TIC-XNet: An Interpretable Framework for Enhanced Detection of Pediatric Tic Events Through Structured Evidence and Improved Temporal Alignment

By
Liping Li
Jianping Wang
Kunying Zhou
Qian Li
Haoyu Wu
Huimin Song
Xiaoxia Fang
June 23, 2026

Frontiers In Psychiatry

Overview

The TIC-XNet framework demonstrates superior performance in detecting tic events in children with tic disorders by integrating video and physiological signals. It achieves high accuracy and interpretable outputs.

Background

Pediatric tic disorders are neurodevelopmental conditions that present challenges in clinical evaluation due to their variable expression across contexts. Current assessment methods rely on subjective observations, highlighting the need for objective, continuous monitoring approaches.

Data Highlights

Metric	TIC-XNet	Comparator Models
Window-level AUC	0.915 ± 0.019	Lower
Event-level recall	Higher	Lower
Event-level precision	Higher	Lower
Missed events	Fewer	More
Post-buffering prediction latency	Lower	Higher

Key Findings

TIC-XNet achieved a window-level AUC of 0.915 ± 0.019.
It demonstrated higher event-level recall and precision compared to comparator models.
The framework resulted in fewer missed events and lower post-buffering prediction latency.
Translated outputs from TIC-XNet showed higher decision fidelity and stability under perturbation.
Subject-level translated signals were associated with tic severity.

Clinical Implications

The TIC-XNet framework provides a method for detecting tic events, integrating multimodal data.

Conclusion

TIC-XNet represents an advancement in the detection of tic events in children, combining predictive performance with interpretable evidence outputs.

TIC-XNet: An Interpretable Framework for Enhanced Detection of Pediatric Tic Events Through Structured Evidence and Improved Temporal Alignment

Clinical Report: TIC-XNet for Enhanced Detection of Pediatric Tic Events

Overview

Background

Data Highlights

Key Findings

Clinical Implications

Conclusion

Related Resources & Content

Original Source(s)

TIC-XNet: a structured evidence translation framework for interpretable multimodal pediatric tic event detection with improved temporal alignment and fidelity

Related Content

PTSD May Be Linked to Accelerated Aging

Higher admission serum total carbon dioxide is independently associated with early seizure recurrence in children with benign convulsions with mild gastroenteritis

Aid in Dying Nears One Third of US