4 ms·
The discussed paper is http://www.nature.com/articles/srep18112 http://www.nature.com/articles/srep18112 "The global landscape of cognition: hierarchical aggreg
by throwupper247 11y ago
The discussed paper is http://www.nature.com/articles/srep18112 http://www.nature.com/articles/srep18112 "The global landscape of cognition: hierarchical aggregation as an organizational principle of human cortical networks and functions"
Consciousness is mentioned only once:
>Our current data-driven results holistically demonstrated expanding amalgamation of function [sic!] over structural
depth [...]
>Fascinatingly, results suggest why the insular and opercular regions are some of the best candidates for
awareness or consciousness [62] [63] or why parts of frontal cortex are involved in higher cognition.
Wait, why is function uncountable?
>These conclusions are elaborated here:
>Sensory regions closest to inputs (bin1) in our hierarchy [include Insula and opercular]
>[...] hypothesized to play a role in multi-sensory integration and interoceptive awareness,
>[...] since it is the hub of the senses
>[...] the coordinating role awareness plays in cognition
>Further, middle regions, intermediate in distance from all sensory inputs (bin5) included hippocampus
and MTL known for memory and celebrity neurons
>Deep regions less connected to and farthest from inputs (bin10) [...] may support more abstract cognitive functions.
Aren't the first two quotes contradicting? Surely, awareness or consciousness is an amalgamation of function and distributed.
I'd rather thought the hippocampus was deemed central to cognition, as memory seems essential for re-cognition. Self-awareness is deemd a function of higher cognition and therefore an abstract concept. Why is it close to the concrete input stage? They talk about recurrent networks, does the hippocampus give feedback to the "input stage", maybe slowly as in conditioning? They "simplified recurrence into a continuum of statistical connectivity to sensory inputs" to probalisticly show a directional hierarchy.
Their methods are interesting, but far beyond me. They seem to build a model of the brain into a NN and test that, where a modified NN is the control group.