OSCR

Regulation of the decision threshold by the locus coeruleus.

Code ↔ Paper

2 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 2 matches
  1. [1] § Methods › Drift diffusion model ↔ R/pulse_diffusion_model.R, lines 539–600 · score 0.71 · Wiener diffusion, diffusion model, lower boundary, upper boundary, drift rate, variable
  2. [2] § Methods › Drift diffusion model ↔ R/drift_diffusion_model.R, lines 378–448 · score 0.68 · drift diffusion model, Wiener diffusion, lower boundary, drift rate, noise, DDM

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 · 600 lines · 22 KB · GPL-3.0 · 1 match

  1. #####################################
  2. # Helper functions for pulse model object
  4. pulse_fp_obj <- function(pars,
  5. dat=NULL,
  6. transform_pars=F,
  7. check_constraints=T,
  8. debug=F,
  9. ...){
  11. ### check constraints
  13. checks = private$objective_checks(pars,
  14. transform_pars,
  15. check_constraints,
  16. reverse_v=F,
  17. debug)
  18. pars = checks[[1]]
  19. pass = checks[[2]]
  21. if (!is.na(pass) & !pass) {
  22. nll = 1e10
  23. if(debug) cat("nll =", nll, "\n")
  24. return(nll)
  25. }
  27. if(is.null(dat)) {
  28. dat = self$data
  29. }
  31. ### loop through conditions to get likelihood
  33. nll = 0
  34. if (private$par_matrix[, .N] < dat[, .N]) {
  35. for(i in 1:private$par_matrix[, .N]){
  36. sub_dat = copy(dat)
  37. for(j in 1:length(private$sim_cond)){
  38. sub_dat = sub_dat[get(private$sim_cond[j]) == private$par_matrix[i, get(private$sim_cond[j])]]
  39. }
  40. sub_nll = do.call(pulse_nll, c(list(choice = sub_dat[, response],
  41. rt = sub_dat[, rt],
  42. stimuli = private$stim_list[[i]]),
  43. as.list(private$par_matrix[i, -(1:length(private$sim_cond))]),
  44. private$fixed,
  45. v_scale=private$v_scale,
  46. bounds=private$bounds,
  47. urgency=private$urgency,
  48. ...))
  49. nll = nll + sub_nll
  50. }
  51. } else {
  53. nll = do.call(pulse_nll, c(list(choice=dat[,response]),
  54. list(rt=dat[, rt]),
  55. stimuli=list(private$stim_list[[1]]),
  56. as.list(private$par_matrix),
  57. private$fixed,
  58. v_scale=private$v_scale,
  59. bounds=private$bounds,
  60. urgency=private$urgency,
  61. ...))
  62. }
  64. if (is.nan(nll)) browser()
  66. if(debug) cat("nll =", round(nll, 3), "\n")
  68. nll
  70. }
  73. #' @noRd
  74. #' @importFrom foreach %do% %dopar%
  75. pulse_x2_obj = function(pars,
  76. data_q=NULL,
  77. n_sim=1,
  78. min_p=1e-10,
  79. transform_pars=F,
  80. check_constraints=T,
  81. debug=F,
  82. seed=-1L,
  83. ...) {
  85. if (seed > 0) {
  86. set.seed(seed)
  87. }
  89. ### check constraints
  91. checks = private$objective_checks(pars,
  92. transform_pars,
  93. check_constraints,
  94. reverse_v=F,
  95. debug)
  96. pars = checks[[1]]
  97. pass = checks[[2]]
  99. if (!is.na(pass) & !pass) {
  100. chisq = 1e10
  101. if(debug) cat(chisq, "\n")
  102. return(chisq)
  103. }
  105. if (is.null(data_q)) {
  106. data_q = copy(self$data_q)
  107. }
  110. ### loop through conditions to get chisquare
  112. pars_only_mat = copy(private$par_matrix)
  113. pars_only_mat = pars_only_mat[, -(1:(length(private$sim_cond)))]
  114. rt_q_cols = (length(data_q)-length(private$p_q)+2):length(data_q)
  116. chisq = 0
  118. for (i in 1:private$par_transform[, .N]) {
  120. # simulate trials
  121. par_list = as.list(pars_only_mat[i])
  123. this_sim = setDT(do.call(sim_pulse, c(n=n_sim,
  124. list(stimuli=private$stim_list[[i]]),
