Artificial Intelligence and the detection of pediatric concussion using epigenomic analysis
Ray O Bahado-Singh1, Sangeetha Vishweswaraiah1, Anıl Er2
1Department of Obstetrics and Gynecology, Oakland University William Beaumont School of Medicine, Royal Oak, MI, USA.
Brain Research
|October 20, 2019
Summary
Mild traumatic brain injury (mTBI), or concussion, involves significant DNA methylation changes in pediatric patients. Artificial intelligence accurately predicts concussion using these epigenetic markers and clinical data.
Area of Science:
- Neuroscience
- Genetics
- Biomarker Discovery
Background:
- Concussion (mild traumatic brain injury) is a common injury requiring better understanding of biological mechanisms.
- Development of reliable biomarkers for concussion diagnosis is an ongoing scientific challenge.
Purpose of the Study:
- To investigate whole genome-wide DNA methylation patterns in pediatric concussion.
- To identify epigenetic biomarkers for mild traumatic brain injury (mTBI) detection.
- To apply Artificial Intelligence (AI) for concussion prediction using epigenetic and clinical data.
Main Methods:
- Whole genome-wide blood DNA cytosine (CpG) methylation was analyzed using the Illumina Infinium MethylationEPIC assay in pediatric concussion cases and controls.
- Pathway analysis was conducted using Ingenuity Pathway Analysis to understand molecular mechanisms.
- Artificial Intelligence (AI) platforms, including Deep Learning (DL), were employed to predict mTBI based on methylation markers and combined predictors.
Main Results:
- 449 CpG sites (473 genes) showed significant methylation differences in mTBI patients compared to controls.
- Individual CpGs demonstrated high accuracy (AUC ≥ 0.80) for mTBI prediction, with four achieving excellent accuracy (AUC ≥ 0.90).
- AI models combining epigenomic and clinical data achieved ≥95% sensitivity and specificity for mTBI detection.
Conclusions:
- Significant gene methylation changes occur in response to mTBI in pediatric patients.
- Epigenetic dysregulation affects neurological pathways crucial for brain function, cognition, and behavior.
- AI-driven analysis of combined epigenetic and clinical data offers a highly accurate approach for concussion diagnosis.
