Automated Operation Span Task

Field	Value
Name	Automated Operation Span Task
Version	0.1.0
Date Updated	2026-07-12
PsyFlow Version	current local runtime
PsychoPy Version	current local runtime
Modality	Behavioral
Language	English

1. Task Overview

Automated OSPAN measures complex working-memory capacity by interleaving arithmetic verification with serial letter storage. The main task contains 15 randomized sets: set sizes 3-7 appear three times each, totaling 75 letters and operations.

2. Task Flow

Block-Level Flow

Instruction -> math practice/deadline calibration -> three combined size-2 practice sets -> 15 scored sets -> summary.

Trial-Level Flow

Within each set, solve operation -> judge proposed answer -> view letter for 800 ms, repeated by set size, then recall letters in serial order and view 2000-ms feedback.

Controller Logic

Math practice response times set the subsequent verification deadline to mean + 2.5 SD, bounded from 1 to 10 seconds.

Other Logic

Absolute span sums perfectly recalled set sizes; total correct counts letters recalled in the correct serial position. Math accuracy below 85% invalidates interpretation.

3. Configuration Summary

a. Subject Info

Field	Constraint
Participant ID	Three digits

b. Window Settings

Setting	Value
Resolution	1280 x 800
Units	Degrees

c. Stimuli

Stage	Content
Operation	Simple integer multiplication plus/minus term
Verification	Proposed result, F false/J true
Storage	One of F H J K L N P Q R S T Y
Recall	Two-row letter matrix and serial position

d. Timing

Stage	Duration
Letter	800 ms
Feedback	2000 ms
Recall	30 s ceiling per letter

Triggers

Operation, verification, letter, recall, feedback, response, timeout, and boundaries have dedicated codes.

4. Methods (for academic publication)

Participants completed an automated operation-span task adapted from Unsworth et al. (2005). After arithmetic practice established an individualized verification deadline, participants completed three set-size-2 combined practice sets and 15 scored sets. Scored set sizes 3 through 7 each occurred three times in randomized order. Each item required solving an operation, judging a proposed answer with F=False or J=True, and retaining a letter shown for 800 ms. At set end, letters were recalled in serial order. Absolute span and position-wise total correct were computed, with scores flagged invalid below 85% arithmetic accuracy.

Running

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