Prediction error correlates in the striosome-dopamine circuit emerge from information gain.
17 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 17 matches
- [1] § Methods › Simulation illustrating how phasic and tonic dopamine components differentially drive D1 versus D2 pathway activity ↔ Circuit_design/d1_d2_lead_to_dual_param_control.py, lines 54 to 138 · score 0.94 · D2 pathway activity, receptor activity, phasic peak, arbitrary units, D1, concentration · block
oscr:b:RNCVFYJ7WYVE8-L· Report a mismatch - [2] § Methods › Analyzing the relationship between policy-IG and informativeness as predictors of Pavlovian acquisition using Harris and Gallistel (2025) data ↔ Harris_and_Gallistel_analysis/Harris_Gallistel_analysis.m, lines 23 to 101 · score 0.88 · permanent threshold, sustained threshold, cumulative nDKL, harris gallistel, crossings, odds · block
oscr:b:NQXXK91XPMMQP-8· Report a mismatch - [3] § Methods › Simulation testing the importance of different cortical “modules” to actor/critic/policy-IG learning ↔ Circuit_design/agents.py, lines 17 to 55 · score 0.78 · conflict resolution, environmental stimuli, motor programs, Actor, Critic, agent · block
oscr:b:RVACNCVHPCFTH-X· Report a mismatch - [4] § Methods › Simulation showing the effect of dopamine→striosome connection weight on policy-IG control ↔ Circuit_design/striosome_dopamine_control.py, lines 72 to 81 · score 0.76 · weak inhibition, strong inhibition, dopamine activity, striosome dopamine, excitatory · block
oscr:b:WA9XFSFWNAKYP-8· Report a mismatch - [5] § Methods › Policy-IG/RPE relationship in Gymnasium tasks ↔ Associative_learning_correlations/open_gym_comparison.py, lines 34 to 35 · score 0.75 · Taxi v3, CliffWalking, FrozenLake, v1 · block
oscr:b:7CBWKZM53HFQC-Q· Report a mismatch - [6] § Methods › Simulation exploring the space of possible action-boundary profiles ↔ Action_initiation/clusters_of_brackets.py, lines 154 to 226 · score 0.74 · noise multiplier, normalized policy, profiles, classified, gap, bracketing · block
oscr:b:VTPG9XDZ8A0XW-9· Report a mismatch - [7] § Methods › Simulations comparing TD and policy-IG learning algorithms ↔ Multi_cue_learning/chunking_vs_TD_learning_blocking_task.py, lines 81 to 211 · score 0.73 · delayed start, multi cue, compound, epochs, licking, trace · block
oscr:b:3EEBWCGY88SRX-1· Report a mismatch - [8] § Methods › Simulations showing relationship between Weber law surprise and policy-IG ↔ Weber_law_surprise/Webers_law_and_policyIG.py, lines 23 to 80 · score 0.72 · Weber law surprise, est, log2, interval, exponential, simulated · block
oscr:b:1TY9CHV51F87Z-R· Report a mismatch - [9] § Methods › Simulation testing the importance of different cortical “modules” to actor/critic/policy-IG learning ↔ Circuit_design/layers.py, lines 185 to 234 · score 0.71 · conflict strength, conflict resolution, confidences, preferences, uncertainty, errors · block
oscr:b:BHDMQA54KA06X-9· Report a mismatch - [10] § Methods › Reinforcement learning simulations showing a linear RPE/policy-IG relationship ↔ Associative_learning_correlations/associative_learning_RPE_pIG_correlation.py, lines 33 to 69 · score 0.68 · q_cue, learning_rate, Shannon entropy, simulated, softmax, RPE · block
