Has anyone looked for mutexes in cortical columns? How would one go about this?
-
-
Replying to @Plinz
From my superficial knowledge I would assume that mutex-like synchronization is performed by the mechanisms that control phase and frequency of neural oscillation, employing effects like resonance, harmonics, interference and delayed feedback of activity waves between columns.
1 reply 0 retweets 0 likes -
-
Replying to @Plinz
The exact mechanisms seem still unknown, but you are probably familiar with the tons of papers about phase-amplitude- and phase-phase coupling, especially theta–gamma coupling that provide some details, e.g: https://academic.oup.com/cercor/article/26/9/3744/2389108 … https://elifesciences.org/articles/20515 http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1005180 …
2 replies 0 retweets 0 likes -
Replying to @pavel23
Hypothesis: we can only animate one mental feature instance at a time because the other instances of that feature are inhibited. If the mutexes are unblocked it looks like this: https://vimeo.com/129710408
1 reply 0 retweets 1 like -
Replying to @Plinz
This article https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4623391/ … suggest that there are different processing stages (three distinct activity time windows), with bottom-up feature detection triggering predictive coding causing "top-down reactivation of target memory traces"
1 reply 0 retweets 0 likes -
So far all research suggests that all synchronization is performed by phase-amplitude coupling of different frequencies, not just binding, but also inhibitory effects.
1 reply 0 retweets 0 likes -
I would also guess that the upper "action"-layers of the columns with weaker feature responses in the lower layers are not silenced but "taken over", yielding a similar result and action, which would seem quite robust to me and explain why only one feature hypothesis survives.
1 reply 0 retweets 0 likes -
In this case, and if the six-layer model of the cortical column isn't completely wrong (which some research suggests), then the Layer IV-connections with the thalamus would play a significant role in determining the winning hypothesis and action.
1 reply 0 retweets 0 likes -
Replying to @pavel23
I don't find the top-down/bottom-up binding and hypothesis stabilization unclear. I am just wondering about how to deal with multiple feature instantiation, for instance, multiple letter 'A" in the same word, which may require mutexes. My question is much more specific!
1 reply 0 retweets 0 likes
Btw. I just discussed it for an hour with some local students and I think I am converging on a solution now.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.