Datasets:
File size: 4,177 Bytes
d846fa6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 | {
"meta_info": {
"case_id": "Mumtaz2016_05.edf",
"bench_subset": "NeuroBench-Core",
"difficult": 2.0,
"original_dataset": "Mumtaz2016"
},
"agent_input": {
"data_path": "data/core/Mumtaz2016_05.edf",
"instruction": "Please compute the alpha relative power ratio for each brain region, output the brain-region names in descending order of alpha relative power, determine whether posterior-dominant alpha exists, and explain your conclusion. Use Welch PSD estimation. Define relative power as the integrated power in the target frequency band divided by the integrated total power from 0.5 to 45 Hz. Compute relative power separately for each channel, and define each region's relative-power score as the arithmetic mean of the channel-wise relative-power values within that region."
},
"eval_config": {
"parser_prompt": "### ROLE\nYou are a precise JSON Extraction Engine for neurophysiological data. Your sole task is to convert natural language reports into structured JSON data.\n\n### TASK\nExtract two outputs from the agent report:\n1) region alpha ranking in descending order\n2) whether posterior-dominant alpha exists\n\n### STRICT CONSTRAINTS (MANDATORY)\n1. Output ONLY a valid JSON object.\n2. DO NOT include Markdown code blocks.\n3. DO NOT include any conversational text or explanations.\n4. Keys must be EXACTLY \"region_alpha_rank\" and \"posterior_dominant_alpha\".\n5. For \"region_alpha_rank\", read the final descending region ranking and keep only exact canonical regions from: frontal, central, parietal, occipital.\n6. Lowercase every retained region, remove duplicates, and preserve their relative order. Ignore unsupported labels such as temporal, mastoid, mixed, or compound labels such as right temporal/parietal; do not map or guess them.\n7. Example: [Central, Frontal, Occipital, Right Temporal/Parietal] -> [\"central\",\"frontal\",\"occipital\"].\n8. Map an explicit final YES/present/exists/true conclusion to true and an explicit NO/absent/does not exist/false conclusion to false. If statements conflict, use only the final conclusion.\n9. If no canonical ranked region is identifiable, set region_alpha_rank to null. If the posterior-dominance conclusion is unclear, set posterior_dominant_alpha to null. Do not guess.\n\n### OUTPUT TEMPLATE\n{\"region_alpha_rank\": [\"region1\",\"region2\"], \"posterior_dominant_alpha\": <bool|null>}",
"metrics": [
{
"metric_id": "region_alpha_rank_accuracy",
"type": "sequence_match_check",
"target_key": "region_alpha_rank",
"weight": 40,
"params": {
"gt_value": [
"occipital",
"parietal",
"central",
"frontal"
],
"match_mode": "exact_order"
}
},
{
"metric_id": "posterior_dominant_alpha_accuracy",
"type": "categorical_check",
"target_key": "posterior_dominant_alpha",
"weight": 30,
"params": {
"gt_value": true
}
},
{
"metric_id": "posterior_alpha_explanation_consistency",
"type": "semantic_consistency",
"target_key": "full_text",
"weight": 30,
"params": {
"judge_prompt": "Judge whether the explanation is logically consistent with the true region ranking and the posterior-dominant alpha label. The true top-ranked region is occipital, and posterior-dominant alpha is present. The explanation must make it clear that a posterior region (parietal or occipital) is strongest. If the explanation contradicts the true ranking or the posterior-dominant alpha label, mark it as incorrect. Output only a valid JSON object with exactly one key named status and a boolean value True or False."
}
}
]
}
}
|