基于Di-LSTM算法的注意力缺陷多动障碍症分类

Classification of ADHD classification based on Di-LSTM

对注意力缺陷多动障碍症(ADHD)受试者的准确识别一直是神经科学研究和临床诊断的挑战。基于更好的区分正常人和患者这一目的,文中采用了一种基于字典学习和长短期记忆(Long Short Term Memory,LSTM)网络的Di-LSTM算法,通过利用快速独立成分分析(Fast Independent Component Analysis,FastICA)初始化的在线字典学习,获得相应时间序列并且结合LSTM进行分类实验,实验结果表明,所提方法分类准确率达到了79.01%,特异性为88.9%,灵敏度为62.7%,说明该方法对于识别ADHD患者有所帮助,具有较好的应用前景。

Accurate identification of subjects with Attention Deficit Hyperactivity Disorder(ADHD) has always been a challenge in neuroscience research and clinical diagnosis. Based on better distinguish between normal and patients with this purpose,this paper uses a dictionary based learning and Long Short Term Memory(LSTM) network Di-LSTM algorithm, by using Fast Independent Component Analysis(FastICA) initialization of the online dictionary to study, to obtain corresponding time sequence and combine LSTM classification experiments,experimental results show that the proposed method classification accuracy reached 79.01%,the specificity is 88.9% and the sensitivity is 62.7%,indicating that the method is helpful for the identification of ADHD patients and has a good application prospect.