Modeling ADHD symptoms and medication effects with AI and math
A discrete-continuous-time stochastic framework for attention-deficit/hyperactivity disorder-related symptom modeling and pharmacological intervention via methylphenidate
Machine Learning
Summary
ADHD involves brain processes that often get stuck in patterns causing symptoms like inattention and hyperactivity. The authors created a new math-based model to describe these brain patterns as random paths that tend to stay in unhealthy states. They showed how the medicine methylphenidate can be represented as a control signal that nudges these paths back toward normal brain activity. To connect their math ideas to real brain scans, they built a deep learning model that learns hidden brain activity patterns from imaging data.
What this means in practice
- •For clinical data teams: Use the model to improve analysis of brain imaging and medication effects for ADHD patients.
- •For machine learning engineers: Develop hybrid architectures combining CNN and Vision Transformer for interpreting biological time series data.
Tested on one dataset.
Authors
Oana-Maria-Catalina Crisan
Abstract
This paper presents an original mathematical and computational framework for modeling the efficacy of methylphenidate as an optimal control policy acting on the nonlinear dynamics of attention deficit hyperactivity disorder. Neural processes in ADHD are conceptualized as stochastic processes with trajectories that inherently converge towards stable, yet dysfunctional patterns associated with clinical symptoms. The pharmacological intervention is mathematically formalized as an external control signal that reassesses the system's drift and redirects these trajectories towards neurotypical regimes. To bridge these theoretical stochastic models with empirical clinical data, latent representations are inferred using a novel deterministic hybrid deep learning architecture, combining a one-dimensional convolutional neural network and a Vision Transformer, trained on preprocessed functional magnetic resonance imaging data.