/PAL-ASD_supps-rev2.Rmd, lines 117 to 197 · score 0.51 · Rescorla Wagner, oHGF, empirical priors, RW, simulated, preregistered · block oscr:b:K1Z9QN7YN6J92-4 · Report a mismatch

Paper

Loaded from Europe PMC by your browser; OSCR does not store it: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

R Markdown · 276 lines · 17 KB · GPL-3.0 · 3 matches

  1. ---
  2. title: "PAL ASD results"
  3. author: "I S Plank"
  4. date: "`r Sys.Date()`"
  5. output: html_document
  6. ---
  7. ```{r setup, include=FALSE}
  8. ls.packages = c("knitr",# kable
  9. "ggplot2", # plots
  10. "brms", # Bayesian lmms
  11. "bridgesampling", # bridge_sampler
  12. "tidyverse", # tibble stuff
  13. "ggpubr", # ggarrange
  14. "ggrain", # geom_rain
  15. "bayesplot", # plots for posterior predictive checks
  16. "SBC", # plots for checking computational faithfulness
  17. "rstatix", # anova
  18. "officer", # read_docx
  19. "effectsize", # interpret_bf
  20. "BayesFactor",
  21. "bayestestR" # equivalence_test
  22. )
  23. lapply(ls.packages, library, character.only=TRUE)
  24. # confidence interval functions
  25. lower_ci = function(var) {
  26. unname(quantile(var, probs = 0.025))
  27. }
  28. upper_ci = function(var) {
  29. unname(quantile(var, probs = 0.975))
  30. }
  31. # function to interpret posterior probabilities
  32. interpret_PP = function(p) {
  33. if (p > 0.975) {
  34. "very strong"
  35. } else if (p > 0.95) {
  36. "strong"
  37. } else if (p > 0.90) {
  38. "moderate"
  39. } else if (p > 0.80) {
  40. "small"
  41. } else if (p > 0.70) {
  42. "anecdotal"
  43. } else {
  44. "no"
  45. }
  46. }
  47. brms_dir = "./_brms_models"
  48. # load the models
  49. m.vol = readRDS(file.path(brms_dir, "m_hgf_vol.rds"))
  50. m.om3 = readRDS(file.path(brms_dir, "m_hgf_om3.rds"))
  51. m.ber = readRDS(file.path(brms_dir, "m_hgf_bern.rds"))
  52. # hypotheses
  53. h3a = hypothesis(m.vol, "0 < diagnosis1")
  54. h3b = hypothesis(m.om3, "0 < diagnosis1")
  55. e = hypothesis(m.ber, "sbe4 > 0", alpha = 0.025)
  56. # equivalence
  57. equ.vol = equivalence_test(m.vol)
  58. equ.om3 = equivalence_test(m.om3)
  59. equ.ber = equivalence_test(m.ber)
  60. # get effect sizes (Hedges, 2007)
  61. df.eff.vol = as_draws_df(m.vol) %>%
  62. mutate(
  63. group = -2*b_diagnosis1 / sigma
  64. ) %>% select(group) %>%
  65. pivot_longer(cols = everything(), values_to = "estimate") %>%
  66. group_by(name) %>%
  67. summarise(
  68. ci.lo = lower_ci(estimate),
  69. mean = mean(estimate),
  70. ci.hi = upper_ci(estimate),
  71. interpret = interpret_cohens_d(mean)
  72. )
  73. df.eff.om3 = as_draws_df(m.om3) %>%
  74. mutate(
  75. group = -2*b_diagnosis1 / sigma
  76. ) %>% select(group) %>%
  77. pivot_longer(cols = everything(), values_to = "estimate") %>%
  78. group_by(name) %>%
  79. summarise(
  80. ci.lo = lower_ci(estimate),
  81. mean = mean(estimate),
  82. ci.hi = upper_ci(estimate),
  83. interpret = interpret_cohens_d(mean)
  84. )
  85. ```
  86. # Results
  87. ## Volatility parameters and learning rates
  88. There was no credible difference between the influence of phasic volatility in autistic and comparison adults (*estimate* = `r round(h3a$hypothesis$Estimate,2)` [`r round(h3a$hypothesis$CI.Lower,2)`, `r round(h3a$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h3a$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.vol$mean, 3)` [`r round(df.eff.vol$ci.lo, 3)`, `r round(df.eff.vol$ci.hi, 3)`], see [!FIGURE 4]), contrasting our hypothesis and previous findings. However, there was `r interpret_PP(h3a$hypothesis$Post.Prob)` evidence in favour of a difference with the posterior probability approaching the preregistered cutoff of 95%, falling slightly short of credible evidence