Predicting mixed neurological health risks from liquid crystal monomer mixtures in indoor dust using a network-driven machine learning model.

Liquid crystal monomers (LCMs) are emerging indoor environmental pollutants with potential implications for the human nervous system, and different LCMs usually coexist simultaneously. However, research focusing on the neurological risks associated with LCMs is limited, and no studies have yet addressed the risks associated with exposure to mixtures of LCMs. We analyzed 70 LCMs in 48 residential dust samples, finding trans,trans-4-(4-Ethoxy-2,3-difluorophenyl)-4'-n-propylbicyclohexyl (2OdFP3bcH; 158 ng/g), 4-Pentyl bicyclohexyl trifluoromethoxy benzene (5cH2OdFP; 105 ng/g), and trans-4-Fluoro-4'-(4-n-propylcyclohexyl)biphenyl (3cHFB; 88.4 ng/g) as dominant. The estimating daily intake (EDI) evaluation showed that ingestion and dermal contact were the primary exposure pathways, with infants and children experiencing significantly higher exposure levels than adults. Utilizing network science method, we prioritized neurological risks of LCMs, identifying 1-methyl-4-[4-(4-propylcyclohexyl)cyclohexyl]benzene (MeP3bcH) as the highest-risk compound. Based on the quantitatively calculated neurological risk index from network science, we further established a composite QSAR model coupled with the machine learning algorithm to assess the mixed neurological risk of LCMs mixtures, and found that charge distribution and ionization energy were the key factors influencing the neurologic risk of LCMs mixtures. This research provides both methodological and empirical support for enhancing the understanding of the neurological risks posed by LCM mixtures and for mitigating public exposure risks.