SAM: Automated Attachment Analysis Using the School Attachment Monitor

本研究旨在发展一套以电脑为本的工具，以具成本效益的方法量度全港市民的亲子依恋。国家儿童局指出，"安全的依恋促进健康和福祉"，而幼儿论坛倡导"儿童有权[.]形成安全、持久的依恋关系[...]塑造他们未来福祉的能力"。当这个问题被忽视时，后果是可怕的：有不正常家庭依恋的儿童有更高的攻击行为风险。到成年早期，有攻击性行为的人给社会造成的损失是同龄人的10倍，死亡率几乎高出10倍，部分原因是自杀和暴力行为的风险增加，但也是由于冠心病等身体问题。在人口水平上及早发现依恋问题，将对社会产生重大利益，并大大降低处理由此产生的问题的成本。大规模的依恋不安全感筛查应该成为儿童的常规检查。问题是依恋评估方法既昂贵又耗时。曼彻斯特儿童依恋故事任务（Manchester Child Attachment Story Task，MCAST）是儿童中期使用的标准方法。在MCAST管理期间，评估员使用玩具屋向孩子展示小插曲，描绘轻度压力的情况。然后，他们被要求用代表孩子和照顾者的玩偶来表演故事的其余部分。孩子完成故事的方式和他们在测试中的行为提供了评估他们依恋状态所需的线索。每个MCAST需要30分钟的管理和另外两个小时转化为可用的医疗记录。此外，专业人士必须参加昂贵的课程，然后进行长时间的可靠性培训才能使用MCAST，因此经过认证的附件评估员非常罕见。这意味着MCAST不能大规模应用，因为需要对人口健康和福祉产生重大影响。我们的目标是通过减少MCAST评估所需的时间和成本，使大规模依恋筛查成为可能。我们的方法包括自动化MCAST的关键步骤，以1）减少完成测试所需的时间（更高的效率），2）允许没有接受过MCAST培训的人员参与（更低的成本）。我们还希望MCAST的自动化能够为依恋及其可观察的、机器可检测的行为标记提供新的见解，从而使未来能够更好地测量依恋。我们将开发一种基于计算机的工具，可用于以快速，具有成本效益的方式测量整个人群的依恋，以支持MCAST评估员。孩子们将通过屏幕上的化身引导故事的小插曲。玩偶在空间中的详细动作和位置将被真实的时间捕捉到。我们还将记录孩子们的语音，以分析韵律和发声。使用这些数据，我们将开发新的算法来自动快速地对依恋模式进行分类，以一定的置信度将每个孩子定位在四个依恋类别之一中（安全;不安全的抵抗-矛盾;不安全的回避和不安全的混乱/迷失方向）。为了做到这一点，我们将开发基于社会信号处理（SSP）的新技术，其中Vinciarelli是一位领先的专家。有了SAM，筛选会议和初步数据分析可以在没有训练有素的MCAST评估员的情况下完成;只有当一个孩子被标记为属于其中一个问题类别时，他们才需要，在那里将进行标准的MCAST评估，这是第一次允许对依恋模式进行大规模人群筛查。SAM的发展和在大群体中快速筛查依恋将在儿童精神疾病的治疗中创造一个范式转变。

Our aim in SAM is to develop a computer-based tool which can measure parent-child Attachment across the population in a cost-effective way. The National Children's Bureau states that "secure attachment promotes health and wellbeing" while the Early Childhood Forum advocates "the right of children to [...] form secure, long lasting attachment relationships [...] which shape their future capacities for wellbeing". When the problem is neglected, the consequences are dire: children who have abnormal family attachments are at much higher risk of aggressive behaviours. By early adulthood, individuals with aggressive behaviour cost society 10 times more than their peers and have a mortality rate almost 10 times higher, in part due to increased risk of suicide and violent behaviour, but also due to physical problems such as coronary heart pathologies. Identifying Attachment problems early, at a population level, would be of significant benefit to society and drastically reduce the costs of dealing with the resulting issues. Large-scale screenings of Attachment insecurity should be routine among children. The problem is that Attachment assessment methods are expensive and time-consuming. MCAST (Manchester Child Attachment Story Task) is the standard method used in middle childhood. During MCAST administration, assessors show vignettes to the child, using a dolls-house, which portray mildly stressful situations. They are then asked to act out what happens in the rest of the story using dolls that represent both the child and a caregiver. The way the child completes the story and their behaviour during the test provides the cues necessary to assess their Attachment status. Each MCAST takes 30 minutes to administer and a further two hours to be transformed into a usable medical record. Furthermore, professionals must attend expensive courses followed by lengthy reliability training to use MCAST, so accredited Attachment assessors are a rare commodity. This means that MCAST cannot be applied on a large scale, as needed to make a significant impact on population health and wellbeing.Our goal is to make large-scale Attachment screening possible by reducing time and costs required for MCAST assessment. Our approach consists of automating the key steps of MCAST to 1) reduce the time needed to complete the test (higher efficiency) and, 2) allow the involvement of personnel with no MCAST training (lower costs). We also expect the automation of MCAST to provide new insights into Attachment and its observable, machine detectable behavioural markers, enabling better future measurement of Attachment. We will develop a computer-based tool which can be used to measure Attachment across the population in a rapid, cost-effective way to support MCAST assessors. The children will be guided through the story vignettes by an on-screen avatar. The detailed movements and positions of the dolls in space will be captured in real time. We will also record speech sounds from the children to analyse prosody and vocalisations. Using these data, we will develop novel algorithms to categorise Attachment patterns automatically and rapidly, locating each child in one of the four Attachment categories (Secure; Insecure Resistant-Ambivalent; Insecure Avoidant and Insecure Disorganised/Disorientated) with a level of confidence. To do this, we will develop novel techniques based on Social Signal Processing (SSP), in which Vinciarelli is a leading expertWith SAM, the screening sessions and preliminary data analysis can be done without the presence of trained MCAST assessors; they would only be needed if a child was tagged as being in one of the problem categories, where a standard MCAST assessment would be undertaken, allowing large-scale population screening of Attachment patterns for the first time. The development of SAM and the rapid screening of Attachment in large groups will create a paradigm shift in the treatment of child psychiatric disorders.