4 ms·

I read Jeff Hawkins book On Intelligence recently. Would these observations fit with the idea of visual data being passed up the hierarchy of the cortex far eno

I read Jeff Hawkins book On Intelligence recently. Would these observations fit with the idea of visual data being passed up the hierarchy of the cortex far enough to be recognising objects as obstructions, but not far enough to be conscious of them?

maebert 14y ago
The term "recognising" is a bit misleading here. Blindsight is in fact a very nice case where traditional cognitive science and AI paradigms suggesting a hierarchy of processing (bottom up of primal stimuli until they end up as meaningful concepts, top down of thoughts until they turn into actions (roughly)) doesn't quite fit.
- boothead 14y ago
  Yes perhaps categorising would be better?
KingMob 14y ago
People into higher-order and SDT-based theories of consciousness have argued similar things. One problem with the hierarchy model, though, is that the brain does not use layers as cleanly separated as in software. So, it's quite possible that after exiting the thalamaus and entering area MT or the amygdala, the signal goes on to "higher" regions such as the prefrontal cortex. So even though it's reaching the region, it's still insufficient. It could be because it's not stimulating enough neurons, but it's also possible that there's some qualitative difference as well. For example, sheer numbers of neurons don't guarantee awareness. The cerebellum has half the total number of neurons in the brain, but they're arranged in a very repetitive manner and seem to have nothing to do with awareness. If you had your cerebellum removed (say, for cancer), you would have difficulty with fine motor control, but not much else. It wouldn't alter your awareness, even though you've just lost half your brain's neurons. If you believe the information integration theorists, what matters is heterogeneity and interconnectivity, rather than mere numbers.

maebert 14y ago

The term "recognising" is a bit misleading here. Blindsight is in fact a very nice case where traditional cognitive science and AI paradigms suggesting a hierarchy of processing (bottom up of primal stimuli until they end up as meaningful concepts, top down of thoughts until they turn into actions (roughly)) doesn't quite fit.

boothead 14y ago
Yes perhaps categorising would be better?

KingMob 14y ago

People into higher-order and SDT-based theories of consciousness have argued similar things. One problem with the hierarchy model, though, is that the brain does not use layers as cleanly separated as in software. So, it's quite possible that after exiting the thalamaus and entering area MT or the amygdala, the signal goes on to "higher" regions such as the prefrontal cortex. So even though it's reaching the region, it's still insufficient. It could be because it's not stimulating enough neurons, but it's also possible that there's some qualitative difference as well. For example, sheer numbers of neurons don't guarantee awareness. The cerebellum has half the total number of neurons in the brain, but they're arranged in a very repetitive manner and seem to have nothing to do with awareness. If you had your cerebellum removed (say, for cancer), you would have difficulty with fine motor control, but not much else. It wouldn't alter your awareness, even though you've just lost half your brain's neurons. If you believe the information integration theorists, what matters is heterogeneity and interconnectivity, rather than mere numbers.