but also not lending support to comparable phasic volatility with only `r round(equ.vol$ROPE_Percentage[2]*100, 2)`% of posterior estimates inside the Region of Practical Equivalence (ROPE, highest density interval (HDI) = [`r round(equ.vol$HDI_low[2], 2)`, `r round(equ.vol$HDI_high[2], 2)`]), suggested by Kruschke [!REF2018] for assessing evidence for equivalence. Similarly, `r interpret_PP(e$hypothesis$Post.Prob)` evidence pointed towards increased phasic volatility predicting ASD in the explorative Bernoulli model, with a posterior probability of `r round(e$hypothesis$Post.Prob*100,2)`% (*estimate* = `r round(e$hypothesis$Estimate,2)` [`r round(e$hypothesis$CI.Lower,2)`, `r round(e$hypothesis$CI.Upper,2)`], inside ROPE = `r round(equ.ber$ROPE_Percentage[5]*100, 2)`%, HDI = [`r round(equ.ber$HDI_low[5], 2)`, `r round(equ.ber$HDI_high[5], 2)`]). Last, there was `r interpret_PP(h3b$hypothesis$Post.Prob)` evidence towards higher environmental tonic volatility in autistic than comparison adults (*estimate* = `r round(h3b$hypothesis$Estimate,2)` [`r round(h3b$hypothesis$CI.Lower,2)`, `r round(h3b$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h3b$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.om3$mean, 3)` [`r round(df.eff.om3$ci.lo, 3)`, `r round(df.eff.om3$ci.hi, 3)`], inside ROPE = `r round(equ.om3$ROPE_Percentage[2]*100, 2)`%, HDI = [`r round(equ.om3$HDI_low[2], 2)`, `r round(equ.om3$HDI_high[2], 2)`]).
  89. ```{r dat1, include=FALSE}
  90. # load the models
  91. m.alpha = readRDS(file.path(brms_dir, "m_hgf_alpha.rds"))
  92. aov = readRDS(file.path(brms_dir, "aov_alpha.rds"))
  93. # hypotheses
  94. h4a = hypothesis(m.alpha, "0 > - diagnosis1 + diagnosis1:level1")
  95. h4b = hypothesis(m.alpha, "0 < -diagnosis1 - diagnosis1:level1")
  96. aov.bf = aov@bayesFactor %>% arrange(desc(bf))
  97. # get effect sizes (Hedges, 2007)
  98. df.eff.alpha = as_draws_df(m.alpha) %>%
  99. mutate(
  100. sumvar = sqrt(sigma^2 + sd_subID__Intercept^2 +
  101. sd_subID__level1^2 + sd_subID__change1^2),
  102. group = 2*b_diagnosis1 / sumvar,
  103. h4a = -(-2*b_diagnosis1 + 2*`b_diagnosis1:level1`) / sumvar,
  104. h4b = -(-2*b_diagnosis1 - 2*`b_diagnosis1:level1`) / sumvar
  105. ) %>% select(group, h4a, h4b) %>%
  106. pivot_longer(cols = everything(), values_to = "estimate") %>%
  107. group_by(name) %>%
  108. summarise(
  109. ci.lo = lower_ci(estimate),
  110. mean = mean(estimate),
  111. ci.hi = upper_ci(estimate),
  112. interpret = interpret_cohens_d(mean)
  113. )
  114. ```
  115. For the model regarding learning rates, posterior predictive checks revealed a suboptimal fit of estimated parameters for the change from volatile to postvolatile phase. Thus, we complemented the examination of the posterior distributions with Bayesian ANOVA of the ranked learning rates. Posterior distributions of the preregistered model showed `r interpret_PP(h4a$hypothesis$Post.Prob)` but not credible evidence for increased learning rate of environmental changes (*estimate* = `r round(h4a$hypothesis$Estimate,2)` [`r round(h4a$hypothesis$CI.Lower,2)`, `r round(h4a$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h4a$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.alpha[df.eff.alpha$name == "h4a",]$mean, 3)` [`r round(df.eff.alpha[df.eff.alpha$name == "h4a",]$ci.lo, 3)`, `r round(df.eff.alpha[df.eff.alpha$name == "h4a",]$ci.hi, 3)`]) and `r interpret_PP(1-h4b$hypothesis$Post.Prob)` evidence *against* the hypothesis of decreased learning rate for cue-outcome associations (*estimate* = `r round(h4b$hypothesis$Estimate,2)` [`r round(h4b$hypothesis$CI.Lower,2)`, `r round(h4b$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h4b$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.alpha[df.eff.alpha$name == "h4b",]$mean, 3)` [`r round(df.eff.alpha[df.eff.alpha$name == "h4b",]$ci.lo, 3)`, `r round(df.eff.alpha[df.eff.alpha$name == "h4b",]$ci.hi, 3)`], see [!Figure 5]). The Bayesian ANOVA revealed almost identical fit for two models, both containing the predictors level and change with one of them also including their interaction (without interaction log(*BF*) = `r round(aov.bf$bf[1],2)`, with interaction log(*BF*) = `r round(aov.bf$bf[2],2)`). However, there was only anecdotal evidence in favour of these two models compared to the model with the third-best fit which additionally included the predictor group (log(*BF*) = `r round(aov.bf$bf[3],2)`). Thus, all results regarding the learning rates should be interpreted cautiously.
  116. ## Reaction times, pupil sizes and accuracies
  117. ```{r dat2, include=FALSE}
  118. m.pal = readRDS(file.path(brms_dir, "m_pal.rds"))
  119. m.pup = readRDS(file.path(brms_dir, "m_pup.rds"))
  120. aov.acc = readRDS(file.path(brms_dir, "aov_acc.rds"))
  121. bf.acc = aov.acc@bayesFactor %>% arrange(desc(bf))
  122. # hypotheses
  123. h1b = hypothesis(m.pal, "0 < diagnosis1:expected1")
  124. h1d.1 = hypothesis(m.pal, "diagnosis1:phase1 - diagnosis1:phase2 < 0")
  125. h1d.2 = hypothesis(m.pal, "2*diagnosis1:phase1 + 4*diagnosis1:phase2 < 0")
  126. e2.1 = hypothesis(m.pal, "0 < -2*expected1", alpha = 0.025)
  127. e2.6 = hypothesis(m.pal, "0 < -2*difficulty1 - difficulty2", alpha = 0.025)
  128. h2a = hypothesis(m.pup, "0 < diagnosis1:expected1")
  129. # create accuracies dataframe
  130. load("../data/PAL-ASD_data.RData")
  131. df.acc = df.tsk %>% filter(expected != "neutral") %>% droplevels() %>%
  132. group_by(subID, diagnosis, phase, expected, difficulty) %>%
  133. summarise(acc = mean(acc)*100)
  134. df.diagnosis = df.acc %>%
  135. group_by(diagnosis) %>%
  136. summarise(mean_accuracy = mean(acc, na.rm = T),
  137. sd_accuracy = sd(acc, na.rm = T))
  138. df.expected = df.acc %>%
  139. group_by(expected) %>%
  140. summarise(mean_accuracy = mean(acc, na.rm = T),
  141. sd_accuracy = sd(acc, na.rm = T))
  142. df.difficulty = df.acc %>%
  143. group_by(difficulty) %>%
  144. summarise(mean_accuracy = mean(acc, na.rm = T),
  145. sd_accuracy = sd(acc, na.rm = T))
  146. ## extract predicted differences in ms instead of log data
  147. df.new = df.tsk %>% filter(expected != "neutral") %>% droplevels() %>%
  148. select(diagnosis, phase, expected, difficulty) %>%
  149. distinct() %>%
  150. mutate(
  151. condition = paste(diagnosis, phase, expected, difficulty, sep = "_")
  152. )
  153. df.ms = as.data.frame(
  154. fitted(m.pal, summary = F,
  155. newdata = df.new %>% select(diagnosis, phase, expected, difficulty),
  156. re_formula = NA))
  157. colnames(df.ms) = df.new$condition
  158. # calculate our difference columns
  159. df.ms = df.ms %>%
  160. mutate(
  161. COMP_unexp_exp = rowMeans(across(matches("COMP_.*_unexpected_.*"))) -
  162. rowMeans(across(matches("COMP_.*_expected_.*"))),
  163. ASD_unexp_exp = rowMeans(across(matches("ASD_.*_unexpected_.*"))) -
  164. rowMeans(across(matches("ASD_.*_expected_.*"))),
  165. `h1b_COMP-ASD unexpected-expected` = COMP_unexp_exp - ASD_unexp_exp,
  166. `h1d_ASD-COMP volatile-prevolatile` =
  167. (rowMeans(across(matches("ASD_volatile_.*"))) - rowMeans(across(matches("ASD_prevolatile_.*")))) - (rowMeans(across(matches("COMP_volatile_.*"))) - rowMeans(across(matches("COMP_prevolatile_.*")))),
  168. `h1d_ASD-COMP postvolatile-volatile` =
  169. (rowMeans(across(matches("ASD_postvolatile_.*"))) - rowMeans(across(matches("ASD_volatile_.*")))) - (rowMeans(across(matches("COMP_postvolatile_.*"))) - rowMeans(across(matches("COMP_volatile_.*")))),