  125. par_list,
  126. private$fixed,
  127. v_scale=private$v_scale,
  128. bounds=private$bounds,
  129. urgency=private$urgency,
  130. ...))$behavior)
  132. # get rt quantile matrix
  133. sub_q = copy(data_q)
  134. for(j in 1:length(private$sim_cond)) {
  135. sub_q = sub_q[get(private$sim_cond[j]) == private$par_matrix[i, get(private$sim_cond[j])]]
  136. }
  137. rt_q_mat = as.matrix(sub_q[, rt_q_cols, .(response), with=F])
  138. n_rt = sub_q[, n_response, .(response)][, n_response]
  139. sim_rts = list(this_sim[response == 0, rt],
  140. this_sim[response == 1, rt],
  141. this_sim[is.na(response), rt])
  143. chisq = chisq + quantile_chisquare(sim_rts, rt_q_mat, private$p_q, n_rt)
  145. }
  147. if(is.na(chisq)) chisq = 1e10
  148. if(debug) cat(chisq, "\n")
  149. chisq
  151. }
  153. #' @noRd
  154. #' @importFrom foreach %do% %dopar%
  155. pulse_qmpe_obj = function(pars,
  156. data_q=NULL,
  157. n_sim=1,
  158. min_p=1e-10,
  159. transform_pars=F,
  160. check_constraints=T,
  161. debug=F,
  162. seed=-1L,
  163. ...) {
  165. if (seed > 0) {
  166. set.seed(seed)
  167. }
  169. ### check constraints
  171. checks = private$objective_checks(pars,
  172. transform_pars,
  173. check_constraints,
  174. reverse_v=F,
  175. debug)
  176. pars = checks[[1]]
  177. pass = checks[[2]]
  179. if (!is.na(pass) & !pass) {
  180. qmpe_nll = 1e10
  181. if(debug) cat(qmpe_nll, "\n")
  182. return(qmpe_nll)
  183. }
  185. if (is.null(data_q)) {
  186. data_q = copy(self$data_q)
  187. }
  190. ### loop through conditions to get chisquare
  192. pars_only_mat = copy(private$par_matrix)
  193. pars_only_mat = pars_only_mat[, -(1:(length(private$sim_cond)))]
  194. rt_q_cols = (length(data_q)-length(private$p_q)+2):length(data_q)
  196. qmpe_nll = 0
  198. for (i in 1:private$par_transform[, .N]) {
  200. # simulate trials
  201. par_list = as.list(pars_only_mat[i])
  203. this_sim = setDT(do.call(sim_pulse, c(n=n_sim,
  204. list(stimuli=private$stim_list[[i]]),
  205. par_list,
  206. private$fixed,
  207. v_scale=private$v_scale,
  208. bounds=private$bounds,
  209. urgency=private$urgency,
  210. seed=seed,
  211. ...))$behavior)
  213. # get rt quantile matrix
  214. sub_q = copy(data_q)
  215. for(j in 1:length(private$sim_cond)) {
  216. sub_q = sub_q[get(private$sim_cond[j]) == private$par_matrix[i, get(private$sim_cond[j])]]
  217. }
  218. rt_q_mat = as.matrix(sub_q[, rt_q_cols, .(response), with=F])
  219. n_rt = sub_q[, n_response, .(response)][, n_response]
  220. sim_rts = list(this_sim[response == 0, rt],
  221. this_sim[response == 1, rt],
  222. this_sim[is.na(response), rt])
  224. qmpe_nll = qmpe_nll + qmpe(sim_rts, rt_q_mat, private$p_q, n_rt, min_p=min_p)
  226. }
  228. if(is.na(qmpe_nll)) qmpe_nll = 1e10
  229. if(debug) cat(qmpe_nll, "\n")
  230. qmpe_nll
  232. }
  235. check_pdm_constraints <- function(){
  236. par_matrix_names = names(private$par_matrix)
  238. checks = sum(private$par_matrix[, (a <= 0)]) # a
  239. checks = checks + sum(private$par_matrix[, (t0 < 0)]) # t0
  240. if ("s" %in% par_matrix_names)
  241. checks = checks + sum(private$par_matrix[, s <= 0]) # s
  242. if ("z" %in% par_matrix_names)
  243. checks = checks + sum(private$par_matrix[, (z <= 0) | (z >= 1)]) # z
  244. else
  245. z = 0.5
  246. if ("sv" %in% par_matrix_names)
  247. checks = checks + sum(private$par_matrix[, (sv < 0)]) # sv
  248. if ("sz" %in% par_matrix_names)
