酷兔英语





Can a computer tell when it hurts? It can if you train it, US researchers said on Tuesday.
A team at Stanford University in California used computerlearning software to sort through data generated by brain scans and detect when people were in pain.
"The question we were trying to answer was can we use neuroimaging to objectively detect whether a person is in a state of pain or not. The answer was yes," said Dr. Sean Mackey of the Stanford University School of Medicine.
Currently, doctors rely on patients to tell them whether or not they are in pain. And that is still the gold standard for assessing pain, Mackey said.
But some patients -- the very young, the very old, dementia patients or those who are not conscious -- cannot say if they are hurting, and that has led to a long search for some way to objectively measure pain.
"People have been looking for a pain detector for a very long time, We're hopeful we can eventually use this technology for better detection and better treatment of chronic pain."Mackey said.
For the study, Mackey's team used a linear support vector machine -- a computer algorithm invented in 1995 -- to classify patterns of brain activity and determine whether or not someone is experiencing pain.
To train the computer, eight volunteers underwent brain scans while they were touched first by an object that was hot, and then by one that was so hot it was painful.
The computer used data from these scans to recognize different brain activity patterns that occur when a person is detecting heat, and which ones detect pain.
Mackey cautioned that the study was done in a very controlled lab environment, and it did not look at the differences between chronic and acute pain.


据路透社9月13日报道,本周二(9月12日),美国研究人员称,经过训练的计算机可以帮助人类诊断身体疼痛。

美国加州斯坦福大学的一个研究团队利用计算机学习软件分类整理了由脑部扫描所得的数据,从而对人们处于疼痛的时刻进行了检测。

斯坦福大学医学院的肖恩•麦基博士说:"我们正在尝试解决的问题是,我们能否用神经成像技术来客观检测出一个人是否处于疼痛状态,而答案是肯定的。"

麦基博士说,通常,医生从病人口中得知他们是否感到疼痛,这种方式依然是医生诊断疼痛的"黄金标准"。

但对于一些病人,例如幼儿、老人、痴呆患者或神志不清的人,他们不能清除表达自己到底是不是感到疼痛,这就要求对客观诊断疼痛的方法做一些深入研究。


麦基博士说:"很长时间以来,人们一直都在寻找 '疼痛诊断器'。我们有望最终利用这项技术对慢性疼痛病症做出更好的诊断和治疗。"

研究过程中,麦基博士的团队使用了一个线性支持向量机--发明于1995年的一个计算机程序法--将大脑活动的模式进行分类,从而诊断一个人是否正在经受着疼痛。


为了训练计算机,研究人员首先用一个热的物体触碰参加实验的八位志愿者,然后再用一个更热的会让人产生疼痛感的物体对他们进行触碰,并分别对其大脑进行扫描。

电脑利用扫描得来的数据识别人们在触碰到热和感到疼痛的时不同的大脑活动模式。


麦基博士提醒说,此实验是在一个十分受控的实验环境下做的,它尚不能分辨出慢性疼痛和急性疼痛。