MindTS-MMD

The MINDTS-MMD (Mindfulness-based Intervention for Tourette Syndrome – Multimodal Dataset) is the first Chinese pediatric Tourette syndrome (TS) multimodal dataset designed to support emotion-aware and behavior-adaptive therapeutic modeling. It integrates synchronized video, audio, and text modalities, capturing children’s tic behaviors, emotional expressions, and interactive contexts during mindfulness-based intervention sessions. The dataset includes over 120 hours of recorded sessions from 60 clinically diagnosed TS participants aged 6–12 years, each with expert-verified annotations of tic frequency, Valence–Arousal emotion scores, and contextual metadata such as activity type and therapeutic stage. All recordings were de-identified and processed under medical ethics approval (Approval No. 2022-114-01(K), Xinxiang Central Hospital). The MINDTS-MMD dataset enables multimodal affect recognition, behavioral monitoring, and personalized guidance generation, providing a benchmark resource for AI-assisted mindfulness interventions in pediatric neurodevelopmental care.

MINDTS-MMD（基于正念的抽动秽语综合征干预多模态数据集，Mindfulness-based Intervention for Tourette Syndrome – Multimodal Dataset）是首个面向中国儿童抽动秽语综合征（Tourette Syndrome，TS）的多模态数据集，旨在支撑情感感知与行为适配的治疗建模研究。该数据集整合了同步的视频、音频与文本多模态数据，记录了儿童在正念干预疗程中的抽动行为、情感表达与交互场景。数据集包含60名经临床确诊的6~12岁TS患儿的120余小时干预疗程记录，所有记录均附带经专家核验的标注内容，涵盖抽动频率、效价-唤醒（Valence–Arousal）情感评分，以及活动类型、治疗阶段等上下文元数据。所有录制数据均已完成去标识化处理，并获得新乡市中心医院医学伦理委员会批准（批准号：2022-114-01(K)）。MINDTS-MMD数据集可用于多模态情感识别、行为监测与个性化指导生成，为儿童神经发育照护领域的AI辅助正念干预提供了基准研究资源。