  249. checks = checks + sum(private$par_matrix[, (sz < 0) | (sz >= 1)]) # sz
  250. if ("st0" %in% par_matrix_names)
  251. checks = checks + sum(private$par_matrix[, (st0 < 0) | (st0 >= 1)]) # st0
  252. if ("lambda" %in% par_matrix_names)
  253. checks = checks + sum(private$par_matrix[, (lambda < 0)]) # lambda
  254. if ("aprime" %in% par_matrix_names)
  255. checks = checks + sum(private$par_matrix[, (aprime < 0) | (aprime > 1)]) # aprime
  256. if ("kappa" %in% par_matrix_names)
  257. checks = checks + sum(private$par_matrix[, kappa < 0]) # kappa
  258. if ("tc" %in% par_matrix_names)
  259. checks = checks + sum(private$par_matrix[, tc <= 0]) # tc
  260. if ("uslope" %in% par_matrix_names)
  261. checks = checks + sum(private$par_matrix[, uslope <= 0]) # uslope
  262. if ("udelay" %in% par_matrix_names)
  263. checks = checks + sum(private$par_matrix[, udelay <= 0]) # udelay
  264. if ("umag" %in% par_matrix_names)
  265. checks = checks + sum(private$par_matrix[, umag <= 0]) # umag
  267. return(checks == 0)
  268. }
  271. #' set diffusion model objective function (for internal use)
  272. #'
  273. #' Change the objective function used to fit the diffusion model. Options include:
  274. #' \describe{
  275. #' \item{fp}{use the fast-dm method implemented in the rtdists package}
  276. #' \item{chisq}{simulate the first passage time distribution using the Euler-Maruyama method, and compare simulated and observed distrubtions using the chisquare statistic}
  277. #' }
  278. #'
  279. #' @usage model$set_objective(objective)
  280. #'
  281. #' @param objective string; the objective function to be used; either "fp", "chisq", or "qmpe"
  282. #'
  283. #' @return modifies the field \code{obj}
  284. #'
  285. set_pdm_objective <- function(objective="chisq") {
  287. if (objective == "fp") {
  288. self$obj = private$fp_obj
  289. } else if (objective == "chisq") {
  290. self$obj = private$chisq_obj
  291. } else if (objective == "qmpe") {
  292. self$obj = private$qmpe_obj
  293. } else {
  294. stop("specified objective not supported")
  295. }
  297. invisible(self)
  299. }
  302. #' predict pulse model (for internal use)
  303. #'
  304. #' Predict behavior with given pulse model parameters.
  305. #' This function is only intended for use with a diffusion model object,
  306. #' and should not be called directly outside of the diffusion model class.
  307. #'
  308. #' @usage model$predict(pars=NULL, n=10000, ...)
  309. #'
  310. #' @param pars numeric vector; vector of parameters. If NULL, uses model$solution$pars.
  311. #' @param n integer; number of decisions to simulate for each condition. If the number of conditions is equal to the length of the data, e.g. if using as_function with a continuous predictor, ignores \code{n} and simulates one decision per condition
  312. #' @param method string; "euler" for euler-maruyama simulation or "fp" for Fokker-Planck method
  313. #' @param stim_list list of arrays; 2 x timepoints x trials array of stimuli to simulate for each condition
  314. #' @param trial_code list of integer vector; trial indexes to which the new stimuli are associated with.
  315. #' @param ... additional arguments passed to method (either \code{sim_pulse} or \code{pulse_fp_fpt})
  316. #'
  317. #' @return data.table with simulation conditions, decision (upper or lower boundary) and response time
  318. #'
  319. #' @keywords internal
  320. #'
