Smoking in pregnancy likely to increase newborn behavioural issues, predicts AI

For about 50 years, smoking has been recognized as a risk factor for cancer, stroke, and diabetes. Research has highlighted its harmful effects during pregnancy, including high infant mortality, preterm birth, and low birth weight. Prenatal nicotine exposure is also linked to neurodevelopmental disorders like ADHD and ASD.

Now, scientists have devised a deep learning framework to analyse mouse behaviour in nicotine exposure experiments, aiming for more accurate results. They found that prenatal nicotine exposure heightened the risk of autism spectrum and attention-deficit/hyperactivity disorders in newborns, contributing valuable insights into the link between smoking during pregnancy and neurodevelopmental disorders.

“AI tools can label the body parts of animals in a marker-less video footage and precisely estimate their poses using supervised machine learning,” explains Prof. Katsuhiko Tabuchi. “Since animal behaviours are defined as a specific arrangement of body parts over a short period of time, deep-learning toolkits like SimBA can use the pose estimations obtained with DeepLabCut to classify different types of animal behaviours.”

Researchers conducted experiments using prenatal nicotine-exposed (PNE) and control mice to identify ADHD- and ASD-like behaviours. Cliff avoidance reaction tests showed PNE mice exhibited higher impulsivity, similar to ADHD traits in humans. Using a Y-shaped maze, researchers assessed working memory in mice by counting spontaneous arm switches. PNE mice showed decreased alterations, indicating altered working memory, a trait seen in ADHD. Open-field and social-interaction experiments revealed social deficits and increased anxiety in PNE mice, features of ASD traits.

“We validated the accuracy of our behavioural analysis framework by drawing a careful comparison between the results generated by the model and behaviour assessments made by multiple human annotators, which is considered the gold standard.” highlighted Prof. Tabuchi.

Reference: Mengyun Zhou, ORCID,Wen Qiu, Nobuhiko Ohashi, Lihao Sun,Marie-Louis Wronski,Emi Kouyama-Suzuki,Yoshinori Shirai,Toru Yanagawa,Takuma Mori,Katsuhiko Tabuchi; Journal: Cell; DOI:10.3390/cells13030275