{
  "id": 8314,
  "title": "No ictal segment in patient 4 where latency>15s",
  "url": "/competitions/seizure-detection/discussion/8314",
  "author_name": "",
  "post_date": "2014-05-29T14:13:22.423Z",
  "votes": null,
  "comment_count": 1,
  "views": 842,
  "content": "<p>is this intentional that no ictal segment for patient 4 has been provided where latency&gt;15s i.e., no not-early ictal trial. It would have been helful if we had a few. Only 20 ictal trials have been rpovided where latency&lt;15s. Any ideas?</p>",
  "messages": [
    {
      "id": "47260",
      "postDate": "05/29/2014 14:13:22",
      "content": "<p>is this intentional that no ictal segment for patient 4 has been provided where latency&gt;15s i.e., no not-early ictal trial. It would have been helful if we had a few. Only 20 ictal trials have been rpovided where latency&lt;15s. Any ideas?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "47288",
      "postDate": "05/29/2014 21:32:10",
      "content": "<p>There is wide ranging variability in the durations and distribution of seizures between subjects, and there can be variability in seizure durations within a subject as well. Seizures were chosen randomly for training and test data, so having no seizures greater than 15 seconds in duration in a subject's training data is certainly possible.</p>\n<p>Having a limited number of seizures to train on is usually the clinical reality in these situations, and we designed the contest to reflect that.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 47288,
      "author_name": "bbrinkm",
      "author_url": "",
      "post_date": "05/29/2014 21:32:10",
      "content": "<p>There is wide ranging variability in the durations and distribution of seizures between subjects, and there can be variability in seizure durations within a subject as well. Seizures were chosen randomly for training and test data, so having no seizures greater than 15 seconds in duration in a subject's training data is certainly possible.</p>\n<p>Having a limited number of seizures to train on is usually the clinical reality in these situations, and we designed the contest to reflect that.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "47260": "",
    "47288": ""
  },
  "source": "meta"
}