Visual Search Task

Field	Value
Name	Visual Search Task
Version	v0.1.5-dev
URL / Repository	https://github.com/TaskBeacon/T000033-visual-search
Short Description	Feature vs conjunction visual search with present/absent decisions.
Created By	TaskBeacon
Date Updated	2026-03-18
PsyFlow Version	0.1.9
PsychoPy Version	2025.1.1
Modality	Behavior
Language	English
Voice Name	en-US-AriaNeural (voice disabled by default)

1. Task Overview

This task implements a classical visual-search paradigm grounded in feature-integration and guided-search literature. Participants search for a red T and report whether it is present (F) or absent (J).

The implementation explicitly separates feature-search and conjunction-search displays, with both target-present and target-absent trials, and logs trial-level RT/accuracy under PsyFlow human|qa|sim modes.

2. Task Flow

Block-Level Flow

Step	Description
1. Block setup	BlockUnit.generate_conditions(...) creates label-level condition sequence from task.conditions.
2. Trial execution	run_trial(...) runs fixation -> search array -> ITI per trial.
3. Block summary	Accuracy, mean correct RT, and timeout count are shown between blocks.
4. Final summary	Session-level accuracy/RT/timeout metrics are displayed.

Trial-Level Flow

Step	Description
fixation	Fixation cross (jittered duration) with target reminder text.
search_array	Circular array of letters appears; participant responds present/absent.
iti	Brief fixation-only inter-trial interval before next trial.

Controller Logic

Component	Description
Condition parser	Maps condition tokens to (search_type, target_present) factors.
Runtime realization	run_trial(...) resolves set size/positions/distractor composition from condition label plus seed context; duration ranges are passed to StimUnit.show()/capture_response() for framework-level sampling.
Trial ID resolver	Uses PsyFlow next_trial_id() at runtime for global trial IDs (no task-local Controller class).

Other Logic

Component	Description
Dynamic stimuli	Search items are generated per trial in src/run_trial.py via PsychoPy text primitives.
Trigger alignment	Phase/response/timeout triggers are emitted at fixation/search/ITI and key events.

3. Configuration Summary

a. Subject Info

Field	Meaning
subject_id	3-digit participant identifier.

b. Window Settings

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

c. Stimuli

Name	Type	Description
fixation	text	Central fixation cross used in fixation/ITI/search display.
search_goal	text	Constant top reminder: Find red T.
array_boundary	circle	Visual boundary for ring-based search layout.
dynamic search items	text	Trial-generated letters (T/L) with color/orientation/position factors.
instruction_text, block_break, good_bye	text	Instruction and summary screens.

d. Timing

Phase	Duration
fixation	sampled from fixation_duration range
search array response window	response_deadline
ITI	iti_duration

e. Triggers

Trigger	Code	Meaning
fixation_onset	20	fixation phase onset
search_onset	30	search display onset
response_present	31	present-key response
response_absent	32	absent-key response
search_timeout	33	no response before deadline
iti_onset	40	inter-trial interval onset

f. Adaptive Controller

Parameter Group	Purpose
feature_set_sizes, conjunction_set_sizes	Set-size pools for feature/conjunction trials.
array_radius_px, array_radius_jitter_px	Circular layout geometry.
target_glyph, conjunction_alt_glyph, color fields	Defines target and distractor identities.
Notes	No separate controller class; these parameters are consumed directly inside run_trial(...).

4. Methods (for academic publication)

Participants performed speeded present/absent judgments in a visual-search paradigm where the target was a red T. Trials manipulated search difficulty via feature versus conjunction distractor structure and target presence.

Each trial began with fixation, followed by a response-bounded search display with item positions sampled on a circular layout. Responses were logged with RT, correctness, and timeout flags. Session outcomes were summarized by accuracy, mean correct RT, and timeout frequency.

The implementation follows a literature-first mapping to feature-integration and guided-search frameworks and avoids MID-style cue/anticipation/feedback state templates.