2010年10月27日,星期三

我们如何衡量"积分"?

“集成”是语言处理(和其他领域)中的主要操作。我们必须整合声音的各个部分以提取单词,整合语素的各个部分以获取单词的含义,将词汇语义信息与句法信息相结合,将感官与运动信息相结合,将音频与视觉信息相结合,而所有这些都与背景相关。

Some theorists talk specifically about regions of the brain that perform such 积分s. I've got my favorite sensory-motor 积分 site, Hagoort has a theory about phonological, semantic, and syntactic 积分 in (different portions of) Broca's area, more broadly, Damasio has been talking about "convergence zones" (aka, 积分 sites) for years.

Two thoughts come to mind. One, is there any part of the brain that isn't doing 积分, i.e., how useful is the concept? And two, if the concept does have some value, how do we identify 积分 areas?

I don't know the answer to first question and I have some concerns about the way some in the field approach the second. W.r.t. the latter, a typical approach is to look for regions that increase in activity as a function of "积分 加载". The idea is that by making 积分 harder, we will drive 积分 areas more strongly and this will cause them to pop out in our brain scans. This seems logical enough. But is it true?

假设Broca的区域(在艰难的过程中似乎总是参与其中的区域)在视听语音条件下的激活程度更高,在该条件下,与视听信号匹配时,视听信号不匹配(实际结果)。让我们考虑可能的解释。

1. Broca's area does AV 积分. It is less active when 积分 is easy, i.e., when A and V match than when 积分 is hard, i.e., when they mismatch 因为它 has to work harder to integrate mismatched signals.

2. Broca的区域不进行AV集成。当积分实际发生时,即当A和V匹配(反映其不参与)时,与未进行积分时(即AV失配)相比,它的活动性较低。当然,这种解释需要替代性的解释,以说明为什么Broca在不匹配的情况下会激活更多。有很多可能性:歧义解决,响应选择,错误检测或仅是WTF响应(考虑到Broca区域的响应属性,我有时想知道是否应该将其重新标记为WTF区域)。

Both possibilities seem perfectly consistent with the facts. Similar possibilities exist for other forms of 积分 making me question whether the "加载" logic is really telling us what we think it is telling us.

There is another approach to identifying 积分 zones, namely to look for areas that respond to both types of information independently but respond better when they appear together. In our example, AV 积分 zones would be those areas that respond to auditory speech or visual speech, but respond best to AV speech. I tend to like this approach a bit better.

你怎么认为?

5条评论:

乔纳森·皮尔说过...

一般来说我'我也更偏爱您提到的后一种方法。我认为在二分法(例如匹配与不匹配)而不是参数化的情况下尤其如此,因为似乎还有更多的替代解释空间。

但我认为'关于这些集成方法,这也是一个更普遍的问题。它'经常假设多峰区域表现出超加性响应;即,正如您所说,与单独使用任何一种方式相比,对视听语音的反应都更强。但是从某种意义上说,这至少对我来说是违反直觉的'normal'信息一致的情况。通常情况下,信息的丢失(或更大的歧义性)与神经反应的增强有关……对于语音和语义歧义性以及感知退化的情况,这是正确的。因此,这可能表明,在AV输入一致的情况下,如果可获得最多的信息,则需要的处理较少—because it'处理语音非常容易。

尽管完全不匹配的条件似乎很不自然(可能驱动了更多WTF响应),但是能够从参数上改变每种方式的信息量可能是解决此问题的一种方法。去年在NLC上有一些来自不同组的不错的海报,所以也许我们'很快就会获得更多数据。

乔纳斯说过...

我认为它’关于时间(原文如此)的事实是,狩猎时更考虑时域“integration”。没有任何时间信息的斑点肯定会使我们更接近祖母神经元,所以这是个好消息—如果您相信祖母神经元,那么可能没人会再做。积分—信息,方式等—我会说,随着时间的流逝,大脑区域的同步和不同步将引起语言人员的更多关注。

我只是重申许多聪明的同事之前为我们详细说明的内容。但是,对我来说,寻找整合而不是整合机制的狩猎区显得越来越徒劳。

乔纳森·皮尔说过...

I am of course all for taking into account timing information and trying to get 在 mechanisms of 积分 and not just regions, although surely 在 the end of the day we want to know both. I think we can still place useful anatomical constraints on theories of speech comprehension without getting to grandmother cells. ;-)

无论如何,例如,在考虑M / EEG数据时,可能采取的措施是'integrative'回应?可能其中一些类似于fMRI,具有多模式> unimodal responses in terms of power or peak signal? Synchronization/coupling? Latency of response? All of these? I am also curious as to what direction you predict these would go with increasing 积分 "load".

Although timing information will undoubtedly be crucial, I think that the main points raised by Greg are sort of methodology-independent...i.e., what experimental paradigms can we use to best test for 积分, and how do we measure it?

乔纳斯说过...

字!

布赖恩·巴顿说过...

我认为这是一个很难解决的问题。简而言之,我对使用与负载或任务难度密切相关的任何事物得出结论的方法感到怀疑。"Harder" 积分 means harder task, and it could simply be that our WTF area response has to do with working memory 加载 (or executive function demands) and nothing to do with A/V 积分 per se. That is difficult to disambiguate.

在最后一段,我与你在一起,这表明激活音频和视频,但同时兼顾这两个方面的区域是一个很好的起点,但仅此而已。尽管这两个维度都是可见的,但在V1中是一个简单的示例。 V1对定向线,颜色和对两者的响应都很好(在fMRI量表上)。但是,这并不能区分神经元实际上正在执行整合的可能性和正在执行两个单独的计算,以及对呈现两者的增加的响应只是两个细胞群体都在响应的可能性之间(人们仍在争论中水平"blobs" doing color and "interblobs"进行定位,这全都基于染色方法并且无法进行功能磁共振成像检查)。最后一部分是确定集成区域的大问题-它是'绝对有可能一个潜在的视音频集成站点实际上只是在听觉和视觉信息上执行相同或相似类型的计算,即使它们不是't integrated.