  170. `e21_unexpected-expected` = rowMeans(across(matches(".*_unexpected_.*"))) -
  171. rowMeans(across(matches(".*_expected_.*"))),
  172. `e26_difficult-easy` = rowMeans(across(matches(".*_difficult"))) -
  173. rowMeans(across(matches(".*_easy")))
  174. )
  175. equ.pal = equivalence_test(df.ms %>% select(starts_with("h")),
  176. range = rope_range(m.pal))
  177. df.eff.pal = as_draws_df(m.pal) %>%
  178. mutate(across(starts_with("sd")|starts_with("sigma"), ~.^2)) %>%
  179. mutate(
  180. sumvar = sqrt(rowSums(select(., starts_with("sd")|starts_with("sigma")))),
  181. h1b = 4*`b_diagnosis1:expected1` / sumvar,
  182. h1d.1 = 2*(`b_diagnosis1:phase1` - `b_diagnosis1:phase2`) / sumvar,
  183. h1d.2 = 2*(`b_diagnosis1:phase1` + 2*`b_diagnosis1:phase2`) / sumvar,
  184. e.exp = 2*b_expected1 / sumvar,
  185. e.dif = -(-2*b_difficulty1 - b_difficulty2) / sumvar
  186. ) %>% select(starts_with("e.")|starts_with("h")) %>%
  187. pivot_longer(cols = everything(), values_to = "estimate") %>%
  188. group_by(name) %>%
  189. summarise(
  190. ci.lo = lower_ci(estimate),
  191. mean = mean(estimate),
  192. ci.hi = upper_ci(estimate),
  193. interpret = interpret_cohens_d(mean)
  194. )
  195. # pup stuff
  196. df.new = m.pup$data %>%
  197. select(diagnosis, expected, rts) %>%
  198. distinct() %>%
  199. mutate(
  200. condition = paste(diagnosis, expected, rts, sep = "_")
  201. )
  202. df.ms = as.data.frame(
  203. fitted(m.pup, summary = F,
  204. newdata = df.new %>% select(diagnosis, expected, rts),
  205. re_formula = NA))
  206. colnames(df.ms) = df.new$condition
  207. # calculate our difference columns
  208. df.ms = df.ms %>%
  209. mutate(
  210. COMP_expectancy = rowMeans(across(matches("COMP_expected_.*"))) -
  211. rowMeans(across(matches("COMP_unexpected_.*"))),
  212. ASD_expectancy = rowMeans(across(matches("ASD_expected_.*"))) -
  213. rowMeans(across(matches("ASD_unexpected_.*"))),
  214. h2a = COMP_expectancy - ASD_expectancy
  215. )
  216. equ.pup = equivalence_test(df.ms %>% select(starts_with("e_") | starts_with("h")),
  217. range = rope_range(m.pup))
  218. df.eff.pup = as_draws_df(m.pup) %>%
  219. mutate(
  220. sumvar = sqrt(sigma^2 + sd_subID__Intercept^2),
  221. h2a = 4*`b_diagnosis1:expected1` / sumvar
  222. ) %>% select(starts_with("e.")|starts_with("h")) %>%
  223. pivot_longer(cols = everything(), values_to = "estimate") %>%
  224. group_by(name) %>%
  225. summarise(
  226. ci.lo = lower_ci(estimate),
  227. mean = mean(estimate),
  228. ci.hi = upper_ci(estimate),
  229. interpret = interpret_cohens_d(mean)
  230. )
  231. ```
  232. There was no credible evidence for differences between autistic and comparison adults with regards to the effect of expectancy (*estimate* = `r round(h1b$hypothesis$Estimate,2)` [`r round(h1b$hypothesis$CI.Lower,2)`, `r round(h1b$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h1b$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.pal[df.eff.pal$name == "h1b",]$mean, 3)` [`r round(df.eff.pal[df.eff.pal$name == "h1b",]$ci.lo, 3)`, `r round(df.eff.pal[df.eff.pal$name == "h1b",]$ci.hi, 3)`], inside ROPE = `r round(equ.pal$ROPE_Percentage[1]*100, 2)`%, HDI = [`r round(equ.pal$HDI_low[1], 2)`, `r round(equ.pal$HDI_high[1], 2)`], see [!Figure 1]) or transition effects (prevolatile to volatile: *estimate* = `r round(h1d.1$hypothesis$Estimate,2)` [`r round(h1d.1$hypothesis$CI.Lower,2)`, `r round(h1d.1$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h1d.1$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.pal[df.eff.pal$name == "h1d.1",]$mean, 3)` [`r round(df.eff.pal[df.eff.pal$name == "h1d.1",]$ci.lo, 3)`, `r round(df.eff.pal[df.eff.pal$name == "h1d.1",]$ci.hi, 3)`], inside ROPE = `r round(equ.pal$ROPE_Percentage[2]*100, 2)`%, HDI = [`r