  321. predict_pulse_model = function(pars=NULL, n=1, method="euler", stim_list=NULL, trial_code=NULL, ...){
  323. if (is.null(stim_list) | (length(stim_list) != private$par_transform[, .N])) {
  324. stim_list = private$stim_list
  325. }
  327. if(is.null(pars)) {
  328. if (is.null(self$solution)) {
  329. stop("if parameters have not been fit (i.e. no model solution), must supply pars!")
  330. }
  331. pars = self$solution$pars
  332. }
  333. private$set_params(pars, reverse_v=F)
  335. ### loop through conditions
  337. d_pred = data.table()
  339. if (private$par_matrix[, .N] < self$data[, .N]) {
  341. for (i in 1:private$par_matrix[, .N]) {
  343. this_stim_trials = dim(stim_list[[i]])[3]
  345. # check stimulus lengths and trial code
  346. if (is.null(trial_code)) {
  347. if ((this_stim_trials != dim(private$stim_list[[i]])[3])) {
  348. stop(paste("new stimulus array length for condition", i, "is not equal to number of trials"))
  349. } else {
  350. this_trial_code = 1:this_stim_trials
  351. }
  352. } else {
  353. if (this_stim_trials != length(trial_code[[i]])) {
  354. stop(paste("new stimulus array length for condition", i, "is not equal to length of trial code"))
  355. } else {
  356. this_trial_code = trial_code[[i]]
  357. }
  358. }
  360. # get predicted behavior
  361. if (method == "euler") {
  363. this_par_list = as.list(private$par_matrix[i, -(1:length(private$sim_cond))])
  364. this_par_values = as.numeric(this_par_list)
  365. this_par_names = names(this_par_list)
  367. this_sim = setDT(self$simulate(n,
  368. stimuli=stim_list[[i]],
  369. this_par_values,
  370. ...)$behavior)
  371. } else {
  373. stop("method not implemented")
  375. }
  377. d_pred = rbind(d_pred,
  378. data.table(private$par_matrix[i, 1:length(private$sim_cond)],
  379. this_sim))
  381. }
  383. }
  385. d_pred
  387. }
  390. #' simulate pulse model (for internal use)
  391. #'
  392. #' simulate pulse DDM with given stimulus and model parameters..
  393. #' This function is only intended for use with a pulse model object,
  394. #' and should not be called directly outside of the pulse model class.
  395. #' Please use \code{sim_pulse} as a standalone function.
  396. #'
  397. #' @usage model$simulate(n, stimuli, par_values, par_names=NULL, ...)
  398. #'
  399. #' @param n integer; number of decisions to simulate per stimulus
  400. #' @param pars numeric vector; vector of parameters
  401. #' @param par_names character vector; vector of parameter names. If pars is not named list, must supply parameter name vector!
  402. #' @param ... additional arguments passed to \code{sim_pulse}
  403. #'
  404. #' @return data.table with simulation conditions, decision (upper or lower boundary) and response time
  405. #'
  406. #' @keywords internal
  407. #'
  408. simulate_pulse_model = function(n, stimuli, pars, ...) {
  410. if (length(pars) != length(self$par_names)) {
  411. stop("supplied parameter vector must be the same length as the number of parameters in the model")
  412. }
  414. names(pars) = self$par_names
  416. if (is.null(stimuli) | class(stimuli) != "array") stop("Must provide stimuli as a 3-d array (2 x timepoint x trials)")
  419. do.call(sim_pulse, c(n=n,
  420. list(stimuli=stimuli),
  421. as.list(pars),
  422. v_scale=private$v_scale,
  423. bounds=private$bounds,
  424. urgency=private$urgency,
  425. ...))
  427. }
  430. init_pulse_model = function(dat,
  431. model_name="pdm",
  432. stim_var=NULL,
