Tagging EEG features within exam reports to quickly generate databases for research purposes.

Area of Science:

• Neurology
• Medical Informatics

Background:

• Existing electroencephalogram (EEG) classification systems can be cumbersome.
• A novel, tag-based approach aims to streamline EEG report analysis.

Purpose of the Study:

• To assess the effectiveness of a new tag-based method for classifying EEG recording features.
• To determine the prevalence of specific EEG features in patients with toxic-metabolic encephalopathy.

Main Methods:

• EEG reports were tagged during creation with features like background activity, epileptiform discharges, and periodic discharges.
• A PHP script parsed tagged reports, and a spreadsheet analysis calculated feature prevalence.
• The study focused on patients with toxic-metabolic encephalopathy.

Main Results:

• Tagging and parsing were time-efficient, integrated into daily workflows.
• Out of 5784 EEG recordings, 218 (3.8%) were tagged for toxic-metabolic aetiology.
• Severe slowing (5-6 Hz) was the most frequent background feature (36.2%), while epileptiform abnormalities were rare (4.6%). Triphasic waves were present in 19.7% of recordings.

Conclusions:

• The tag-and-parse method is rapid and easily integrated into clinical practice.
• EEG slowing is a prevalent feature in toxic-metabolic encephalopathies, unlike triphasic waves.
• The free, open-source EEG Report Parser offers an adaptable alternative to existing software, minimizing costs and training.