oscr:b:3EFRNKMNEHH5Q-Q· Report a mismatch - [11] § Methods › Analyzing the relationship between policy-IG and dopamine responses using Eshel et al. (2016) data ↔ Associative_learning_correlations/eshel_et_al_fitting.py, lines 33 to 76 · score 0.64 · p_cue, offset, power, firing, sigmoid, unexpected · block
oscr:b:RF8Q3AEG6PCAS-3· Report a mismatch - [12] § Methods › Analyzing the relationship between policy-IG and dopamine responses to novel stimuli using Kutlu et al. (2021) data ↔ Associative_learning_correlations/experimental_data_fit_Kutlu_et_al_novelty.py, lines 67 to 94 · score 0.63 · trial bin, dopamine response, decline, neutral, stimuli, fit · block
oscr:b:KVTB5BZW07SB8-B· Report a mismatch - [13] § Methods › Simulation illustrating the emergence of policy-IG correlates ↔ Naturalistic_foraging/naturalistic_rodent_foraging.py, lines 317 to 395 · score 0.61 · causal learning, eligibility, traces, Rolling, timestep, foraging · block
oscr:b:T0CW1A6SN1WTC-N· Report a mismatch - [14] § Methods › Simulation illustrating different heterogenous policy-IG reactions to the same stimuli ↔ Circuit_design/four_stream_policies_over_time.py, lines 31 to 101 · score 0.61 · streams, drift, Limbic, Oculomotor, strength, pulses · block
oscr:b:87NKQGWWYAGYY-D· Report a mismatch - [15] § Methods › Analyzing the relationship between policy-IG and dopamine responses using Stauffer et al. (2014) data ↔ Associative_learning_correlations/experimental_data_fit_Stauffer_et_al.py, lines 12 to 54 · score 0.56 · interpolate utility, spline, intercept, slope, fitted, nats · block
oscr:b:MVRZR3JEJ2A50-T· Report a mismatch - [16] § Results › Prediction: policy-IG learning resembles contingency and temporal difference learning ↔ Multi_cue_learning/chunking_vs_TD_learning_blocking_task.py, lines 216 to 286 · score 0.53 · TD algorithm, TD learning, stimulation, cue · block
oscr:b:ZBMPDD48AM39A-8· Report a mismatch - [17] § Methods › Policy-IG/RPE relationship during foraging with surprising costs ↔ Associative_learning_correlations/foraging_with_large_costs_RPE_pIG_relationship.py, lines 86 to 148 · score 0.51 · learning agent, surprise, arbitrarily, foraging, spaced, episodes · block
oscr:b:D4WH4JA7BN396-3· 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
Python · 286 lines · 13 KB · no license · 2 matches
- import numpy as np
- import matplotlib.pyplot as plt
- np.random.seed(0)
- class Policy:
- def __init__(self, p_lick_target_init=0.5):
- self.p_lick_target = p_lick_target_init
- self.p_lick_current = 0.5
- self._update_probabilities()
- def _update_probabilities(self):
- self.p_lick = np.clip(self.p_lick_current, 1e-9, 1 - 1e-9)
- self.p_no_lick = 1.0 - self.p_lick
- self.actions = ["Lick", "No Lick"]
- self.probabilities = np.array([self.p_lick, self.p_no_lick])
- self.entropy = self._calculate_entropy()
- def _calculate_entropy(self):
- if np.allclose(self.p_lick, 0) or np.allclose(self.p_lick, 1):
- return 0.0
- if self.p_lick <= 1e-9 or self.p_no_lick <= 1e-9:
- return 0.0
- return -(self.p_lick * np.log(self.p_lick + 1e-9) + self.p_no_lick * np.log(self.p_no_lick + 1e-9))
- def update_learning(self, learning_rate):
- self.p_lick_current += learning_rate * (self.p_lick_target - self.p_lick_current)
- self._update_probabilities()
- def get_current_p_lick_for_display(self):
- return self.p_lick_current
- class Chunk:
- def __init__(self, name, cues, p_lick_target_for_policy=0.5):
- self.name = name
- self.cues = cues
