Видео

Very nice talk!. Very exciting topic! Undoubtedly it seems that the cerebellum will never cease to amaze us.

hey thanks for uploading this! presenting this paper in lab in a couple weeks so this is super helpful!

Great presentation

Really great presentation. Thank you.

This is an amazing lecture as a neuro physiotherapist I understood the overshooting or undershooting saccade we see when we use a tobii eye gazing camera to record eye movements. thank you, and I agree with your mum so what. How can we improve eye-directed body movements and balance? it would be amazing if you could reply. Farshideh Bondarenko

fantastic talk and discussion!

Thank you.

Dear Reza, Thank you for organizing these talks! This one was amazing! Not sure if Prof. Li would be able to answer here; My question is about the right graph at minute 19:26. Prof. later mentions that the experiments confirm that the same attractors with the same dynamics are activated when we do reversals, from Task 1 to Task 2 back to Task 1. This is inline with some kind of change point detection and structure learning hypothesis that are abundant in fear conditioning experiments for example; However this particular graph shows that even after reversals, many thousands of trials are needed to get above chance success rate for the animal. With a robust structural learning model, this should only take a few trials. I was wondering what is the reason for this discrepancy.

Looks like aim labs.

If I want to look at something at my left I have to move my shoulders first then my head would follow because only to move my eyes to the left is a times impossible with holding my head straight. As well there is no such thing as straight my head is tilted as my null point of view. I was diagnosed with cervical dystonia, spasmodic torticollis. And not to mention that as a child I had strabismus.

1. Reafference would appear to occur in the high cerebellar activity of entrainment of motor response to novel activity so useful in prediction. Prediction might be regarded as the final or evolutionarily sole value of a brain, opposed to mere neural entrainment occurring in cnidarians, as contrast, in motile organisms. For others outside mammalian brain-mapping, i note that highly socially, behaviorally, conscious sauropsids, and through to perhaps all Vertebrata, develop dense brain structure most likely endowing/including the plasticity useful for continual learning and modifiable sensorimotor response. In your brain, by the way, one half of the approximate totality of neurons, are contained in the cerebellum, which, despite its relatively smaller dimensions, indicates preserved utility. Speech, such as youtube comment, is relatively insignificant, being useful for symbolic informational utility, and subject, somewhat of an epiphenomenon, to the extreme deceptiveness to which many members of our species largely use it. Epiphenomena are very likely very transiently extant occurrences, immediately subject to selection. To digress, as youtube is so popular as to include much of human cognitive variations, were it not for our dense sympatric exuberance, we might have some time ago, speciated, separated through our sensorimotor distaste between those seeking to understand reality and those seeking permanent immersion in delusion, infantilizing oneself. Only we adolescents who remain rather critical of contrafactual ideation as social dogma, may be poised at a developmental stage useful for avoidance of reversion to medieval eroneous formation of heuristics, a term for distillation of novel stimuli and previously consolidated distillations, resimulating our individually highly limited past. Such differently-valenced and always contingent simplified distillation may confer great utility to some exafferent signal usage in the human brain. This is only generalized fodder for any as pissed off as i concerning what occurs in the anthropocentric fantasia outside laboratories.

1. RE: "twitch". Muscle itself as a word, arose from the perception of movement under skin that appeared similar to " musculus", mouse, in Latin. 2. Evolution is implicitly or explicitly defined as any proliferative change occurring in a population. Adaptation to niches in circadian cycle, that is, conditions favorable to survival/reproductive success during specific periods of light, dark, heat variance, even the obvious but unexplored lunar light variance affecting differential intentional utilities, too correlated to every physiological, hormonal, and circadian variability to be regarded as "error" when so obviously entrained. Coevolution, including predator-prey, sociocognitive, sociobehavioral optimalities, hormonal correlations, introduces some complexity to assessing both developmental and lifespan- related plasticity in wake-sleep states. What is not adaptive, is induced cyclic behavioral rigidity, beyond that essential for development of plasticity in response itself. So the question of optimality in neural, hormonal, and even glial monitoring of neurodevelopment, the latter also not well explored in development and functionality, may have to be approached across time, model organisms, and periodicity for clearer factor observation, hypotheses, experiments, conclusions. 3. The present difficulties in measurement of such plasticity in successful mature ( or any developmental stage) correctively self-entraining behavioral, and neuroconnective optimality does preclude observation of well-adapted cerebellar development of unconstrained individuals. Since cerebellum is involved in novel sensorimotor experience as well as sleep consolidation thereof, INCLUDING physical and incorrectly separated cognitive event consolidation, some noninvasive monitoring of Purkinje and cascading occurrences, including hippocampal, memory-encoding, and salience-encoding relational casades. Cerebellar activity, if reduced in intensity, continues to occur in familiar complex activity. These are only preliminary notes, formed within the first few minutes of the lecture.

Is this a new channel? Good content, love to see more! Definitely gonna get more subs!

uhm uhm uhm uhn uhm uhm uhm uhn uh uhm uhn uhm uhn

amazing talk!

Pure gold!

amazing talk Reza! Really enjoyed it. can the integration be done non locally, with the help of other circuits, maybe out of the CB, in a recurrent fashion, By which I mean that the integration result goes out of the CB, and then comes back again as a new input? The integration seems possible with addition of a memory unit, which can be represented in the hidden state of an RNN

Крутое видео получилось

This was fun. Thank you for uploading!

Fascinating.

at 53:46 is it a variance or covarianve C*cov(x)*CT

at 41:14 wd times xd not wn times xn

This is just different formulations of normal PCA. I did not find robust PCA in any part of this lecture. Totally misleading video title

beautiful lesson!

59:05..... WOOOOOOOO-HOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Thanks. Then is there any reason to use regular PCA at all?

Money talk, Paul! And sooo great to see the Shadmehr lab diving into this new direction. Tongues are amazing!

Great work Paulie. Proud!!

This is a crystal clear lecture on classical PCA. But why is it named robust PCA? I was expecting explanations on this: arxiv.org/pdf/0912.3599.pdf

What a great insight! Did not see Kalman from this way before

very knowledgeable professor, I really love this teaching style. Thanks for sharing.

very good lecture