健康与遗传心理学研究室知识库

注意缺陷多动障碍（attention-deficit hyperactivity disorder ，ADHD）是一种常见的儿童神经和精神发育障碍。传统的诊断方法以主观评测为主。客观测评方法是注意缺陷多动障碍（ADHD）诊断与研究领域中新近发展的一种评估技术。但现有客观评测方法普遍存在数据采集不够方便，干扰因素多；分析和使用方面不够科学和难以普及的不足。 本研究的目的是研究探索使用一种民用级运动感知摄像机替换传统的运动传感器；使用信号处理频域分析替代传统简单身体动作描述数据的ADHD客观评测方法，以获取更为全面准确的数据，充分挖掘其信息在ADHD病情判断及诊断方面意义，为ADHD测评技术提供新角度，并做出革新。 本研究由两项子研究构成。研究一选取ADHD阳性和非ADHD儿童各30名，在持续操作测验情境下提取身体运动频域信息，依据此信息分别对两组儿童的数据实施独立样本t检验和ROC计算。研究二考察了14名经医生诊断的ADHD阳性及疑似儿童，在持续操作测验（CPT）情境下使用动作感知摄像机记录其身体运动能量释放信号数据，经频域分析，将15个频道的120种组合下能量统计数据与其《临床总体印象量表》（CGI-S）总分及《ADHD评分表》（ADHD-RS）总分及两个分维度得分做皮尔逊相关检测。研究一显示，所有组合t检验p值均小于0.05，其中有103项p值小于0.01。频率范围越高，其p值越小。使用ROC计算，120组AUC中最大值为0.94，最小值为0.87。研究二的将结果显示，能量频域组合与CGI-S得分相关系数检验结果中有30组r>0.6(p < 0.01),最大值为0.69（p = 0.006）；与ADHD-RS总分的相关系数检验结果中，有13组r > 0.6(p < 0.01)，最大值为0.69（p = 0.006）。r值图表呈现倒U型，但两种量表的峰值位置不同。 综上所述（1）使用身体运动能量信号频域数据可以有力区分ADHD及非ADHD儿童。（2）ADHD儿童身体运动能量信号的某些频域数据与ADHD病情严重程度呈显著相关。（3）使用该数据可对不同主观能评测量表做出能量角度的特征呈现。（4）该研究提出的基于身体运动信号的测量观点对于ADHD次测量学领域及临床领域具有重要意义。

Attention-deficit hyperactivity disorder (ADHD) is a neural and mental development disorder with a very considerable prevalence. The traditional ADHD evaluation method is generally subjective. Objective measurement is a new technology in ADHD clinical and research fields. But current objective measurement also has the deficits of inconvenience of using, bias of data collecting and insufficient analysis. This study is designed to use a consumer-grade motion sensing camera to replace the wearable motion tracking sensors and use the frequency-domain analysis to replace the descriptive parameters of subjects' body movement to collect more comprehensive data and find out the information inside of the data. In this way, it may bring a new point of view and evolution to ADHD assessment. This project is with two subjective studies. Study 1 recruited ADHD positive (n = 30) and non-ADHD(n = 30)subjects. The body movement signal is collected while the children doing CPT which is transformed from depth bitmaps serial, and were exhausted into 120 combinations by consecutive frequency channels. We deployed independent samples' t-test and ROC to all the 120 combinations. Study 2 inspected the Pearson correlation of body movement signal data while doing CPT and the score of Clinical Global Impression (CGI-S) and ADHD Rating Scale (ADHD-RS). All 120 t-tests of study 1 has p value less than 0.05, among them, there are 103 with p<0.01. The curve of the p value shows a descending trend, by the frequency's increasing, the confidence also increases. The results of study 2 showed 30 combinations' out of 120 correlation coefficients of CGI-S score and body movement signal has r > 0.6 (p < 0.01),maximum value is 0.69(p = 0.006); correlation coefficients of body movement signal and ADHD-RS has 13 out of 120 with r>0.6(p<0.01), the maximum r value is 0.69(p=0.006). The curve of 120 combinations has a shape of Kuznets curve, but the two scale has different peak locations. To sum up, (1)With body movement signal frequency-domain data, the ADHD children can be distinguished. This technology has a great potential to help to make diagnosis. (2) With this objective measurement, different rating scale's features can be observed.(3)ADHD children's some frequency channels of body movement has strong correlation with ADHD severity. (4) The point of view of body movement energy is of great significance to ADHD measurement field and clinical field.