Abstract:
‘The different areas of the cerebral cortex are linked by a network of white matter, comprising the myelinated axons of pyramidal cells. Is this network a neural net, in the sense that representations of the world are embodied in the structure of the net, its pattern of nodes, and connections? Or is it a communications network, where the same physical substrate carries different information from moment to moment? This question is part of the larger question of whether the brain is better modeled by connectionism or by symbolic artificial intelligence (AI), but we review it in the specific context of the psychophysics of stimulus comparison and the format and protocol of information transmission over the long-range tracts of the brain.’
It’s the first one.
This network is neural net, in the sense that representations of the world are embodied in the structure of the net, its pattern of nodes, and connections.
As long as the connections are fixed (they are) and some areas are permanently connected to senses and motor drivers, the structure collects the patterns of the world in fixed places in that network.
Kind of true, but there is a lot multiplexing going on. In some sense, connections are fixed in any network, unless it’s wireless, but different segments are active depending on the input / configuration. These two options are not really well defined, they are more like coarse analogies.
From a temporal propogation perspective does the network/symbolic representation not change depending on the effective mode of operation (the brain wave frequency) ?
Consider that myelinated connections have a propogation delay and neurons also have propogation delays when the frequency of operation changes (e.g. source trigger characteristics) the way in which the resulting signals pass through that network may also change - albeit subtle. Enough to come up with new ideas (slower weaker out of phase connections brought to the fore) or think through fixed strategies quicker (faster in phase only what you know, existing strong paths only) ?
If the propogation delay of the neuron is effectively changed (e.g. output burst frequency) and the myelinated segment is not changed then the net result of timing may change the resulting winner inhibitory response or the resulting path.
Existing neural net’s seem to be currently defined with hard wired time so that they effectively embed and operate with an effectively fixed brain wave frequency ? They sort of lack the ability to think laterally and are only on 100% caffeine mode all the time, unable to think outside the box, so to say ?
Symbolic modeling approaches also assumes the symbols are not changed and hold constant, regardless of propogation timing effects ?
Does a changes of the thought frequency from say 7Hz to 15Hz not alter the resulting signal/wave/propogation path at all ?
The section on page 11 “transmission time is inversely related to the diameter of axons [63]” does not follow through taking into account the timing effects of increased myelination of axons over time and the net effect this has on the timing of the network as a whole [ref]. The network timing changes with faster winners. As the network learns the “known” routes may become faster, but how do we then think outside the box, back to those weaker slower paths… getting high…?
Idk but I like to imagine the type of network changes with scale.
on the scale of single neurons/hypercolumns its a neural net, but on larger scales, it may be behaving like a register that temporarily stores and broadcasts symbols (SDRs),