A Robust Machine Learning Based Framework for the Automated Detection of ADHD Using Pupillometric Biomarkers and Time Series Analysis

Abstract

Accurate and efficient detection of attention-deficit/hyperactivity disorder (ADHD) is critical to ensure proper treatment for affected individuals. Current clinical examinations, however, are inefficient and prone to misdiagnosis, as they rely on qualitative observations of perceived behavior. We propose a robust machine learning based framework that analyzes pupil-size dynamics as an objective biomarker for the automated detection of ADHD. Our framework integrates a comprehensive pupillometric feature engineering and visualization pipeline with state-of-the-art binary classification algorithms and univariate feature selection. The support vector machine classifier achieved an average 85.6% area under the receiver operating characteristic (AUROC), 77.3% sensitivity, and 75.3% specificity using ten-fold nested cross-validation (CV) on a declassified dataset of 50 patients. 218 of the 783 engineered features, including fourier transform metrics, absolute energy, consecutive quantile changes, approximate entropy, aggregated linear trends, as well as pupil-size dilation velocity, were found to be statistically significant differentiators (p < 0.05), and provide novel behavioral insights into associations between pupil-size dynamics and the presence of ADHD. Despite a limited sample size, the strong AUROC values highlight the robustness of the binary classifiers in detecting ADHD—as such, with additional data, sensitivity and specificity metrics can be substantially augmented. This study is the first to apply machine learning based methods for the detection of ADHD using solely pupillometrics, and highlights its strength as a potential discriminative biomarker, paving the path for the development of novel diagnostic applications to aid in the detection of ADHD using oculometric paradigms and machine learning.

Wavelet-based Approach for Diagnosing Attention Deficit Hyperactivity Disorder (ADHD)

“…Overall, our proposed method shows higher performance compared to the feature engineering method. Thus, the discriminatory descriptors derived though the assessment of self-similar behavior in the pupil diameter data show higher discriminatory power in detecting ADHD compared to the features derived from the original data domain used in Das and Khanna 4 .…”

“…We compared performance of the proposed self-similarity-based method with a feature engineering-based method proposed in Das and Khanna 4 . The feature engineering method engineered 22 customized features.…”

Accurate Identification of ADHD among Adults Using Real-Time Activity Data

“…Deep learning algorithms can also be used to distinguish ADHD patients from clinical controls using EEG data [ 52 ]. Moreover, indicators such as pupil size can be used to determine whether or not a patient has ADHD [ 16 ]. In connection to ADHD, the gender and age difference factors could also be used [ 12 ].…”

“…The current way of diagnosing ADHD is a subjective evaluation and thorough clinical observation which are mostly imprecise and erroneous, so there is a need for more objective methods. Since the diagnosis is exclusively based on observed behavior and reported symptoms, there is a possibility of over-and-under-diagnosis and also there are no valid objective tests to identify ADHD [ 16 ].…”

Wavelet-based Approach for Diagnosing Attention Deficit Hyperactivity Disorder (ADHD)

“…It is important to note that, in our analysis, we did not include the ADHD-on-medication group, as the inclusion of this group introduces bias into the ADHD-diagnosed group [5]. Thus, the dataset used in this analysis is reduced to 50 instances (22 controls and 28 ADHD-diagnosed children off-medication).…”