Development of a spontaneous pain indicator based on brain cellular calcium using deep learning

Area of Science:

• Neuroscience
• Artificial Intelligence
• Pain Medicine

Background:

• Chronic pain affects millions globally, with spontaneous pain being particularly difficult to diagnose and treat.
• Current preclinical models often fail to accurately represent human spontaneous pain, hindering treatment development.

Purpose of the Study:

• To develop a deep learning algorithm for detecting spontaneous pain from brain activity in preclinical models.
• To create a clinically relevant platform for real-time evaluation of pain and analgesic efficacy.

Main Methods:

• Developed AI-bRNN (Average training, Individual test-bidirectional Recurrent Neural Network), a deep learning algorithm.
• Utilized two-photon microscopy to record cellular Ca2+ activity in awake, head-fixed mice.
• Applied the algorithm to various cell types, brain areas, and somatosensory inputs (pain and itch).

Main Results:

• AI-bRNN accurately detected spontaneous pain intensity and timing in chronic pain models.
• The algorithm demonstrated versatility across different cell types (neurons, glia) and brain regions (cortex, cerebellum).
• AI-bRNN successfully evaluated analgesic efficacy in real time.

Conclusions:

• AI-bRNN provides a quantitative, real-time preclinical evaluation platform for pain medicine.
• This approach can accelerate the development of new diagnostic and therapeutic strategies for human chronic pain.
• The algorithm bridges the gap between preclinical findings and clinical application in pain research.