  433. stim_sep="",
  434. stim_dur=.01,
  435. stim_interval=.1,
  436. stim_pre=0,
  437. dt=.001,
  438. include=NULL,
  439. depends_on=NULL,
  440. as_function=NULL,
  441. start_values=NULL,
  442. fixed_pars=NULL,
  443. extra_condition=NULL,
  444. v_scale=1,
  445. bounds=NULL,
  446. urgency=NULL,
  447. objective="chisq",
  448. max_time=10,
  449. verbose=TRUE,
  450. ...){
  452. if (is.null(stim_var)) {
  453. stop("must provide \"stim_var\" argument, referencing the data column that contains the stimulus train.")
  454. }
  457. # set default parameter values
  458. all_pars = c("v", "a", "t0",
  459. "s", "z", "dc",
  460. "sv", "sz", "st0",
  461. "lambda", "aprime", "kappa", "tc",
  462. "uslope", "udelay", "umag")
  463. values = c(1, 2, .3,
  464. 1, .5, 0,
  465. 0, 0, 0,
  466. 0, 0.5, 1, .25,
  467. 0, 0, 0)
  468. lower = c(-100, .1, 1e-10,
  469. 1e-10, .05, -100,
  470. 0, 0, 0,
  471. 0, 0, 0, 1e-10,
  472. 0, 0, 0)
  473. upper = c(100, 25, 5,
  474. 100, .95, 100,
  475. 1, 1, 1,
  476. 100, 1, 5, 5,
  477. 10, 10, 10)
  478. default_pars = c("v", "a", "t0")
  479. start_if_include = c(sv=0.1,
  480. sz=0.1,
  481. st0=0.1,
  482. lambda=1,
  483. uslope=1,
  484. umag=1,
  485. udelay=1)
  487. super$initialize(dat,
  488. model_name,
  489. par_names=all_pars,
  490. par_values=values,
  491. par_lower=lower,
  492. par_upper=upper,
  493. default_pars=default_pars,
  494. start_if_include=start_if_include,
  495. include=include,
  496. depends_on=depends_on,
  497. as_function=as_function,
  498. start_values=start_values,
  499. fixed_pars=fixed_pars,
  500. max_time=max_time,
  501. extra_condition=extra_condition,
  502. bounds=bounds,
  503. urgency=urgency,
  504. verbose=verbose,
  505. ...)
  507. private$v_scale = v_scale
  509. self$set_objective(objective)
  511. private$stim_list = list()
  512. for(i in 1:private$par_transform[, .N]) {
  513. dsub = copy(self$data)
  514. for (c in private$sim_cond) {
  515. dsub = dsub[get(c) == private$par_transform[i, get(c)]]
  516. }
  518. up_stims = dsub[, get(stim_var[1])]
  519. if (length(stim_var) > 1) {
  520. down_stims = dsub[, get(stim_var[2])]
  521. } else {
  522. down_stims = NULL
  523. }
  525. private$stim_list[[i]] = pulse_stimulus(up_stims,
  526. down_stims,
  527. pattern=stim_sep,
  528. dur=stim_dur,
  529. isi=stim_interval,
  530. pre_stim=stim_pre,
  531. dt=dt,
  532. as_array=T)
  534. }
  536. }
  539. #' Pulse diffusion model R6 Class
  540. #'
  541. #' @description
  542. #'
  543. #' R6 Class that defines a pulse diffusion model (Brunton et al., 2012, Science) to be applied to a set of behavioral data.
  544. #'
  545. #' @details
  546. #'
  547. #' For details regarding other methods, see:
  548. #' \itemize{
  549. #' \item{pm$set_objective: \code{\link{set_pdm_objective}}}
  550. #' \item{pm$fit: \code{\link{fit_diffusion_model}}}
  551. #' \item{pm$predict: \code{\link{predict_pulse_model}}}
  552. #' \item{pm$simulate: \code{\link{simulate_pulse_model}}}
  553. #' }
  554. #'
  555. #' @usage pm <- pulse_model$new(dat, stim_var, stim_sep="", stim_dur=.01, stim_interval=.1, model_name="pdm", include=NULL, depends_on=NULL, as_function=NULL, start_values=NULL, fixed_pars=NULL, extra_condition=NULL, bounds=0L, objective="fp")
  556. #' @usage pm$set_objective(objective="fp")
  557. #' @usage pm$fit(use_bounds=TRUE, transform_pars=FALSE, ...)
  558. #' @usage pm$predict(pars=NULL, n=10000, ...)