- self.policy = Policy(p_lick_target_init=p_lick_target_for_policy)
- self.h_max = np.log(len(self.policy.actions)) if len(self.policy.actions) > 0 else 0
- self.q_intrinsic = 0
- self.is_learning_active = False
- self.update_q_intrinsic()
- def update_q_intrinsic(self):
- self.q_intrinsic = self.h_max - self.policy.entropy
- def learn_policy_step(self, learning_rate):
- if self.is_learning_active:
- self.policy.update_learning(learning_rate)
- self.update_q_intrinsic()
- # params
- H_MAX_GLOBAL = np.log(2)
- EPISODES_PER_EPOCH = 40
- TOTAL_EPOCHS = 4
- POLICY_LEARNING_RATE_CHUNKER = 0.15
- ALPHA_Q = 0.15
- GAMMA_Q = 0.9
- TEMPERATURE_Q = 1.0 # softmax temperature
- STATE_CHEESE_NO_CUE = 0
- STATE_RED_BELL_ONLY = 1
- STATE_BLUE_BELL_ONLY = 2
- STATE_RED_BLUE_COMPOUND = 3
- NUM_STATES_Q = 4
- ACTION_LICK_Q = 0
- ACTION_NO_LICK_Q = 1
- NUM_ACTIONS_Q = 2
- def choose_action(q_values, temperature):
- # subtract max for numerical stability
- z = q_values / temperature
- z = z - np.max(z)
- exp_z = np.exp(z)
- probs = exp_z / np.sum(exp_z)
- return np.random.choice(NUM_ACTIONS_Q, p=probs)
- def simulate_delayed_start_plot():
- chunk_dictionary = {}
- total_sim_episodes = TOTAL_EPOCHS * EPISODES_PER_EPOCH
- policy_baseline_chunk = Policy(p_lick_target_init=0.5)
- chunker_traces = {
- "CheeseOnly": [np.nan] * total_sim_episodes,
- "Red+cheese": [np.nan] * total_sim_episodes,
- "Red+blue+cheese": [np.nan] * total_sim_episodes
- }
- q_table = np.zeros((NUM_STATES_Q, NUM_ACTIONS_Q))
- qlearner_traces = {
- "V(S_Cheese)": [np.nan] * total_sim_episodes,
- "V(S_Red)": [np.nan] * total_sim_episodes,
- "V(S_BlueOnly)": [np.nan] * total_sim_episodes,
- "V(S_RedBlueCompound)": [np.nan] * total_sim_episodes
- }
- P_LICK_TARGET_UNIVERSAL = 0.95
- P_LICK_TARGET_CHEESE_ONLY = P_LICK_TARGET_UNIVERSAL
- P_LICK_TARGET_RED = P_LICK_TARGET_UNIVERSAL
- P_LICK_TARGET_RED_BLUE = P_LICK_TARGET_UNIVERSAL
- stim_epoch = 4
- stim_duration_episodes = int(EPISODES_PER_EPOCH * 0.5)
- stim_start_episode_global_idx = (stim_epoch - 1) * EPISODES_PER_EPOCH
- stim_end_episode_global_idx = stim_start_episode_global_idx + stim_duration_episodes
- global_episode_idx = 0
- for epoch in range(1, TOTAL_EPOCHS + 1):
- # chunker learning activation
- if epoch > 1 and "CheeseOnly" in chunk_dictionary:
- chunk_dictionary["CheeseOnly"].is_learning_active = False
- if epoch > 2 and "Red+cheese" in chunk_dictionary:
- chunk_dictionary["Red+cheese"].is_learning_active = False
- if "Red+blue+cheese" in chunk_dictionary:
- chunk_dictionary["Red+blue+cheese"].is_learning_active = False
- if epoch == 1:
- if "CheeseOnly" not in chunk_dictionary:
- chunk_dictionary["CheeseOnly"] = Chunk("CheeseOnly", {"Cheese"}, P_LICK_TARGET_CHEESE_ONLY)
- chunk_dictionary["CheeseOnly"].is_learning_active = True
- elif epoch == 2:
- if "Red+cheese" not in chunk_dictionary:
- chunk_dictionary["Red+cheese"] = Chunk("Red+cheese", {"RedBell"}, P_LICK_TARGET_RED)
- chunk_dictionary["Red+cheese"].is_learning_active = True
- elif epoch == 3:
- if "Red+blue+cheese" not in chunk_dictionary:
- chunk_dictionary["Red+blue+cheese"] = Chunk("Red+blue+cheese", {"RedBell", "BlueBell"},
- P_LICK_TARGET_RED_BLUE)
- for episode_in_epoch in range(EPISODES_PER_EPOCH):
- # chunker logic
- if epoch == stim_epoch and stim_start_episode_global_idx <= global_episode_idx < stim_end_episode_global_idx and "Red+blue+cheese" in chunk_dictionary:
