Intentional Binding Task

Field	Value
Name	Intentional Binding Task
Version	v0.1.0
URL / Repository	https://github.com/TaskBeacon/T000085-intentional-binding-task
Short Description	Conventional Libet-clock task measuring perceived temporal attraction between voluntary actions and their auditory outcomes.
Created By	TaskBeacon
Date Updated	2026-08-01
PsyFlow Version	0.1.12
PsychoPy Version	2025.2.4
Modality	Behavior
Language	Chinese
Voice Name	None (voice disabled)

1. Task Overview

This task implements the conventional four-condition Intentional Binding paradigm. Participants watch a clock hand rotating once every 2.5 s and report the perceived position of either a self-paced key press or a 1000 Hz tone. Comparing signed report errors between single-event baselines and action-causes-tone trials yields action binding, tone binding, and total binding estimates. Because temporal binding is not uniquely diagnostic of agency, the exported measures should be interpreted as action-outcome temporal attraction rather than a stand-alone proof of subjective agency.

2. Task Flow

Block-Level Flow

Stage	Implementation
Session setup	Load the selected mode config, collect or inject subject_id, initialize PsyFlow, the trigger runtime, window, keyboard, and preloaded stimulus bank.
General instruction	Show instruction; explain the rotating clock, event-specific report, free action timing, and pointer report. Continue with space.
Block order	Shuffle baseline_action, baseline_tone, operant_action, and operant_tone deterministically from overall_seed and participant ID.
Block instruction	Show one config-defined condition instruction that states whether an action and/or tone will occur and whether action or tone time must be reported.
Practice	Run 6 practice trials for the current condition, then show practice_complete. QA and simulation profiles omit practice.
Scored block	Run 30 scored trials for the current condition using built-in BlockUnit.generate_conditions(...) with a single condition label.
Completion	Save one reduced row per logical trial and show completed/valid counts plus mean absolute report error.

Trial-Level Flow

Phase	Visible/Auditory Content	Timing and Response
warning	warning: centered Chinese new-trial message	1.0 s; condition-specific onset trigger
fixation	fixation: central +	0.5 s
clock_stationary	60 ticks, labels 0, 5, ..., 55, center dot, and a red hand at one of 12 start positions	0.5 s
clock_running - action trials	Same clock; hand rotates clockwise once per 2.5 s	Space is accepted for up to 6.0 s; the action ends baseline-action trials or starts the operant delay
clock_running - tone baseline	Same rotating clock without valid keys	Tone onset is sampled in 0.2-2.0 s from a deterministic trial seed
action_tone_delay	Rotating clock continues	0.2 s after an operant action
tone	Rotating clock plus PsychoPy-generated 1000 Hz tone	0.1 s
post_event_clock	Rotating clock continues without feedback	Continues until 2.0 s has elapsed from the event being judged
clock_report	Static clock face plus report_action or report_tone	Click one of 60 clock positions within 15 s
iti	blank white screen	0.5 s

Controller Logic

Component	Logic
Adaptive controller	None. The task contains no staircase, reward update, or performance adaptation.
Condition scheduling	Each block passes its single canonical condition through PsyFlow's built-in generator; no custom condition generator is used.
Nuisance sampling	src/utils.py deterministically samples the 12-position clock start and baseline-tone latency from overall_seed plus PsyFlow trial_id.
Dynamic clock	ClockHand.draw() derives the current position from a PsychoPy clock; StimUnit retains ownership of drawing frames, flips, response capture, triggers, and logging.

Other Logic

Measure	Definition
Clock unit	2500 / 60 = 41.667 ms
Signed judgment error	Shortest circular displacement from actual event position to reported position, wrapped to [-30, 30) and converted to milliseconds
Action binding	Mean operant_action error minus mean baseline_action error
Tone binding	Mean baseline_tone error minus mean operant_tone error
Total binding	Action binding plus tone binding
Valid trial	Required action occurred (when applicable) and one clock-position report was collected

3. Configuration Summary

Settings below come from config/config.yaml.

a. Subject Info

Field	Meaning
subject_id	Three-digit participant identifier from 101 to 999

b. Window Settings

Parameter	Value
Size	1280 x 800
Units	Degrees of visual angle
Background	White
Fullscreen	False
Monitor geometry	35.5 cm width, 57 cm viewing distance

c. Stimuli

Name	Type	Description
instruction	Text	General Chinese task instructions
block_*	Text	Four condition-specific Chinese instruction screens
warning	Text	New-trial warning
fixation	Text	Central plus sign
clock_rim, clock_center, clock_label	Circle/Text	Config-defined anchors for the generated 60-position clock
Generated ticks/hand	Line	60 radial ticks and a red clockwise rotating hand
outcome_tone	Sound	Generated 1000 Hz sine tone asset, 100 ms
report_action, report_tone	Text	Event-specific report prompts above the clock
Report targets	Circle	60 invisible pointer targets centered on clock positions 0-59
blank	Text	Empty ITI stimulus

d. Timing

Phase	Duration
Warning	1.0 s
Fixation	0.5 s
Stationary clock	0.5 s
Clock period	2.5 s per revolution
Action response ceiling	6.0 s
Baseline-tone latency	Uniform 0.2-2.0 s
Action-to-tone delay	0.2 s
Tone	0.1 s
Clock after judged event	2.0 s total from critical event
Clock report ceiling	15.0 s
ITI	0.5 s

e. Triggers

Event	Code
Experiment start / end	1 / 99
Block start / end	10 / 90
Condition warnings	20-23
Fixation / stationary clock	30 / 31
Condition clock onsets	40-43
Action / action timeout	50 / 54
Tone / post-event clock	51 / 52
Report onset / selection / timeout / complete	60 / 61 / 62 / 63
ITI	70

4. Methods (for academic publication)

Participants completed four blocked conditions presented in participant-seeded random order: baseline action, baseline tone, operant action judgment, and operant tone judgment. Each condition began with six practice trials and contained 30 scored trials. A trial started with a 1.0 s warning and 0.5 s fixation. A 60-position Libet clock then appeared with its hand stationary for 0.5 s at one of 12 five-unit start positions. The hand subsequently rotated clockwise with a 2.5 s period. In action-containing conditions, participants pressed the space bar at a freely chosen time within a 6.0 s ceiling. In baseline-tone trials, no action was made and the computer presented the tone 0.2-2.0 s after rotation onset.

In operant trials, the voluntary action caused a 1000 Hz, 100 ms tone after 200 ms. The clock continued until approximately 2.0 s had elapsed from the event designated for judgment. Participants then clicked the perceived event position on a stationary clock labeled from 0 to 59. Baseline blocks contained only the judged event: action without tone or computer-generated tone without action. No accuracy feedback or reward was provided. Clock start position and baseline-tone latency were deterministically sampled per trial, while trial identity, phase context, response capture, and triggers were managed by PsyFlow.

For each valid trial, the report error was the shortest signed circular displacement between the reported and actual event positions multiplied by 41.667 ms per clock unit. Action binding is estimated by subtracting baseline-action error from operant-action error; tone binding is estimated by subtracting operant-tone error from baseline-tone error; total binding is their sum. The procedure follows the conventional Libet-clock timing and blocked judgments described by Barlas and Obhi (2013), with operant/baseline interpretation informed by Dewey and Knoblich (2014). Inferences about subjective agency should remain cautious because comparable temporal binding can arise without intentional action, as shown by Suzuki et al. (2019).

Run

python main.py human
python main.py qa --config config/config_qa.yaml
python main.py sim --config config/config_scripted_sim.yaml
python main.py sim --config config/config_sampler_sim.yaml