round(equ.pal$HDI_low[2], 2)`, `r round(equ.pal$HDI_high[2], 2)`]; volatile to postvolatile: *estimate* = `r round(h1d.2$hypothesis$Estimate,2)` [`r round(h1d.2$hypothesis$CI.Lower,2)`, `r round(h1d.2$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h1d.2$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.pal[df.eff.pal$name == "h1d.2",]$mean, 3)` [`r round(df.eff.pal[df.eff.pal$name == "h1d.2",]$ci.lo, 3)`, `r round(df.eff.pal[df.eff.pal$name == "h1d.2",]$ci.hi, 3)`], inside ROPE = `r round(equ.pal$ROPE_Percentage[3]*100, 2)`%, HDI = [`r round(equ.pal$HDI_low[3], 2)`, `r round(equ.pal$HDI_high[3], 2)`]) based on reaction times. Similarly, there was `r interpret_PP(h2a$hypothesis$Post.Prob)` credible difference in the effect of expectancy on pupil sizes in autistic compared to non-autistic adults (*estimate* = `r round(h2a$hypothesis$Estimate,2)` [`r round(h2a$hypothesis$CI.Lower,2)`, `r round(h2a$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(h2a$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.pup[df.eff.pup$name == "h2a",]$mean, 3)` [`r round(df.eff.pup[df.eff.pup$name == "h2a",]$ci.lo, 3)`, `r round(df.eff.pup[df.eff.pup$name == "h2a",]$ci.hi, 3)`]), in contrast to our hypothesis. Across groups, there was `r interpret_PP(e2.1$hypothesis$Post.Prob)` evidence that expected trials led to faster responses than unexpected trials (*estimate* = `r round(e2.1$hypothesis$Estimate,2)` [`r round(e2.1$hypothesis$CI.Lower,2)`, `r round(e2.1$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(e2.1$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.pal[df.eff.pal$name == "e.exp",]$mean, 3)` [`r round(df.eff.pal[df.eff.pal$name == "e.exp",]$ci.lo, 3)`, `r round(df.eff.pal[df.eff.pal$name == "e.exp",]$ci.hi, 3)`]) and easy expressions led to faster responses than difficult expressions (*estimate* = `r round(e2.6$hypothesis$Estimate,2)` [`r round(e2.6$hypothesis$CI.Lower,2)`, `r round(e2.6$hypothesis$CI.Upper,2)`], *posterior probability* = `r round(e2.6$hypothesis$Post.Prob*100,2)`%, *δ* = `r round(df.eff.pal[df.eff.pal$name == "e.dif",]$mean, 3)` [`r round(df.eff.pal[df.eff.pal$name == "e.dif",]$ci.lo, 3)`, `r round(df.eff.pal[df.eff.pal$name == "e.dif",]$ci.hi, 3)`]), suggesting that our task manipulations were successful.
  233. Accuracies were generally high, with a grand average of `r round(mean(df.acc$acc),1)`% accurate responses across diagnostic groups. An explorative Bayesian ANOVA of the ranked accuracies revealed influences of group, expectancy and difficulty but no interaction effects (log(*BF*) = `r round(bf.acc$bf[1],2)`). Accuracies were higher in the comparison than the autistic group (ASD: `r round(df.diagnosis[df.diagnosis$diagnosis == "ASD",]$mean_accuracy,1)` ± `r round(df.diagnosis[df.diagnosis$diagnosis == "ASD",]$sd_accuracy,1)`%; COMP: `r round(df.diagnosis[df.diagnosis$diagnosis == "COMP",]$mean_accuracy,1)` ± `r round(df.diagnosis[df.diagnosis$diagnosis == "COMP",]$sd_accuracy,1)`%) and in the expected than the unexpected trials (expected: `r round(df.expected[df.expected$expected == "expected",]$mean_accuracy,1)` ± `r round(df.expected[df.expected$expected == "expected",]$sd_accuracy,1)`%; unexpected: `r round(df.expected[df.expected$expected == "unexpected",]$mean_accuracy,1)` ± `r round(df.expected[df.expected$expected == "unexpected",]$sd_accuracy,1)`%). Furthermore, accuracies were lower in the difficult condition (easy: `r round(df.difficulty[df.difficulty$difficulty == "easy",]$mean_accuracy,1)` ± `r round(df.difficulty[df.difficulty$difficulty == "easy",]$sd_accuracy,1)`%; hard: `r round(df.difficulty[df.difficulty$difficulty == "difficult",]$mean_accuracy,1)` ± `r round(df.difficulty[df.difficulty$difficulty == "difficult",]$sd_accuracy,1)`%).