  559. #' @usage pm$simulate(n, stimuli, par_values, par_names=NULL, ...)
  560. #'
  561. #' @param dat data table; contains at least 2 columns: rt - response time for trial, response - upper or lower boundary (1 or 0)
  562. #' @param stim_var string; name of stimulus column in dat. If two names, first is upper boundary evidence and second is lower boundary evidence. If one variable name, assumes binary evidence with 1=evidence to upper boundary and 0=evidence to lower boundary
  563. #' @param stim_sep string; pattern separating timepoint by timepoint evidence within evidence string
  564. #' @param stim_dur numeric; stimulus pulse duration (in seconds)
  565. #' @param stim_interval numeric; inter-stimulus interval (in seconds)
  566. #' @param stim_pre numeric; time before stimulus pulse within stimulus bin
  567. #' @param dt numeric; time step to simulate/solve for first passage times
  568. #' @param model_name string; name to identify model, default = "ddm"
  569. #' @param include character; vector of parameters to include in model. drift rate v, boundary a, and non-decision time t0 are included by default always. Can specify ddm parameters starting point z, drift rate variability sv, non decision time variability st0, starting point variability sz, and wiener diffusion noise s. Also can include collapsing bound parameters: degree of collapse aprime (weibull only), slope of collapse kappa, time constant of collapse tc.
  570. #' @param depends_on named character; if a parameter value depends on a task condition, for example drift rate depends on task difficulty, specify here: c("v" = "task difficulty"). Can specify multiple dependent parameters as a vector
  571. #' @param as_function list; list specifying parameters whose value is a function of other parameters
  572. #' @param start_values named numeric; to change default starting parameter values, specify as named vector. E.g. c("v"=2, "a"=3) to set drift rate to 2 and boundary to 3
  573. #' @param fixed_pars named numeric; to fix a parameter at a specified value, use a named vector as with start_values
  574. #' @param extra_condition character; vector of task condition names. Will calculate first passage times for each condition. Recommended only when comparing a model without depends_on with a model that contains a depends_on parameter.
  575. #' @param bounds string: either "fixed" for fixed bounds, or "weibull" or "hyperbolic" for collapsing bounds according to weibull or hyperbolic ratio functions
  576. #' @param objective character; "fp" for fokker-planck simulation, "chisq" for euler simluaton with chisq metric, "qmpe" for euler simultion with qmpe metric
  577. #' @param verbose logical: If TRUE, print messages related to model initation. Default = TRUE
  579. #'
  580. #' @return definition of pulse diffusion model object
  581. #'
  582. #' @export
  583. pulse_model = R6::R6Class("pulse_model",
  584. inherit=base_diffusion_model,
  585. public=list(
  586. initialize=init_pulse_model,
  587. set_objective=set_pdm_objective,
  588. predict=predict_pulse_model,
  589. simulate=simulate_pulse_model,
  590. get_stimulus=function() return(private$stim_list)
  591. ), private=list(
  592. dt=NULL,
  593. as_function=NULL,
  594. stim_list=NULL,
  595. v_scale=NULL,
  596. check_par_constraints=check_pdm_constraints,
  597. fp_obj = pulse_fp_obj,
  598. chisq_obj = pulse_x2_obj,
  599. qmpe_obj = pulse_qmpe_obj
  600. ))

