Congratulations to Bin Wang on his paper, the first from my lab, out now in @PhysRevLett ! https://t.co/XoEU5lDdTe 1/

ALL the PIs in our center signed the following letter (this is an English translation), because it looks like our government is continuing to push everyone toward the brink with their eyes open wide.

Coupling of olfactory receptors leads to sufficient structure, without clustering, to make random convergence onto the mushroom body effective. In summary, we show that peripheral preprocessing prepares *any* odor signal to be effectively readout by the central brain. 9/9

... but the theory assumes that neural responses in the input layer are clustered by the odor "meaning". That assumption turns out to not be true. Electrical coupling in the periphery comes to rescue the beautiful neural architecture and the theory. 8/

Previous experimental and theoretical studies proposed that learned odor discrimination is based on random connections to a region called the "mushroom body". This neural architecture seems perfectly suited for the task. https://t.co/kKltOIaxHB 7/

By analyzing the fly connectome, Palka showed how the brain takes advantage of the amplification, and optimally combines information from all receptors to determine if *any odor mixture* is 'good' or 'bad'. This generalizes previous studies showing readout of specific odors. 6/

The next step was to understand what this amplification means for odor processing in the fly's central brain. From the antenna, odor signals are relayed to two brain regions, responsible for detecting innately meaningful odors and new odors. 5/

The model solutions gave a direct link between the electrical coupling and amplified neural responses to specific odor mixtures. 4/

We built and solved a nonlinear model of coupled receptors driven by odor plumes. The model is based on years of work by Chih-Ying's lab (including our collaborator @ShiuanTze Wu), characterizing the peripheral odor map. https://t.co/icJGGZfLOn 3/

The array of olfactory receptor neurons in flies is highly structured. Our collaborator Chih-Ying Su @SuLabUCSD showed previously that receptors are electrically coupled, a coupling that affects animal behavior. https://t.co/phQgGQeNHK 2/

I am delighted to share our preprint led by Palka @PuriPalka. Our study of preprocessing of olfactory information in flies provides insight to general principles of sensory information processing in shallow neural architectures. https://t.co/OlLDEwcs7u 🪰 #tweetprint 👇🏽 1/

Congratulations to Palka @PuriPalka for receiving the @UCSanDiego GPSA Interdisciplinary Research Award! Pictured with Palka are Chih-Ying @SuLabUCSD and I looking and feeling very proud.

https://t.co/4WuAV6ciu4

The distortion and disintegration of grid cells' patterned activity challenges the idea that those cells are a "neural GPS" used for spatial navigation. @gilyginosar and friends review the evidence. Out today @NeuroCellPress https://t.co/sjUEskBYG8

Anarchy has fully taken over Tel Aviv ;)

All universities in Israel are on strike starting Monday, protesting the government's anti-democratic takeover of the judiciary. No instruction until further notice.

A delay in notification will cause problems for many international scientists who need to plan their travel around visa applications (to enter Canada) and renewals (to return home). @CosyneMeeting

Looking forward to visiting the Hebrew University tomorrow, and to present our work led by @PuriPalka , with @ShiuanTze and @SuLabUCSD

Hi @airserbia , I tried buying a ticket on your website. It says that the payment is processing for 40 minutes. I am now trying to call customer service, no answer for 20 minutes. Can someone tell me if the ticket was purchased?

Please RT! We're #hiring a #postdoc to develop graph-based models and high-dimensional ML methods to analyze dynamical networks. Collab w/@schoeneberglab @FleurMFerguson using cutting-edge 4D lattice light-sheet imaging. Join our multi-disciplinary team! https://t.co/VcZf5XfmP5

Tomorrow, Wednesday morning #SfN2022 Poster BB15 Bin Wang will present our work on distributed predictive processing. Based on our model, we think that loose balance allows prediction error and stimulus representations to mix.