Effect of Training on Visual Identification of High Frequency Oscillations-A Delphi-Style Intervention

Area of Science:

• Neuroscience
• Epileptology
• Biostatistics

Background:

• Accurate identification of high-frequency oscillations (HFOs) is crucial for understanding epilepsy.
• Interrater reliability in HFO detection can be variable, impacting clinical application.
• Standardized training methods are needed to improve consistency in HFO evaluations.

Purpose of the Study:

• To evaluate the impact of a Delphi-method feedback intervention on visual HFO identification.
• To assess changes in interrater reliability and generalizability of HFO evaluations post-intervention.
• To determine the effectiveness of this training for epileptologists with varying experience levels.

Main Methods:

• A morphology detector identified potential HFOs (80-250 Hz ripple band) for visual review.
• 19 epileptologists participated in three evaluation sessions with a Delphi-style feedback intervention.
• A delayed-intervention paradigm was used; generalizability theory assessed interrater reliability.

Main Results:

• The Delphi intervention significantly reduced inter-reviewer variability (p=0.037, p=0.003).
• Early intervention group showed significant variability reduction (p=0.041); delayed group did not (p=0.414).
• Projected reviewers needed for strong generalizability decreased from 35 to 16 post-intervention.

Conclusions:

• Delphi-style feedback robustly enhances interrater reliability and generalizability of HFO evaluations.
• The intervention improved consistency, making reviewer reliability comparable to experienced specialists.
• This method facilitates HFO evaluation standardization for improved clinical care.