The role of artificial intelligence and machine learning (AI/ML) in the discovery of new drugs for neurological diseases in paediatrics

Abstract

Drug discovery and development is a highly complex and costly process, often requiring over a decade and billions of dollars to bring a single therapeutic to market. The development of artificial intelligence (AI) and machine learning (ML) methods has overcome many obstacles in drug discovery, advancing these processes by reducing the time needed to identify potential drugs through the prediction of pharmacokinetics, toxicity, and potential side effects. AI also improves clinical trial design by enhancing patient recruitment and data analysis, thereby reducing costs and increasing the success rates of drug discovery. AI utilises ML, deep learning (DL), and natural language processing (NLP) to analyse vast datasets, enabling the rapid identification of drug targets, prediction of compound efficacy, and optimisation of drug design.

AI and ML are also transforming drug discovery in neurological diseases, which are often complex and heterogeneous, affecting millions, but can also be monogenic and rare, with only a few patients. These diseases encompass over a thousand disorders that place a significant burden on human health and finances. Although computational tools have been developed and applied in neurological disease research for decades, the current era of AI leveraging ML offers significant potential to accelerate the identification and discovery of new drugs, although in some cases this is limited by our understanding of the disease. AI is also used in brain imaging and diagnostics, but predicting the permeability of the blood-brain barrier (BBB) is one of the most important applications of AI in drug discovery for brain disorders, as drug delivery to the brain is often hindered by the need to cross the BBB. Additional challenges arise from the scarcity of optimal models for preclinical drug testing and the frequently observed lack of translation from preclinical to human clinical trials. Recent advancements in assay-ready organoid platforms and microfluidic chips offer considerable potential for the use of human brain organoids in drug development.

Compared with adults, paediatric patients typically experience a significant delay, often nearly a decade, in accessing novel medications. However, AI and precision medicine are increasingly influencing paediatric pharmacotherapy, where age-dependent pharmacokinetic variability requires highly individualised therapeutic strategies. Clinical trials for paediatric ADHD drugs focus on evaluating the efficacy, safety, and optimal dosing of stimulants and non-stimulants in children and adolescents, with recent trials exploring novel formulations for improved symptom management. Several studies have developed ML and DL models to identify children with ADHD using electroencephalography (EEG) data. However, it is important to note that the use of artificial intelligence in the identification and treatment of ADHD remains limited.

Despite these achievements, AI-driven drug discovery still faces several limitations, including data biases, regulatory barriers, and ethical issues. Overcoming these limitations will be crucial to realising the full potential of AI in this field.