File size: 1,884 Bytes
e41c1f7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
{
    "meta_info": {
        "case_id": "MentalArithmetic_01.edf",
        "bench_subset": "NeuroBench-Core",
        "difficult": 1.0,
        "original_dataset": "MentalArithmetic"
    },
    "agent_input": {
        "data_path": "data/core/MentalArithmetic_01.edf",
        "instruction": "Please first extract only the EEG channels from the raw signal, and then apply a 0.5-40Hz FIR bandpass filter to these channels. After filtering, obtain each channel's Alpha-band power and total filtered-signal power by integrating the PSD over frequency. Calculate Alpha relative power separately for each channel, then average the channel-wise ratios and report the final percentage value clearly in your response."
    },
    "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 the 'Alpha Relative Power Percentage' from the agent's report.\n\n### STRICT CONSTRAINTS (MANDATORY)\n1. Output ONLY a valid JSON object.\n2. DO NOT include Markdown code blocks (e.g., no backticks or json tags).\n3. DO NOT include any conversational text, explanations, or units (%).\n4. The Key in the JSON must be EXACTLY \"alpha_ratio\".\n5. The Value must be a FLOAT (e.g., if the report says 25.5%, output 25.5).\n6. If the value is missing or cannot be found, return {\"alpha_ratio\": null}.\n\n### OUTPUT TEMPLATE\n{\"alpha_ratio\": <float>}",
        "metrics": [
            {
                "metric_id": "alpha_ratio_accuracy",
                "type": "numeric_check",
                "target_key": "alpha_ratio",
                "weight": 100,
                "params": {
                    "gt_value": 41.33018074506845,
                    "tolerance": 4.133018074506846
                }
            }
        ]
    }
}