PAL-ASD_results.Rmd at commit d399a35, under GPL-3.0 · at the source

Overview

  1. Department of Psychiatry and Psychotherapy, LMU University Hospital, LMU Medizin, Ludwig-Maximilians-Universität München,Nußbaumstraße 7, Munich, 80336 Germany
  2. Department of Sociology, Faculty of Behavioral and Social Sciences, University of Groningen,Grote Rozenstraat 1/2, Groningen, 9712 TS Netherlands
  3. Department of Psychology, Faculty of Psychology and Neuroscience, Maastricht University,Universiteitssingel 40, Maastricht, 6229 ER Netherlands
  4. Department of Psychology, LMU Munich,Leopoldstr. 13, 80802 Munich, Germany
  5. NeuroImaging Core Unit Munich (NICUM), LMU Munich,Nußbaumstraße 7, Munich, 80336 Germany
Institutions: Ludwig-Maximilians-Universität München (Germany); University of Groningen (Netherlands); Maastricht University (Netherlands)
Journal: Molecular autism, volume 17, issue 1, article 32
Dates: received 6 November 2025; accepted 13 July 2026; published online 21 July 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1186/s13229-026-00730-3 · PMID 42482104 · PMCID PMC13435570 · OpenAlex W7169849249
Open access: gold, a free copy (OpenAlex)
Status: code verified (how the code is found and verified)
Categories: behavior only (modality), human (organism), autism (population), cognitive (subfield)
Methods: Machine learning, Statistics, Physiology & signal measures
Keywords: Probabilistic associative learning, Hierarchical gaussian filter, Autism spectrum disorder, Volatility, Predictive coding, Bayesian brain framework
MeSH: Autism Spectrum Disorder*, Autistic Disorder*, Brain*, Emotions*, Adult, Bayes Theorem, Facial Expression, Female, Humans, Male, Reaction Time, Young Adult (* major topic)
Topic: Autism Spectrum Disorder Research (Cognitive Neuroscience, Neuroscience), according to OpenAlex
Funding: Klinikum der Universität München (6933)
Citations: not cited yet (Europe PMC); 58 references in the paper

Abstract

Background: Autism spectrum disorder (ASD) affects social interaction, communication and behavioural flexibility. Recent theories grounded in the Bayesian Brain Framework propose that autistic individuals process environmental uncertainty differently, overweighting volatility when updating predictions about the world. However, the robustness of these predictive processing differences and their relevance to core symptoms remains unclear.

Methods: Lawson and colleagues [1] reported a tendency to overestimate environmental volatility in autistic adults based on belief states extracted using a Hierarchical Gaussian Filter. We extended their paradigm to a domain central to the autistic experience: rather than distinguishing houses from faces, participants detected valence in emotional facial expressions. Preregistered hypotheses were evaluated using Bayesian linear mixed models based on data of 22 autistic and 22 non-autistic participants.

Results: Despite using an identical computational model to extract belief states, we did not find robust differences between autistic and non-autistic adults, though autistic participants showed a non-credible trend towards increased processing of phasic volatility. Largely comparable probabilistic associative learning was also reflected in response times and pupil sizes.

