Prediction of Autism Treatment Response from Baseline fMRI using Random Forests and Tree Bagging

TL;DR

This study develops a novel machine learning pipeline using random forests and tree bagging to predict autism treatment response from baseline fMRI data, aiming to personalize therapy choices.

Contribution

It introduces a new predictive approach combining random forest and bagging for treatment response prediction in ASD using fMRI data.

Findings

Achieved highest prediction accuracy compared to standard methods.

Successfully identified informative brain voxels for response prediction.

Validated approach with leave-one-out cross-validation on ASD children.

Abstract

Treating children with autism spectrum disorders (ASD) with behavioral interventions, such as Pivotal Response Treatment (PRT), has shown promise in recent studies. However, deciding which therapy is best for a given patient is largely by trial and error, and choosing an ineffective intervention results in loss of valuable treatment time. We propose predicting patient response to PRT from baseline task-based fMRI by the novel application of a random forest and tree bagging strategy. Our proposed learning pipeline uses random forest regression to determine candidate brain voxels that may be informative in predicting treatment response. The candidate voxels are then tested stepwise for inclusion in a bagged tree ensemble. After the predictive model is constructed, bias correction is performed to further increase prediction accuracy. Using data from 19 ASD children who underwent a 16 week trial of PRT and a leave-one-out cross-validation framework, the presented learning pipeline was tested against several standard methods and variations of the pipeline and resulted in the highest prediction accuracy.