AN ARTIFICIAL INTELLIGENCE FRAMEWORK FOR PAIN DECODING AND DRUG EFFICACY EVALUATION BASED ON NEURAL POPULATION DYNAMICS

Pain is a major global health burden, yet preclinical pain assessment still relies heavily on behavior-based assays that are vulnerable to stress, habituation, and limited reproducibility, thereby constraining the development of safer and more effective analgesics. To overcome these limitations, we propose an artificial intelligence (AI)–based framework for pain decoding and drug efficacy evaluation grounded in neural population dynamics rather than overt behavior. Using in vivo calcium imaging, we longitudinally recorded hippocampal CA1 neuronal population activity in freely behaving mice during nociceptive conditions and following opioid analgesic administration. Multi-scale analyses revealed that analgesic treatment robustly reshapes CA1 population dynamics, including suppression of overall activity, reorganization of temporal structure, stabilization of low-activity states, and compression of neural state-space trajectories, without reliance on repeated behavioral testing.
Building on these findings, we developed an AI-assisted neural decoding framework that integrates population-level neural features with advanced machine-learning architectures to infer pain- and analgesia-related brain states directly from neural activity. A brain-state–driven neuronal distillation strategy enriched drug-sensitive neural ensembles, improving decoding robustness and biological interpretability. Comparative model evaluation demonstrated that ensemble-based decoders achieved high discriminability, reliable probability calibration, and strong generalization across animals and experimental contexts.
This framework enables continuous, stimulus-minimized quantification of analgesic efficacy, temporal dynamics, and drug-specific neural signatures. By shifting evaluation from behavior to neural population states, this approach enhances temporal resolution, reduces experimental variability, and improves ethical compliance, establishing a scalable platform for pain research and preclinical analgesic drug development.