- chunk_dictionary["Red+blue+cheese"].is_learning_active = True
- elif "Red+blue+cheese" in chunk_dictionary and global_episode_idx >= stim_end_episode_global_idx:
- chunk_dictionary["Red+blue+cheese"].is_learning_active = False
- for chunk_obj in chunk_dictionary.values():
- chunk_obj.learn_policy_step(POLICY_LEARNING_RATE_CHUNKER)
- # record chunker traces
- if epoch >= 1 and "CheeseOnly" in chunk_dictionary:
- chunker_traces["CheeseOnly"][global_episode_idx] = chunk_dictionary["CheeseOnly"].q_intrinsic
- if epoch >= 2 and "Red+cheese" in chunk_dictionary:
- chunker_traces["Red+cheese"][global_episode_idx] = chunk_dictionary["Red+cheese"].q_intrinsic
- if epoch >= 3 and "Red+blue+cheese" in chunk_dictionary:
- chunker_traces["Red+blue+cheese"][global_episode_idx] = chunk_dictionary["Red+blue+cheese"].q_intrinsic
- # q-learner logic
- q_learner_current_trial_states = []
- enable_blue_to_red_propagation = False
- if epoch == 1:
- q_learner_current_trial_states.append(STATE_CHEESE_NO_CUE)
- elif epoch == 2:
- q_learner_current_trial_states.append(STATE_RED_BELL_ONLY)
- elif epoch == 3:
- rand_e3 = np.random.rand()
- if rand_e3 < 0.5:
- q_learner_current_trial_states.append(STATE_BLUE_BELL_ONLY)
- else:
- q_learner_current_trial_states.append(STATE_RED_BLUE_COMPOUND)
- elif epoch == 4:
- q_learner_current_trial_states.extend(
- [STATE_BLUE_BELL_ONLY, STATE_RED_BLUE_COMPOUND, STATE_RED_BELL_ONLY])
- if global_episode_idx >= stim_start_episode_global_idx:
- enable_blue_to_red_propagation = True
- if q_learner_current_trial_states:
- current_q_s = np.random.choice(q_learner_current_trial_states)
- action_q = choose_action(q_table[current_q_s], TEMPERATURE_Q)
- reward_q = 0
- next_s_max_q = 0
- if current_q_s == STATE_BLUE_BELL_ONLY and enable_blue_to_red_propagation:
- s_blue = STATE_BLUE_BELL_ONLY
- a_blue = action_q
- r_blue = 0
- s_next_after_blue = STATE_RED_BELL_ONLY
- val_of_s_next_after_blue = np.max(q_table[s_next_after_blue])
- q_table[s_blue, a_blue] += ALPHA_Q * (
- r_blue + GAMMA_Q * val_of_s_next_after_blue - q_table[s_blue, a_blue])
- elif current_q_s == STATE_BLUE_BELL_ONLY and not enable_blue_to_red_propagation:
- r_blue_only = 0
- q_table[current_q_s, action_q] += ALPHA_Q * (
- r_blue_only + GAMMA_Q * 0 - q_table[current_q_s, action_q])
- else:
- if current_q_s == STATE_CHEESE_NO_CUE and action_q == ACTION_LICK_Q:
- reward_q = 1.0
- elif current_q_s == STATE_RED_BELL_ONLY and action_q == ACTION_LICK_Q:
- reward_q = 1.0
- elif current_q_s == STATE_RED_BLUE_COMPOUND and action_q == ACTION_LICK_Q:
- reward_q = 1.0
- q_table[current_q_s, action_q] += ALPHA_Q * (
- reward_q + GAMMA_Q * next_s_max_q - q_table[current_q_s, action_q])
- # record Q traces
- if epoch >= 1:
- qlearner_traces["V(S_Cheese)"][global_episode_idx] = np.max(q_table[STATE_CHEESE_NO_CUE])
- if epoch >= 2:
- qlearner_traces["V(S_Red)"][global_episode_idx] = np.max(q_table[STATE_RED_BELL_ONLY])
- if epoch >= 3:
- qlearner_traces["V(S_BlueOnly)"][global_episode_idx] = np.max(q_table[STATE_BLUE_BELL_ONLY])
- qlearner_traces["V(S_RedBlueCompound)"][global_episode_idx] = np.max(q_table[STATE_RED_BLUE_COMPOUND])
- global_episode_idx += 1
- return chunker_traces, qlearner_traces, q_table
- chunker_res_delay, qlearner_res_delay, final_q_delay = simulate_delayed_start_plot()
