EEfRT Task

Field	Value
Name	EEfRT Task
Version	v0.2.1-dev
URL / Repository	https://github.com/TaskBeacon/T000019-eefrt
Short Description	Effort Expenditure for Rewards Task with probabilistic reward outcomes.
Created By	TaskBeacon
Date Updated	2026-02-24
PsyFlow Version	0.1.9
PsychoPy Version	2025.1.1
Modality	Behavior
Language	Chinese
Voice Name	zh-CN-YunyangNeural

1. Task Overview

This task implements an EEfRT-style paradigm in which participants choose between low-effort/low-reward and high-effort/high-reward options under varying reward probabilities. After choosing, participants perform an effort execution stage and then receive completion and reward feedback. The task captures effort allocation, completion rate, and reward-sensitive decision behavior.

2. Task Flow

Block-Level Flow

Step	Description
1. Parse mode and config	main.py loads human, qa, or sim runtime mode and the selected YAML config.
2. Initialize runtime	Window, keyboard, trigger runtime, and stimulus bank are initialized.
3. Prepare offers	For each block, a custom condition generator builds deterministic (probability, hard_reward, fallback, reward-draw) trial specs.
4. Execute trials	Trials run through src/run_trial.py with full context logging.
5. Block summary	Block metrics are shown (high-effort rate, completion rate, block reward).
6. Final summary	Final task metrics are shown and trial data are saved.

Trial-Level Flow

Step	Description
Offer fixation	Central fixation before offer display.
Offer choice	Participant chooses low vs high effort option with response deadline.
Ready stage	The chosen effort requirement and deadline are shown.
Effort execution	Participant presses effort key repeatedly within the deadline.
Effort feedback	Completion or failure message is presented.
Reward feedback	Probabilistic reward outcome is presented (or no reward if not completed).
Inter-trial interval	Fixation interval before the next trial.

Controller Logic

Component	Description
Adaptive controller	Not used. EEfRT does not require dynamic RT-window adjustment.
Offer condition generation	src/utils.py generates deterministic block trial specs (offer grid + fallback choice + reward draw sample).
Runtime scoring	Trial outcomes and rewards are computed in src/run_trial.py; summaries are aggregated from trial data in main.py.

Runtime Context Phases

Phase Label	Meaning
offer_fixation	Fixation before the choice screen.
offer_choice	Choice screen for low vs high effort options.
effort_execution_window	Effort performance response window.
effort_feedback	Effort completion feedback stage.
inter_trial_interval	ITI fixation stage.

3. Configuration Summary

a. Subject Info

Field	Meaning
subject_id	Participant identifier.

b. Window Settings

Parameter	Value
size	[1280, 720]
units	pix
screen	0
bg_color	black
fullscreen	false
monitor_width_cm	35.5
monitor_distance_cm	60

c. Stimuli

Name	Type	Description
fixation	text	Central fixation cross.
choice_header / choice_left / choice_right	text	Offer and choice information panel.
ready_text	text	Pre-effort confirmation prompt.
effort_prompt / effort_counter	text	Effort instruction and live progress display.
effort_success_feedback / effort_fail_feedback	text	Completion feedback.
reward_win_feedback / reward_nowin_feedback / reward_incomplete_feedback	text	Reward outcome feedback.
block_break / good_bye	text	Block and final summaries.

d. Timing

Phase	Duration
cue	1.0 s
choice	5.0 s
ready	1.0 s
effort deadline (easy)	7.0 s
effort deadline (hard)	21.0 s
feedback	1.0 s
reward_feedback	1.0 s
iti	1.0 s

4. Methods (for academic publication)

Participants complete an EEfRT-style effort-based choice task. On each trial, reward probability and low/high effort options are presented. Participants choose an option and then perform the selected effort requirement using repeated keypresses within a time limit.

If the effort criterion is met, reward outcome is sampled according to trial probability. If the criterion is not met, reward is set to zero. Trial-level records include offer parameters, selected option, effort completion status, reaction timing, and reward outcome.

The implementation supports human, qa, and sim modes with consistent trial logic and responder-context instrumentation for reproducibility and auditability.

This refactor removes the generic task controller object and replaces it with explicit condition-generation utilities so the EEfRT offer logic and stochastic outcomes are easier to audit.