Talk:The big idea

From A conversation about the brain
Jump to: navigation, search

CT:

During your talk I heard the claim that most of the physiology has been solved. I was reminded of perhaps my favorite paper in vision science, "What is the other 85% of V1 doing?" in this paper Field and Olshausen argue, I think convincingly, that given the techniques used to study V1 a large portion of the function of primary visual cortex is terra incognita.

AG:

I was only referring to the fact that this proposal relies on mechanisms described a long time ago. I take your point. I am not trying to say that all neurophysiology is a solved problem.

CT:

I am very sympathetic to the idea that the brain can be described via linear systems, ideal observers, and simple computational models. I want it to be true and assume it to be true in my approach to psychology but one type of behavior give me pause -- categorical perception.

It gives me pause because not only is categorical perception a violation of Weber's law but it also seems (in some circumstances i.e., phonemes) immune to motivation. Perhaps I am missing something, but it seems that a linear system, even one that contains room for motivation, would be stumped by the discontinuity in voice onset time between /ba/ and /pa/ phonemes or the inability to distinguish and/or produce a difference in /r/ and /l/ sounds despite the motivation of the language learner to perceive or produce these phonemes.

AG:

I think I have not quite understood the issue here, but it sounds like an important one to resolve. In terms of the /r/, /l/ distinction, it makes me think of the issues I discuss in relation to anti-correlated stereograms where information is potentially available but not 'found' in some circumstances. We should discuss this more.