pulse_diffusion_model.R at commit 57e9c3b, under GPL-3.0 · at the source

Overview

Authors: Hongjie Xia1, Maxime Maheu2,3, Gary A. Kane4, Benjamin B. Scott3,4
ORCID iDs: Hongjie Xia, Benjamin B. Scott
  1. Department of Biology, Boston University,Boston, MA USA
  2. Département d’Études Cognitives, École Normale Supérieure, Université PSL,Paris, France
  3. Department of Psychological and Brain Sciences, Boston University,Boston, MA USA
  4. Center for Systems Neuroscience, Boston University,Boston, MA USA
Institutions: Boston University (United States); Université Paris Sciences et Lettres (France)
Journal: Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology, volume 51, issue 9, pages 1680-1689
Dates: received 30 October 2025; accepted 19 March 2026; published online 13 April 2026; in print August 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1038/s41386-026-02399-x · PMID 41974997 · PMCID PMC13389323 · OpenAlex W4414513604
Open access: hybrid, a free copy (OpenAlex)
Preprint: doi:10.1101/2025.09.23.678165
Status: code verified
Categories: mouse (organism), cognitive (subfield)
Methods: Statistics, Machine learning, fMRI & imaging
Keywords: Decision, Perception
MeSH: Decision Making*, Locus Coeruleus*, Norepinephrine*, Adrenergic alpha-2 Receptor Agonists, Animals, Clonidine, Male, Mice, Receptors, Adrenergic, alpha-2 (* major topic)
Topic: Receptor Mechanisms and Signaling (Molecular Biology, Biochemistry, Genetics and Molecular Biology), according to OpenAlex
Funding: H2020 Marie Skłodowska‐Curie Actions (Postdoctoral fellowship); National Institute of Mental Health (R56MH132732)
Citations: not cited yet (Europe PMC); 80 references in the paper

Abstract

A fundamental challenge for decision-making under uncertainty lies in balancing speed and accuracy. Humans and animals solve this problem by adjusting decision thresholds—the criterion that determines how much information is required before committing to a choice. While brain regions associated with this process have been identified, the neural circuits that directly alter decision thresholds remain unknown. Here, we investigate the role of the locus coeruleus (LC) norepinephrine (NE) system in controlling this balance. Through cell-type-specific chemogenetic manipulations, we discovered that LC-NE activation increased decision thresholds. This effect is replicated by administration of the α2-adrenergic receptor (α2-AR) agonist clonidine. Notably, α2-AR activation altered decision threshold specifically, without reproducing other LC-NE activation effects such as promoting task engagement. Together, these results suggest that LC-NE regulates decision thresholds, possibly via downstream α2-ARs.

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 2 matches between paragraphs and lines of code.

gkane26/rddm

License: GPL-3.0
State: the link answers, verified on 29 September 2026
Evidence: files inventoried
Commit: 57e9c3bbaa10038b4af7d6b9e4675f5ec1ef17d0, 5 May 2022
Languages: R (14), C++ (11), C/C++ (6)
Size: 77 files, 31 scripts
Software Heritage: archived
Found in: the text, “Drift diffusion model”
Holds: README, license file, environment (DESCRIPTION), documentation, 1 notebook
Not found: CITATION.cff, tests, continuous integration
Tools: data.table (2 files), tidyverse (2 files)
Availability: 1 check, the latest on 29 September 2026: the link answers
  • 29 September 2026: the link answers
33 files
At the source: github.com/gkane26/rddm

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;
  • 31 scripts, each with its path and the digest of its content;
  • 2 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

All data will be provided upon request.

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, 29 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 4 authors, 2 keywords, 9 MeSH terms, 2 funders, 79 references.