- # plotting
- total_eps_delay = TOTAL_EPOCHS * EPISODES_PER_EPOCH
- eps_numbers_delay = np.arange(1, total_eps_delay + 1)
- fig, ax1 = plt.subplots(figsize=(10, 6))
- colors = {'cheese': 'green', 'red': 'red', 'blue': 'blue', 'purple': 'purple'}
- ax1.plot(eps_numbers_delay, chunker_res_delay["CheeseOnly"], color=colors['cheese'], linestyle='-', linewidth=2,
- label="Cheese chunk")
- ax1.plot(eps_numbers_delay, chunker_res_delay["Red+cheese"], color=colors['red'], linestyle='-', linewidth=2,
- label="Red+cheese chunk")
- ax1.plot(eps_numbers_delay, chunker_res_delay["Red+blue+cheese"], color=colors['purple'], linestyle='-', linewidth=2,
- label="Red+blue+cheese chunk")
- ax1.plot(eps_numbers_delay, qlearner_res_delay["V(S_Cheese)"], color=colors['cheese'], linestyle='--', linewidth=2,
- alpha=0.7, label="Cheese value (TD algorithm)")
- ax1.plot(eps_numbers_delay, qlearner_res_delay["V(S_Red)"], color=colors['red'], linestyle='--', linewidth=2, alpha=0.7,
- label="Red cue value (TD algorithm)")
- ax1.plot(eps_numbers_delay, qlearner_res_delay["V(S_BlueOnly)"], color=colors['blue'], linestyle='--', linewidth=2,
- alpha=0.7, label="Blue cue value (TD algorithm)")
- ax1.set_xlabel("Episode Number", fontsize=10)
- ax1.set_ylabel('Value (chunk or cue)', color='black', fontsize=10)
- ax1.tick_params(axis='y', labelsize=9)
- ax1.tick_params(axis='x', labelsize=9)
- ax1.set_ylim(-0.05, 1.10)
- epoch_labels_delay = ["E1: Learn Cheese", "E2: Learn Red", "E3: Blue Intro", "E4: Stim. + Consolidate"]
- stim_epoch_plot_delay = 4
- stim_duration_plot_delay = int(EPISODES_PER_EPOCH * 0.5)
- stim_band_plot_start_delay = (stim_epoch_plot_delay - 1) * EPISODES_PER_EPOCH + 0.5
- stim_band_plot_end_delay = stim_band_plot_start_delay + stim_duration_plot_delay
- for i in range(1, TOTAL_EPOCHS):
- xc = i * EPISODES_PER_EPOCH
- ax1.axvline(x=xc + 0.5, color='dimgray', linestyle=':', linewidth=1.0)
- ax1.axvspan(stim_band_plot_start_delay, stim_band_plot_end_delay, color='yellow', alpha=0.3)
- tick_pos_delay = [i * EPISODES_PER_EPOCH + EPISODES_PER_EPOCH / 2 for i in range(TOTAL_EPOCHS)]
- valid_tick_pos_delay = [tp for tp in tick_pos_delay if tp <= total_eps_delay]
- valid_tick_labels_delay = [epoch_labels_delay[i] for i, tp in enumerate(tick_pos_delay) if tp <= total_eps_delay]
- ax1.set_xticks(valid_tick_pos_delay)
- ax1.set_xticklabels(valid_tick_labels_delay, rotation=10, ha="right", fontsize=8)
- import matplotlib.patches as mpatches
- handles, labels = ax1.get_legend_handles_labels()
- stim_patch = mpatches.Patch(color='yellow', alpha=0.3, label='Artificial stimulation')
- if 'Artificial stimulation' not in labels:
- handles.append(stim_patch)
- labels.append('Artificial stimulation')
- desired_order_map = {
- "Cheese chunk": 0, "Red+cheese chunk": 1, "Red+blue+cheese chunk": 2,
- "Cheese value (TD algorithm)": 3, "Red cue value (TD algorithm)": 4,
- "Blue cue value (TD algorithm)": 5, "Artificial stimulation": 6
- }
- sorted_handles_labels = sorted(zip(handles, labels), key=lambda x: desired_order_map.get(x[1], 99))
- if sorted_handles_labels:
- sorted_handles, sorted_labels = zip(*sorted_handles_labels)
- ax1.legend(sorted_handles, sorted_labels, fontsize=8, loc='center left', bbox_to_anchor=(1.02, 0.5))
- else:
- ax1.legend(fontsize=8, loc='center left', bbox_to_anchor=(1.02, 0.5))