Limitations: While participants reacted faster to expected than unexpected trials in the beginning, this effect decreased over the course of the experiment; thus, possibly indicating that the individual blocks of stable cue-outcome associations were too short.

Conclusions: We were unable to find credible differences in environmental volatility and learning rate updates using a task designed to probe a key autistic difficulty, facial emotion recognition. The current results call into question the generalisability and clinical relevance of prior findings.

Supplementary Information: The online version contains supplementary material available at 10.1186/s13229-026-00730-3.

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 13 matches between paragraphs and lines of code.

IreneSophia/EMBA_PAL

License: GPL-3.0
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: d399a3524c0e563706de2ff0bfc0969d4b122c07, 24 July 2026
Languages: MATLAB (88), R (27), Python (6)
Size: 8,902 files, 121 scripts
Software Heritage: not archived
Found in: the text, “Method”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: tidyverse (25 files), brms (20 files), ggplot2 (18 files), BayesFactor (16 files), easystats (15 files), ggpubr (15 files), rstatix (15 files), Stan (9 files), Parallel Computing Toolbox (5 files), pandas (5 files), statsmodels (4 files), Matplotlib (3 files), NumPy (3 files), Statistics and Machine Learning Toolbox (1 file), Psychtoolbox (1 file), reshape2 (1 file), SciPy (1 file)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
123 files

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;
  • 121 scripts, each with its path and the digest of its content;
  • 13 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

Datasets cited

Data availability

The datasets generated and analysed during the current study are available on OSF, [https://osf.io/krp4s].

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 1, 27 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 6 authors, 6 keywords, 12 MeSH terms, 1 funder, 45 references.

Cite

This paper

Plank, I. S., Pior, A., Yurova, A., Nowak, J., Shi, Z., & Falter-Wagner, C. M. (2026). Predicting emotional valence in autism: a preregistered study in the Bayesian Brain framework. Molecular autism, 17(1), 32. https://doi.org/10.1186/s13229-026-00730-3

BibTeX

@article{plank2026predicting,
author = {Plank, Irene Sophia and Pior, Alexandra and Yurova, Anna and Nowak, Julia and Shi, Zhuanghua and Falter-Wagner, Christine M.},
title = {{Predicting emotional valence in autism: a preregistered study in the Bayesian Brain framework}},
journal = {Molecular autism},
year = {2026},
month = jul,
volume = {17},
number = {1},
pages = {32},
publisher = {BMC},
issn = {2040-2392},
doi = {10.1186/s13229-026-00730-3},
url = {https://doi.org/10.1186/s13229-026-00730-3},
pmid = {42482104},
pmcid = {PMC13435570}
}

RIS

TY - JOUR
AU - Plank, Irene Sophia
AU - Pior, Alexandra
AU - Yurova, Anna
AU - Nowak, Julia
AU - Shi, Zhuanghua
AU - Falter-Wagner, Christine M.
TI - Predicting emotional valence in autism: a preregistered study in the Bayesian Brain framework
T2 - Molecular autism
J2 - Mol Autism
PY - 2026
DA - 2026/07/21
VL - 17
IS - 1
SP - 32
SN - 2040-2392
PB - BMC
DO - 10.1186/s13229-026-00730-3
UR - https://doi.org/10.1186/s13229-026-00730-3
LA - en
ER -

CSL-JSON

{
"id": "10.1186/s13229-026-00730-3",
"type": "article-journal",
"title": "Predicting emotional valence in autism: a preregistered study in the Bayesian Brain framework",
"container-title": "Molecular autism",
"author": [
{
"family": "Plank",
"given": "Irene Sophia"
},
{
"family": "Pior",
"given": "Alexandra"
},
{
"family": "Yurova",
"given": "Anna"
},
{
"family": "Nowak",
"given": "Julia"
},
{
"family": "Shi",
"given": "Zhuanghua"
},
{
"family": "Falter-Wagner",
"given": "Christine M."
}
],
"container-title-short": "Mol Autism",
"volume": "17",
"issue": "1",
"page": "32",
"DOI": "10.1186/s13229-026-00730-3",
"PMID": "42482104",
"PMCID": "PMC13435570",
"ISSN": "2040-2392",
"publisher": "BMC",
"URL": "https://doi.org/10.1186/s13229-026-00730-3",
"language": "en",
"issued": {
"date-parts": [
[
2026,
7,
21
]
]
}
}

