Fidgety Philip and the Suggested Clinical Immobilization Test: Annotation data for developing a machine learning...

Area of Science:

• Neuroscience
• Psychiatry
• Computer Science

Background:

• The Fidgety Philip cartoon illustrates hyperactive behaviors associated with ADHD diagnosis.
• Existing diagnostic methods have gaps when compared to modern computational approaches.
• Objective quantification of hypermotor-restlessness (H-behaviors) during sitting is challenging.

Purpose of the Study:

• To develop a neutral, pictogram-guided phenotyping language (PG-PL) for describing body-segment movements during sitting.
• To establish a reliable method for manual annotation of human movement patterns.
• To create a foundation for a machine learning algorithm for automated movement detection.

Main Methods:

• Seven research assistants annotated Fidgety Philip cartoons and snapshots using freehand and PG-PL methods.
• Inter-observer agreement was assessed to ensure reliability of the PG-PL.
• PG-PL annotations of video clips were used to develop a machine learning algorithm for automated movement detection.

Main Results:

• The pictogram-guided phenotyping language (PG-PL) demonstrated value for manual annotation of human movement patterns.
• Satisfactory inter-observer agreements were achieved, validating the PG-PL's consistency.
• A machine learning algorithm for automated movement detection was successfully developed using PG-PL annotations.

Conclusions:

• The PG-PL provides a standardized, objective method for describing restless movements.
• This approach bridges the gap between clinical observation and computational analysis in ADHD research.
• The developed methodology and algorithm can be further refined for differentiating movement patterns.