- ax1.grid(False)
- fig.tight_layout(rect=[0, 0, 0.75, 1])
- plt.rcParams['pdf.fonttype'] = 42
- plt.savefig('chunking_vs_TD_learning_blocking_task.pdf')
- plt.show()
chunking_vs_TD_learning_blocking_task.py at commit 1b65f2b, no license · at the source
Overview
- Computational Science Program, University of Texas at El Paso, El Paso, TX USA
- Department of Biological Sciences, University of Texas at El Paso, El Paso, TX USA
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, with 17 matches between paragraphs and lines of code.
dirkbeck/dopamine_modeling
1b65f2be6ecdaf5aca5000cacffee0c062b74119, 22 February 2026Availability: 1 check, the latest on 27 September 2026: the link answers
- 27 September 2026: the link answers
42 files
- Action_initiation/
clusters_of_brackets.py , Python, 354 lines, 1 match - Action_initiation/
hierarchical_melody_note , Python, 200 lines_primative_simulation.py - Action_initiation/
policy_shift_action_init , Python, 120 linesiation.py - Action_initiation/
sensivity_by_baseline_po , Python, 70 lineslicy_IG.py - Amo_et_al_analysis/
Amo_et_al_analysis.m , MATLAB, 284 lines - Associative_learning_cor
relations/ , Python, 254 lines, 1 matchassociative_learning_RPE _pIG_correlation.py - Associative_learning_cor
relations/ , Python, 263 linesassociative_learning_RPE _pIG_large_cost_and_nove lty.py - Associative_learning_cor
relations/ , Python, 76 lines, 1 matcheshel_et_al_fitting.py - Associative_learning_cor
relations/ , Python, 77 linesexperimental_data_fit_Fi orillo_et_al.py - Associative_learning_cor
relations/ , Python, 94 lines, 1 matchexperimental_data_fit_Ku tlu_et_al_novelty.py - Associative_learning_cor
relations/ , Python, 116 linesexperimental_data_fit_Ku tlu_et_al_utility.py - Associative_learning_cor
relations/ , Python, 62 lines, 1 matchexperimental_data_fit_St auffer_et_al.py - Associative_learning_cor
relations/ , Python, 187 lines, 1 matchforaging_with_large_cost s_RPE_pIG_relationship.p y - Associative_learning_cor
relations/ , Python, 67 linesnovelty_colormap.py - Associative_learning_cor
relations/ , Python, 112 lines, 1 matchopen_gym_comparison.py - Associative_learning_cor
relations/ , Python, 97 linestheoretical_2D_curve.py - Associative_learning_cor
relations/ , Python, 156 linestheoretical_3D_curve.py - Circuit_design/
ACG_simulations_main_exe , Python, 90 linescution.py - Circuit_design/
agents.py , Python, 444 lines, 1 match - Circuit_design/
d1_d2_both_necessary.py , Python, 249 lines - Circuit_design/
d1_d2_lead_to_dual_param , Python, 138 lines, 1 match_control.py - Circuit_design/
distribution_across_neur , Python, 88 linesons.py - Circuit_design/
environment.py , Python, 52 lines - Circuit_design/
four_stream_policies_ove , Python, 101 lines, 1 matchr_time.py - Circuit_design/
layers.py , Python, 304 lines, 1 match - Circuit_design/
plotting_functions.py , Python, 205 lines - Circuit_design/
striosome_dopamine_contr , Python, 94 lines, 1 matchol.py - Decision_making_biases/
Ushapeperformance.py , Python, 148 lines - Decision_making_biases/
decision_making_manifold , Python, 121 lines.py - Decision_making_biases/
risk_impulsivity_and_con , Python, 148 linesflict.py - Disorders/
conceptual_disorder_poli , Python, 42 linescy_IG_axis.py - Disorders/
coupled_infoIG_RL_sim.py , Python, 193 lines - Harris_and_Gallistel_ana
