Attention over vulnerable brain regions associating cerebral palsy disorder and biological markers

Muhammad Hassan1, Jieqiong Lin1, Ahmed Ameen Fateh1

  • 1Department of Radiology, Shenzhen Children's Hospital, Shenzhen, Guangdong, China.

PubMed

Insights

This study reveals how biological traits like age, gender, and weight impact brain development in children with cerebral palsy (CP) and healthy controls. Weight estimation showed the most significant differences between the two groups.

Area of Science:

  • Neuroscience
  • Developmental Biology
  • Medical Imaging

Background:

  • Cerebral palsy (CP) is a neurological disorder stemming from fetal brain ischemia and hypoxia.
  • Monitoring early brain development in children with CP is crucial for effective intervention.
  • Understanding the impact of biological traits (BTs) on brain development in both healthy controls (HC) and CP-affected children is essential.

Purpose of the Study:

  • To associate biological traits (BTs) with neurodevelopment in healthy controls (HC) and children with cerebral palsy (CP).
  • To investigate the underlying neurodevelopmental differences associated with BTs in HC and CP.
  • To establish a benchmark for the association of BTs with brain development in HC and CP.

Main Methods:

  • Utilized AWG-Net with customized spatial-channel (CSC) and multi-head self (MHA) attentions for analyzing brain development.
  • Incorporated cumulative-dense structures to propagate susceptible brain regions to deeper layers.
  • Trained the model using T1-w, T2-w, Flair, and Sag MRI sequences, annotated with age, gender, and weight.

Main Results:

  • Age, gender, and weight significantly correlate with brain development in both HC and CP groups, with weight showing distinct patterns.
  • Identified specific vulnerable brain regions associated with age and gender, noting commonalities and discrepancies between HC and CP.
  • AWG-Net demonstrated differential learning trends for HC and CP concerning brain development and gender, with slight variations for weight.

Conclusions:

  • Weight estimation is more affected by CP-vulnerable brain regions than age and gender.
  • T1-weighted MRI is suitable for assessing age, weight, and gender in neurodevelopmental studies.
  • Learned trends in brain development and gender association differ between healthy controls and children with cerebral palsy.
Abstract