NTU 与 IMH 开发 LEMon-AID，从日常对话识别精神病高风险青少年
新加坡 NTU 和 IMH 做的 LEMon-AID，听日常聊天就能筛出精神病高风险青少年，429 人研究里敏感度 93%，比常规 AI 方法强。
南洋理工大学与心理卫生学院的研究团队开发了 AI 框架 LEMon-AID，通过分析日常对话中细微的情绪模式，识别处于精神病超高风险状态的青少年，即使对方没有明确提及症状。研究纳入 429 名 14 至 29 岁的参与者，框架在识别高风险人群时达到最高 93% 的敏感度，预测准确率为 67%，优于传统 AI 方法。成果发表于 npj Mental Health，研究团队指出还需进一步临床验证才能用于早期干预。
Everyday conversations could hold clues to early mental health risks.
Researchers from #NTUsg and the Institute of Mental Health (IMH) have developed an AI framework that analyses subtle emotional patterns in casual speech to identify youths at ultra-high risk of developing psychosis, even when symptoms are not explicitly mentioned.
In a study involving 429 participants aged 14 to 29, the framework, LEMon-AID, achieved up to 93% sensitivity in identifying at-risk youths and 67% predictive accuracy, outperforming conventional AI approaches.
Published in @npjMentalHealth, the findings could pave the way for earlier identification and intervention, although further clinical validation is needed.
Read more: https://t.co/C3e7O11WTf
#NTUsgResearch #NTUsgPartnerships #AIatNTUsg #Healthcare #MedTech #Technology #MentalHealthMonth @NTU_LKCMedicine