lysis/ , MATLAB, 546 lines, 1 matchHarris_Gallistel_analysi s.m - Harris_and_Gallistel_ana
lysis/ , Python, 257 linesharris_gallistel_analysi s.py - Multi_cue_learning/
chunking_vs_TD_learning_ , Python, 286 lines, 2 matchesblocking_task.py - Multi_cue_learning/
chunking_vs_TD_learning_ , Python, 118 linestwo_cue_task.py - Naturalistic_foraging/
naturalistic_rodent_fora , Python, 562 lines, 1 matchging.py - Ramping_and_state_value/
basic_ramping_and_telepo , Python, 83 linesrt.py - Ramping_and_state_value/
dopamine_and_state_value , Python, 221 lines_correlation.py - Ramping_and_state_value/
probabilistic_task_rampi , Python, 74 linesng.py - Weber_law_surprise/
Webers_law_and_policyIG. , Python, 161 lines, 1 matchpy - README.md, Text, 159 lines
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: dirkbeck/
dopamine_modeling
Read it in the paper: doi.org/10.1038/s41467-026-73994-1.
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;
- 41 scripts, each with its path and the digest of its content;
- 17 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
- doi:10.5061/
dryad.hhmgqnkjw , at Dryad; found in the text, “Testing for discrete plateau structure in…” - osf:vmwzr, at OSF; found in the text, “Analyzing the relationship between policy-IG…”
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/s41467-026-73994-1.
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, 2 authors, 5 keywords, 9 MeSH terms, 2 funders, 184 references.
Cite
This paper
Beck, D. W., & Friedman, A. (2026). Prediction error correlates in the striosome-dopamine circuit emerge from information gain. Nature communications, 17(1), 8066. https://
BibTeX
@article{beck2026predict
author = {Beck, Dirk W and Friedman, Alexander},
title = {{Prediction error correlates in the striosome-dopamine circuit emerge from information gain}},
journal = {Nature communications},
year = {2026},
month = jun,
volume = {17},
number = {1},
pages = {8066},
publisher = {Nature Publishing Group},
issn = {2041-1723},
doi = {10.1038/
url = {https://
pmid = {42364984},
pmcid = {PMC13454626}
}
RIS
TY - JOUR
AU - Beck, Dirk W
AU - Friedman, Alexander
TI - Prediction error correlates in the striosome-dopamine circuit emerge from information gain
T2 - Nature communications
J2 - Nat Commun
PY - 2026
DA - 2026/
VL - 17
IS - 1
SP - 8066
SN - 2041-1723
PB - Nature Publishing Group
DO - 10.1038/
UR - https://
LA - en
ER -
CSL-JSON
{
"id": "10.1038/
"type": "article-journal",
"title": "Prediction error correlates in the striosome-dopamine circuit emerge from information gain",
"container-title": "Nature communications",
"author": [
{
"family": "Beck",
"given": "Dirk W"
},
{
"family": "Friedman",
"given": "Alexander"
}
],
"container-title-short":
"volume": "17",
"issue": "1",
"page": "8066",
"DOI": "10.1038/
"PMID": "42364984",
"PMCID": "PMC13454626",
"ISSN": "2041-1723",
"publisher": "Nature Publishing Group",
"URL": "https://
"language": "en",
"issued": {
"date-parts": [
[
2026,
6,
27
]
]
}
}
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:10.1523/jneurosci.1527-22.2023
- Anticipation of Appetitive Operant Action Induces Sustained Dopamine Release in the Nucleus Accumbens.Journal: The Journal of neuroscience : the official journal of the Society for NeuroscienceIn common: cognitive, 27 references
- [2] doi:10.1038/s41467-025-61466-x [code]
