Nazhvani Adeleh Dehghani, Boostani Reza, Afrasiabi Somayeh, Sadatnezhad Khadijeh
Computer Science Engineering & IT Department, Faculty of Electrical and Computer Engineering, Shiraz University, Shiraz, Iran.
Clin Neurol Neurosurg. 2013 Nov;115(11):2329-35. doi: 10.1016/j.clineuro.2013.08.009. Epub 2013 Sep 1.
Children with Bipolar Mood Disorder (BMD) and those with Attention Deficit Hyperactivity Disorder (ADHD) share many clinical signs and symptoms; therefore, achieving an accurate diagnosis is still a challenge, especially in the first interview session. The main focus of this paper is to quantitatively classify the ADHD and BMD patients using their Visual Evoke Potential (VEP) features elicited from their Electroencephalogram (EEG) signals.
In this study, 36 subjects were participated including 12 healthy ones, 12 patients with ADHD and 12 ones with BMD. The age of ADHD patients was 16.92±6.29 and for the BMD ones was 17.85±3.68. Their scalp EEG signals in the presence of visual stimulus were recorded using 22 silver electrodes located according to the 10-20 international recording protocol. To extract their VEP, first a preprocessing step was executed to remove the power line and movement artifacts. Afterward, the wavelet denoising and synchronous averaging were applied to the preprocessed trials in order to elicit the P100 component. To obtain interpretable features from the evoked patterns, amplitude and latency were extracted and applied to the 1-Nearest Neighbor (1NN) classifier due to the locally scattered distribution of the VEP features.
The evaluation was performed according to leave-one(subject)-out method and the experimental results were led to 92.85% classification accuracy which is a fairly promising achievement to distinguish the BMD, ADHD, and healthy subjects from each other.
From the physiological point of view, this result point out to the existence of significant difference in the neural activities of their visual system in the ADHD, BMD, and healthy subjects in response to a periodic optical stimulus.
双相情感障碍（BMD）患儿与注意力缺陷多动障碍（ADHD）患儿有许多共同的临床体征和症状；因此，实现准确诊断仍然是一项挑战，尤其是在首次问诊时。本文的主要重点是利用从脑电图（EEG）信号中提取的视觉诱发电位（VEP）特征对ADHD和BMD患者进行定量分类。
在本研究中，共有36名受试者参与，其中包括12名健康受试者、12名ADHD患者和12名BMD患者。ADHD患者的年龄为16.92±6.29岁，BMD患者的年龄为17.85±3.68岁。使用根据10-20国际记录协议放置的22个银电极记录他们在视觉刺激下的头皮EEG信号。为了提取他们的VEP，首先执行预处理步骤以去除电源线和运动伪影。之后，将小波去噪和同步平均应用于预处理后的试验，以引出P100成分。为了从诱发模式中获得可解释的特征，提取了幅度和潜伏期，并由于VEP特征的局部分散分布而将其应用于1-最近邻（1NN）分类器。
根据留一（受试者）法进行评估，实验结果得出92.85%的分类准确率，这是区分BMD、ADHD和健康受试者的一个相当有前景的成果。
从生理学角度来看，这一结果表明，在ADHD、BMD和健康受试者中，其视觉系统在周期性光刺激下的神经活动存在显著差异。
