Related Experiment Video
Updated: Nov 13, 2025
19:57
The Use of Trace Eyeblink Classical Conditioning to Assess Hippocampal Dysfunction in a Rat Model of Fetal Alcohol Spectrum Disorders
Published on: August 5, 2017
8.6K
Detection of prenatal alcohol exposure using machine learning classification of resting-state functional network
Carlos I Rodriguez1, Victor M Vergara2, Suzy Davies3
1The Mind Research Network, 1101 Yale Blvd. NE, Albuquerque, NM 87106, United States.
Alcohol (Fayetteville, N.Y.)
|March 15, 2021
Summary
This study used functional network connectivity (FNC) from fMRI scans and machine learning to detect prenatal alcohol exposure (PAE) in rats. Results show FNC data can identify PAE in female rats, suggesting a potential diagnostic tool for Fetal Alcohol Spectrum Disorder (FASD).
Area of Science:
- Neuroscience
- Biomedical Engineering
- Public Health
Background:
- Fetal Alcohol Spectrum Disorder (FASD) presents significant diagnostic challenges due to varied criteria and lack of biomarkers.
- Prenatal alcohol exposure (PAE) causes neurodevelopmental abnormalities, necessitating reliable detection methods.
- Previous research indicated sex- and region-dependent alterations in functional network connectivity (FNC) due to PAE.
Purpose of the Study:
- To investigate the efficacy of using resting-state functional magnetic resonance imaging (fMRI) derived FNC data with machine learning for detecting PAE in a rodent model.
- To explore sex-specific differences in the ability of FNC data to classify PAE.
- To identify potential neural networks involved in PAE classification.
Main Methods:
- Resting-state fMRI data from adult rats with moderate PAE or control (SAC) exposure were analyzed for FNC.
- Spatial group independent component analysis (GICA) was used to extract resting state networks.
- Support vector machine (SVM) algorithms, including a quadratic kernel (QSVM), were employed for binary classification with leave-one-out-cross-validation (LOOCV).
Main Results:
- A QSVM classifier achieved significant accuracy in detecting PAE, particularly in females (79.2%).
- Classification accuracy for the aggregated sample was 62.5%, and for males, it was 58.3%.
- Analysis of QSVM weights highlighted the involvement of hippocampal and cortical networks in successful PAE classification.
Conclusions:
- Resting-state FNC data, analyzed with QSVM, shows promise as a non-invasive method for identifying PAE.
- The findings suggest that FNC patterns in adult female rats can be effectively used for PAE detection.
- This approach holds potential for the translational development of objective biomarkers for FASD identification.
