# Exp01 SSVEP-YOLO Target Selection

> Generated from repository source files. Do not edit this page directly; edit the source file listed below and rerun `vp run docs:generate`.

> Source: `experiments/exp01-ssvep-yolo-target-selection/protocol.md`

## Experiment 1: SSVEP-YOLO Dynamic Target Selection

Status: `pre-data protocol draft`

## Goal

Test whether YOLO-generated candidate boxes can serve as reliable SSVEP targets.

## Conditions

1. Fixed UI button SSVEP baseline.
2. Static scene candidate boxes.
3. Recorded YOLO video candidate boxes.
4. Real-time YOLO candidate boxes.

## Variables

* Object count: 2, 4, 6.
* Box size: small, medium, large.
* Box spacing: far, near, partially overlapping.
* EEG window: 0.5s, 1.0s, 1.5s, 2.0s.
* Channels: O1, Oz, O2, POz, Pz and subsets.

## Metrics

* Accuracy.
* ITR.
* Response time.
* False selection rate.
* Candidate jitter.
* Target lost rate.
* Selection correction count.
* Fatigue and NASA-TLX.

## Analysis Plan

See `../analysis-plan.md`. Complete `../pre-data-freeze-checklist.md` and a session record before collecting evidence-producing trials.

Required pre-session records:

* `session_id`;
* `hardware_inventory_id`;
* `threshold_freeze_record_id`;
* protocol version and software commit;
* display timing settings and stimulus rendering method;
* EEG montage, preprocessing, and SSVEP decoder ID;
* detector/replay source ID for recorded or live YOLO conditions;
* randomization seed/file and block order.

Trial inclusion rule:

* Include every attempted trial in the log.
* Mark trials as `completed`, `timed_out`, `rejected`, `safety_blocked`, or `aborted`.
* Do not remove target-loss, candidate jitter, low-confidence, or wrong-selection trials unless the exclusion rule is frozen before the session.

Primary endpoint:

* correct accepted target selection by visual condition and object count.

Baseline interpretation:

* fixed UI button SSVEP establishes the controlled selection reference;
* static scene candidate boxes isolate object-box layout;
* recorded YOLO candidate boxes add realistic visual dynamics under replay control;
* real-time YOLO candidate boxes test online scene-aware target generation.

Advancement rule:

* real-time YOLO candidate boxes should enter closed-loop grasping only after pre-set accuracy, target-loss, and jitter criteria are met. Numeric thresholds are `needs confirmation`.
* progression decisions must cite threshold IDs from `experiments/templates/threshold-freeze-record.md`.

## Evidence Output

Results update `library/EVIDENCE_MATRIX.md` rows for SSVEP role and dynamic command space.

***

> Source: `experiments/exp01-ssvep-yolo-target-selection/metrics.md`

## Metrics

| Metric | Definition | Status |
| --- | --- | --- |
| accuracy | correct accepted target selections / attempted trials, with denominator reported | definition draft frozen |
| ITR | standard SSVEP ITR from `../analysis-plan.md`; report raw-window and practical-selection variants | definition draft frozen; final `T` values need session binding |
| response time | stimulus onset or candidate freeze start to accepted decision, reported by condition | definition draft frozen |
| false selection rate | wrong accepted selections / attempted trials | definition draft frozen |
| target lost rate | trials where target candidate is unavailable during selection window | definition draft frozen |
| candidate jitter | movement of candidate box during selection window using pre-frozen pixel or visual-angle definition | definition draft frozen; units need display binding |
| workload | NASA-TLX or equivalent linked by questionnaire ID | protocol choice needs confirmation |

***

> Source: `experiments/exp01-ssvep-yolo-target-selection/variables.md`

## Variables

| Factor | Levels | Notes |
| --- | --- | --- |
| object count | 2 / 4 / 6 | match candidate capacity to SSVEP code set |
| box size | small / medium / large | define in visual angle before data collection |
| spacing | far / near / overlap | quantify pixel and visual-angle distance |
| window length | 0.5 / 1.0 / 1.5 / 2.0 seconds | evaluate speed-accuracy tradeoff |
| channel subset | O1 / Oz / O2 / POz / Pz | expand or reduce after pilot |
| visual condition | fixed UI / static box / recorded YOLO / live YOLO | isolates dynamic detection effects |

## Freeze Requirements

* Exact pixel sizes, visual angles, spacing definitions, and viewing distance must be frozen in the session record.
* Recorded YOLO conditions must reference the replay file, detector version, and candidate-track source.
* Live YOLO conditions must reference detector weights, input size, confidence threshold, tracking rule, and freeze policy.
* Channel subsets must map to exact electrode labels in `experiments/hardware-inventory.md`.