[评论] 研究发现一岁婴儿开始具有预测行为的能力

新浪科技

据美国广播公司(ABC)20日报道，对于绝大多数父母来说婴儿的行为是一个难解的谜团，没有人知道这些小家伙的小脑袋中思考着什么或者他们挥动手臂的用意又是什么。

　　研究人员称，在年幼的婴儿看来父母对于他们来说既神秘又让他们倍感困惑，因为，根据一项新的研究，不到六个月大的孩子完全不知道另一个人将要试图做些什么。瑞典的研究人员对六个月大的婴儿、一岁的婴儿以及成人进行了研究，让他们所有人观看一段一只手将玩具放入水桶的录象。当那只手拿起玩具时1岁大的孩子和成人很快便将实现转移到水桶上，这表明他们知道那只手即将做什么。但六个月大的婴儿并没有转移他们的视线。研究人员认为大脑使用特殊的神经布线来预言人们的举动。在婴儿六个月大时大脑中还没有形成这种布线。

　　这些发现并不能使父母更快地安抚婴儿停止哭泣，但它们或许为这种情景揭示某种信息。例如如果妈妈拿起一个可爱的玩具，六个月大的婴儿并不会意识到她打算将玩具递过来，即使她以前这样做过。换言之，新生儿在父母打算帮助他们时并不会意识到这一点。这项研究提醒我们不到六个月大的孩子在理解我们试图做什么时存在困难。

　　华盛顿大学医学院的儿科精神病学教授约翰·康斯坦提诺说：“我们应该认识到他们的行为并不是对他人注意力的操控或过度要求。”到婴儿一岁大时他们能更为容易地理解他们周围的人可能做出的举动。例如，当妈妈拣起一个球时他会料想到妈妈会把球放入水桶内，因为之前他曾看到她这样做过。研究人员解释说，如果他拣起了球他自己会将球放入水桶内。这听起来太过简单或微不足道，但研究人员的研究表明这具有深层次的含义。

　　芝加哥大学早期孩童时代研究中心的贝内特·伯恩哈尔说，婴儿预测其他人行为的能力与理解他们自己行为的能力息息相关。这种能力有时候被称作共享意向。这项研究表明我们不可能在这个年龄前更早地确定孤独症的出现，因为产生孤独症行为的电路在婴儿6到18月大之前还无法正常运转。这项发现或许解释了为什么婴儿在5或6个月大的时候似乎完全正常，但其后却出现了孤独症的症状。这些患孤独症的孩子负责共享意向的电路有可能没有发育完全。

　　研究人员称人们现在对于儿童大脑布线的认识仍然是一片空白。父母应该记住的是：年幼的孩子一无所知。当婴儿成长发育时他们的大脑变的能更好地理解他们周围的世界和他自己。这些知识或许并不能解决一个刚刚做父母的所有问题，但它或许能帮助他们更好一些地理解为什么孩子会做他们做的事情。(杨孝文
   任秋凌）

****************************************************

[Comment]

This conclusion may be meaningful statistically. Some baby could be very different timing for this progress.

大脑布线
     is interesting terminology to help the public readers to understand the concept, although it is not accurate.

[此贴子已经被作者于2006-6-23 0:32:54编辑过]

You are right Thinker_jeff, at least I think so. That concept is not so serious as the terms that has been defined by the authorities. However some thing interesting must be happening during that six-month-long period as I supposed.

An artificial neuron model used for AI I recently thought of might be able to form a memory loop or a sequence of memory loops with not very tight links. However, it can't count time at all which puzzled me of its intelligence. It can only explain why we memorys and why our memory last longer by stimulate the neuron with high frequency. It can't even solve a single problem yet.

May be the model lacks some important parts, can you help me?

[em17],这些很让人吃惊呢.了不起的发现

QUOTE:

以下是引用汪之旻在2006-6-22 14:47:00的发言：

An artificial neuron model used for AI I recently thought of might be able to form a memory loop or a sequence of memory loops with not very tight links. However, it can't count time at all which puzzled me of its intelligence. It can only explain why we memorys and why our memory last longer by stimulate the neuron with high frequency. It can't even solve a single problem yet.

May be the model lacks some important parts, can you help me?

I may not be a right person to give you an advice because I don’t know much about AI. Here are just a few ideas to share.

1. Your neuron model may only be one or a few of basic elements for information processing, that’s why it couldn’t solve a problem which needs much more elements than your model.

2. Some classic book in this field should be good reference for you. Reading more references seems important to you in such stage.

Do these help you at all?

Thanks, I will read more during my break. However, I don't think current AI would be able to count time as man does.