Adaptive Response Technology for Autism Spectrum Disorders Intervention

描述（由申请人提供）：据估计，美国每110名儿童中就有1名患有自闭症谱系障碍（ASD）。自闭症谱系障碍的识别和有效治疗通常被定性为突发公共卫生事件，每年花费350 - 900亿美元。旨在解决较高水平社会和适应性障碍的传统干预措施已被证明对患有ASD的学龄儿童和青少年的效果最低，并且被认为是传统方法未能在适当的强化剂量下系统地将干预策略与自然环境中的特定技能缺陷相匹配的结果。在本拨款申请中，为了解决上述问题，我们制定了一项研究计划，将智能适应技术作为基于虚拟现实（VR）的干预方式，用于治疗儿童和青少年ASD的核心社会缺陷和b)应用适应性行为技能。该技术将具有可控制难度水平的应用能力，可以根据参与者在参与任务时的行为参与预测，实时适应VR环境，适应社会和适应性行为干预场景。行为参与是由情感状态和注意状态实现的，因此，干预技术将对个体儿童参与的变化如何预测任务表现敏感。规则管理的适应将被纳入虚拟现实互动，并根据这些修改如何提高性能进行明确检查。简单地说，这项技术将被应用于根据个人的独特特征自动调整任务特征，希望通过个人调制和支架强化模式的应用，在核心受损领域内的任务表现将得到极大的加强。本研究的具体目标是：1)完善作者在试点工作中已经开发的智能自适应响应技术，设计VR环境下的自适应行为和社会任务；2)建立基于生理的个性化情感模型和基于眼睛注视信息的注意推理机制，并设计基于规则的监督机制，以实现对行为参与敏感的自适应强化策略；3)应用和检验预测模型和适应技术在ASD儿童中改善基于vr和现实社会和适应任务相关的特定学习的效果。这种适应性反应技术在ASD干预中的成功将为系统探索和应用旨在将个体缺陷与针对性干预相匹配的适应性干预范式铺平道路。

DESCRIPTION (provided by applicant): It is estimated that 1 in 110 children in the United States are affected by Autism Spectrum Disorders (ASD). The identification and effective treatment of ASD is often characterized as a public health emergency incurring a $35-90 billion annual cost. Traditional interventions designed to address higher level social and adaptive impairments have been demonstrated to be minimally effective for school-aged children and adolescents with ASD and are thought to be a result of a failure of traditional methodologies to systematically match intervention strategies to specific skill deficits within and across naturalistic settings in appropriately intensive dosages. In this grant proposal, in an attempt to address the above concerns, a research plan is developed to apply intelligent adaptive technology as a Virtual Reality (VR) based intervention modality for treatment of a) core social deficits and b) applied adaptive behavioral skills for children and adolescents with ASD. The technology will be applied with capacities for controllable levels of difficulty that can be adapted with the VR environment into both social and adaptive behavioral intervention scenarios in real-time based on predictions of behavioral engagement of participants while engaged with the tasks. Behavioral engagement is operationalized by affective and attentive states and as such, the intervention technology will be sensitive to how variations in an individual child's engagement predict task performance. Rule-governed adaptation will be incorporated into VR interactions and examined explicitly in terms of how such modifications improve performance. Simply, this technology will be applied to automatically adjust task characteristics based on an individual's unique profile in hopes that task performance within core domains of impairment will be greatly bolstered by application of individually modulated and scaffolded reinforcement modalities. The specific aims of this research are: 1) to refine the intelligent adaptive response technology that the authors have already developed in their pilot work, and design both adaptive behavior and social tasks in a VR environment; 2) to develop physiology-based individualized affective models and attention inference mechanisms from eye gaze information, and to design a rule-based supervisor to allow adaptive reinforcement strategies sensitive to behavioral engagement; and 3) to apply and examine the efficacy of prediction models and the adaptive technology in children with ASD relative to improving specific learning related to VR-based and realistic social and adaptive tasks. The success of such an adaptive response technology in ASD intervention will pave the way for systematic exploration and application of adaptive intervention paradigms aimed at matching individual deficit with targeted intervention.