- A decision-space model explains context-specific decision-making.Journal: Nature communicationsIn common: Statistics and Machine Learning Toolbox, SciPy, Matplotlib, 1 other tool, cognitive, 18 references, author Alexander Friedman
- [3] doi:10.1371/journal.pcbi.1010340 [code]
- The dopamine circuit as a reward-taxis navigation system.Journal: PLoS computational biologyIn common: pandas, SciPy, Matplotlib, 1 other tool, 22 references
- [4] doi:10.1038/s41586-022-05611-2 [code]
- Spontaneous behaviour is structured by reinforcement without explicit reward.Journal: NatureIn common: statsmodels, seaborn, pandas, 3 other tools, 20 references
- [5] doi:10.1126/sciadv.adn4203 [code]
- Mesostriatal dopamine is sensitive to changes in specific cue-reward contingencies.Journal: Science advancesIn common: Statistics and Machine Learning Toolbox, pandas, SciPy, 2 other tools, 18 references
- [6] doi:10.7554/elife.62583 [code]
- Slowly evolving dopaminergic activity modulates the moment-to-moment probability of reward-related self-timed movements.Journal: eLifeIn common: Statistics and Machine Learning Toolbox, 20 references
- [7] doi:10.1016/j.celrep.2024.114080
- Distributional coding of associative learning in discrete populations of midbrain dopamine neurons.Journal: Cell reportsIn common: cognitive, 21 references
- [8] doi:10.1038/s41593-025-02104-z [code]
- Estrogen modulates reward prediction errors and reinforcement learning.Journal: Nature neuroscienceIn common: Optimization Toolbox, Statistics and Machine Learning Toolbox, 18 references
- [9] doi:10.1016/j.cub.2024.09.045 [code]
- Pre-existing visual responses in a projection-defined dopamine population explain individual learning trajectories.Journal: Current biology : CBIn common: statsmodels, seaborn, pandas, 3 other tools, cognitive, 16 references
- [10] doi:10.1038/s41593-023-01401-9 [code]
- Unique functional responses differentially map onto genetic subtypes of dopamine neurons.Journal: Nature neuroscienceIn common: Statistics and Machine Learning Toolbox, 19 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.
Claim this paper
Correct its record
Say what each link of this record is, remove the ones that are not the paper's, add the ones that are missing. The correction becomes a new version of the record, in its Versions section.
Validate its tracing map
You validate the map as this page shows it: 1 repository of the authors' code, each at its verified commit and with its license, 41 scripts, and 17 matches between paragraphs and code (see the Code and Map sections). It then receives a DOI on Zenodo, with you (your ORCID iD) and OSCR as its creators; the code itself is not deposited.
The map's fingerprint: sha256:6864e97bd1d7b27b…
Add the badge to the code's README
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 -->
[](https://openscicode.org/badge/doi/10.1038/s41467-026-73994-1)
<!-- oscr-badge:end -->reStructuredText
.. oscr-badge:start
.. image:: https://openscicode.org/badge/doi/10.1038/s41467-026-73994-1.svg
:target: https://openscicode.org/badge/doi/10.1038/s41467-026-73994-1
:alt: OSCR
.. oscr-badge:endHTML
<!-- oscr-badge:start -->
<a href="https://openscicode.org/badge/doi/10.1038/s41467-026-73994-1"><img src="https://openscicode.org/badge/doi/10.1038/s41467-026-73994-1.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.