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:10.1162/imag.a.942 [code]
Are prediction error modifications domain specific in autism but domain general in ADHD?
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: BayesFactor, brms, Stan, 6 other tools, autism, cognitive, 10 references, 6 authors
[2] doi:10.1002/hbm.70052 [code]
Neural Mechanisms of Social Interaction Perception: Observing Interpersonal Synchrony Modulates Action Observation Network Activation and Is Spared in Autism.
Journal: Human brain mapping
In common: BayesFactor, brms, Stan, 4 other tools, autism, cognitive, 2 references, 2 authors
[3] doi:10.1371/journal.pcbi.1011473 [code]
Testing predictive coding theories of autism spectrum disorder using models of active inference.
Journal: PLoS computational biology
In common: Statistics and Machine Learning Toolbox, behavior only, autism, 12 references
[4] doi:10.1371/journal.pbio.3002978 [code]
Basal ganglia components have distinct computational roles in decision-making dynamics under conflict and uncertainty.
Journal: PLoS biology
In common: BayesFactor, brms, Stan, 12 other tools, cognitive
[5] doi:10.1162/imag.a.1258 [code]
Non-specific increase in alpha power during a neurofeedback session targeting its downregulation.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: BayesFactor, brms, Stan, 8 other tools, 2 references
[6] doi:10.1038/s44271-026-00431-w [code]
Alpha power increases spontaneously during a neurofeedback session.
Journal: Communications psychology
In common: BayesFactor, brms, Stan, 7 other tools, cognitive, 2 references
[7] doi:10.1038/s41467-023-38580-9 [code]
Neural correlates of hierarchical predictive processes in autistic adults.
Journal: Nature communications
In common: Psychtoolbox, Statistics and Machine Learning Toolbox, autism, cognitive, 9 references
[8] doi:10.1523/eneuro.0065-24.2024 [code]
Gaze and Arrows: Does the Gaze-Following Patch in the Posterior Temporal Cortex Differentiate Social and Symbolic Spatial Cues?
Journal: eNeuro
In common: BayesFactor, brms, Stan, 10 other tools, cognitive
[9] doi:10.1111/psyp.70163 [code]
Spontaneous Modulation of Standard EEG Frequency Bands During a Neurofeedback-Like Task.
Journal: Psychophysiology
In common: brms, Stan, Psychtoolbox, 6 other tools, cognitive, 2 references
[10] doi:10.1016/j.isci.2025.114212 [code]
Alpha power indexes working memory load for durations.
Journal: iScience
In common: BayesFactor, brms, easystats, 10 other tools, cognitive

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.

Add the badge to the code's README

OSCR: the badge of this paper

If you are an author or a maintainer of this paper's code, you can add this badge. It says what OSCR knows of the paper (its code, its validated tracing map) and links the code to this page. Paste one of these snippets under the README's title. Keep the two marks around it, so that the badge can be replaced in place and is never added twice. OSCR changes nothing in your repository.

Markdown
<!-- oscr-badge:start -->
[![OSCR](https://openscicode.org/badge/doi/10.1186/s13229-026-00730-3.svg)](https://openscicode.org/badge/doi/10.1186/s13229-026-00730-3)
<!-- oscr-badge:end -->
reStructuredText
.. oscr-badge:start

.. image:: https://openscicode.org/badge/doi/10.1186/s13229-026-00730-3.svg
   :target: https://openscicode.org/badge/doi/10.1186/s13229-026-00730-3
   :alt: OSCR

.. oscr-badge:end
HTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.1186/s13229-026-00730-3"><img src="https://openscicode.org/badge/doi/10.1186/s13229-026-00730-3.svg" alt="OSCR"></a>
<!-- oscr-badge:end -->

On GitHub, open the README, choose Edit (the pencil), paste the snippet under the title, then "Commit changes".

If you choose to, you can have the badge kept up to date for you, with oscr badge --write from the command line or with the GitHub Action in your repository's workflows. Every other form and option (a card, AsciiDoc, LaTeX, shields.io) is on the badge 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: code errors and mismatches between the code and the paper about this paper and its code, from readers and its authors, beside the paper and its tracing map.

Report a mismatch between the code and the paper · Report a code error (signed in)

Reproductions: reports from readers who ran the authors' code and did not get the paper's result, what they ran, with which environment, commit and data.

Report a reproduction failure (signed in)

Versions of its code: the releases of its code tied to a version of this paper (the preprint, the accepted manuscript, the version of record…), each with the tracing map versioned with it.

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