Cite

This paper

Xia, H., Maheu, M., Kane, G. A., & Scott, B. B. (2026). Regulation of the decision threshold by the locus coeruleus. Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology, 51(9), 1680-1689. https://doi.org/10.1038/s41386-026-02399-x

BibTeX

@article{xia2026regulation,
author = {Xia, Hongjie and Maheu, Maxime and Kane, Gary A. and Scott, Benjamin B.},
title = {{Regulation of the decision threshold by the locus coeruleus}},
journal = {Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology},
year = {2026},
month = apr,
volume = {51},
number = {9},
pages = {1680--1689},
publisher = {Nature Publishing Group},
issn = {0893-133X},
doi = {10.1038/s41386-026-02399-x},
url = {https://doi.org/10.1038/s41386-026-02399-x},
pmid = {41974997},
pmcid = {PMC13389323}
}

RIS

TY - JOUR
AU - Xia, Hongjie
AU - Maheu, Maxime
AU - Kane, Gary A.
AU - Scott, Benjamin B.
TI - Regulation of the decision threshold by the locus coeruleus
T2 - Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology
J2 - Neuropsychopharmacology
PY - 2026
DA - 2026/04/13
VL - 51
IS - 9
SP - 1680
EP - 1689
SN - 0893-133X
PB - Nature Publishing Group
DO - 10.1038/s41386-026-02399-x
UR - https://doi.org/10.1038/s41386-026-02399-x
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41386-026-02399-x",
"type": "article-journal",
"title": "Regulation of the decision threshold by the locus coeruleus",
"container-title": "Neuropsychopharmacology : official publication of the American College of Neuropsychopharmacology",
"author": [
{
"family": "Xia",
"given": "Hongjie"
},
{
"family": "Maheu",
"given": "Maxime"
},
{
"family": "Kane",
"given": "Gary A."
},
{
"family": "Scott",
"given": "Benjamin B."
}
],
"container-title-short": "Neuropsychopharmacology",
"volume": "51",
"issue": "9",
"page": "1680-1689",
"DOI": "10.1038/s41386-026-02399-x",
"PMID": "41974997",
"PMCID": "PMC13389323",
"ISSN": "0893-133X",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41386-026-02399-x",
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
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:10.1038/s41562-026-02533-1 [code]
Fluctuations in arousal reflect latent state transitions that facilitate behavioural optimization.
Journal: Nature human behaviour
In common: cognitive, 6 references
[2] doi:10.1126/sciadv.aec9291 [code]
Computational mechanisms of perception in autism revealed using games inspired by rodent operant tasks.
Journal: Science advances
In common: data.table, tidyverse, cognitive, 1 reference, author Benjamin B Scott
[3] doi:10.1126/sciadv.adz6495
Pupil-linked arousal heterogeneously modulates cell-type-specific sensory processing.
Journal: Science advances
In common: mouse, 6 references
[4] doi:10.1371/journal.pone.0355165 [code]
Pupillary dynamics during hands-off L2 driving and transitions of control under high cognitive load.
Journal: PloS one
In common: data.table, tidyverse, cognitive, 4 references
[5] doi:10.1186/s40478-026-02287-x [code]
Cellular signatures of melanocortin pathway genes across the locus coeruleus.
Journal: Acta neuropathologica communications
In common: tidyverse, mouse, 4 references
[6] doi:10.7554/elife.110294 [code]
Arousal modulates functional connectivity through structured and hemispherically asymmetric community architecture during wakefulness.
Journal: eLife
In common: 5 references
[7] doi:10.1016/j.isci.2026.117110
Orbitofrontal cortex contributes to context-dependent timing under ambiguity.
Journal: iScience
In common: cognitive, 4 references
[8] doi:10.1073/pnas.2536535123 [code]
Orbitofrontal noradrenaline supports adaptive learning-rate adjustment in probabilistic reversal learning.
Journal: Proceedings of the National Academy of Sciences of the United States of America
In common: cognitive, 4 references
[9] doi:10.7554/elife.110685 [code]
Sensory adaptation and pupil-linked arousal support flexible evidence accumulation during perceptual decision making.
Journal: eLife
In common: cognitive, 4 references
[10] doi:10.1162/imag.a.1200 [code]
An fMRI examination of the role of the Locus Coeruleus in state regulation in ADHD.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: 4 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.