Monday, November 29, 2021

Signal to Noise for the Win


A new model for how the brain perceives unique odors
Oct 2021, phys.org

So here is a scientist, a physicist by name, with an interest in the information-processing abilities of biological, and neurological systems. It didn't take him long, I would imagine, to realize that olfaction is a prime model for complex information-processing systems; it was the first sense, used by the first bacteria to detect chemicals in the primordial soup, later used by animals to make a big, complex mammal brain. If you want to look at how information processing happens in biological systems, this would be the ideal place to look.

But first, you have to throw into the garbage everything you know about olfaction already, which should be easy for a computational information scientist. What they did here was to look far out, all the way out -- beyond molecules and their myriad physico-chemical characteristics, of which the molecules themselves number in the billions; beyond genetics and their 30% variation across the global population; beyond cultural effects that are almost too surreal to quantify for methodical research purposes, like where Americans prefer peppermint since it's associated with candy, yet older folks from England don't like mint since it's associated with pain-relief products of their era, or the easier comparison of preference for durian fruit or Époisses de Bourgogne cheese.

Too messy, said the computational information scientist. And so they put it all in the blender, all those variables together. And they called it noise. Boil it all down, cancel it all out, all that dirty data of molecules and genes and cultural and personal association. Throw it all in the same bin, and call it noise. That's what they did.

Actually, they didn't call it noise, they called it "context" --

 "If you experience odors in a similar context, even if they were initially rather different in the responses they evoked in the nose, they begin to be represented by similar neural responses so they become the same in your head," Balasubramanian says.

The researchers found that their simplified model could be used to reproduce the same types of results seen in olfaction experiments. It's something that Balasubramanian did not expect to see, as he thought that such a complex process would require "learning and plasticity" in order to adapt and change neural synapses to modify the brain's representation of smells. "We may have found a general strategy of using certain kinds of randomized signals to entrain those effects," he says about their results. "It doesn't have to be just olfaction; it can be elsewhere, too." -medicalexpress

Did you see that? "It doesn't have to be olfaction; it can be elsewhere too." Olfactively-piqued, computational information neuroscientist, where have you been? Proving that the nose-brain is the neural model par excellence, while showing us how it actually works, both at the same time.

*If you want to know more about why olfaction is the ideal model for growing an artificial brain from scratch, it's a constant theme in Hidden Scents.

via University of Pennsylvania: Gaia Tavoni et al, Cortical feedback and gating in odor discrimination and generalization, PLOS Computational Biology (2021). DOI: 10.1371/journal.pcbi.1009479

Post Script:
This study shouldn't be mentioned without this other study, where they taught an artificial network how to smell, via Massachusetts Institute of Technology's McGovern Institute for Brain Research: Peter Y. Wang et al, Evolving the olfactory system with machine learning, Neuron (2021). DOI: 10.1016/j.neuron.2021.09.010. http://dx.doi.org/10.1016/j.neuron.2021.09.010

Post Post Script:
Might as well put this here, since it has to go somewhere -- what happens when you take an information scientist and give them an olfactory science problem? This is what happens. They come up with an answer that is so simple it just makes you look stupid. This is an example, although not a true example, of the other half of the coming dark ages. After a global pandemic, it's inevitable to experience a kind of dark ages, where lots of people died, but way more people got sick, and also a lot of people retired. That's a post-pandemic-pandemic of institutional knowledge loss, like a collective long covid brain fog on our culture -- we forget how to do stuff, because the guy who did it for the past 30 years isn't doing it anymore. That guy either died, got too sick to work, or retired for a million other reasons, of which many of them could be pandemic-related. And there's lots of those guys (and even more of them gals). Who knows what that will look like for us today or tomorrow, but it's happening as we speak, and years from now we might notice, we might even call it the great forgetting. The flip side to the dark ages is the renaissance, which comes from all the new people in new roles and at new jobs. These people are coming in new, with nobody around to teach them how to do things "right," and although that makes for a bumpy road ahead, it also gets you things like this discovery, one of the hardest problems of olfaction taken on by someone who has not much at all to do with olfaction (although he should, because olfaction has been an information science problem all along).

Wednesday, November 24, 2021

Olfaction In Silico


Artificial networks learn to smell like the brain
Nov 2021, phys.org

No kidding, the sensory apparatus that resembles a deep learning neural network can be simulated with a deep learning neural network -- "Artificial networks trained to classify odor identity recapitulate the connectivity inherent in the olfactory system."

The part of our brain that smells is also the most primitive. Before brains were a thing, bacteria performed chemosensory calculations on the primordial soup. As the soup became more complex, so did sensory equipment. Chemosensitive receptors on the surface of a bacterium became antennae, and then became noses, and those noses became seeing, hearing, even speaking brains. But the first version is the one used for smelling. So it shouldn't be a surprise that the first place we see a direct link between the mammalian brain and our artificial instantiation is via olfaction. Nonetheless, the scientists were "surprised to see it replicate biology's strategy so faithfully."

"By showing that we can match the architecture very precisely, I think that gives more confidence that these neural networks can continue to be useful tools for modeling the brain," says Robert Yang, assistant professor in MIT's departments of Brain and Cognitive Sciences and Electrical Engineering and Computer Science [and who collaborated on this project with Columbia neuroscientists Richard Axel and Larry Abbott, btw].

They use an antennae model, and the indispensable fruit fly, but it's all close enough to humans, I mean that's why we use the fruit fly in the first place. They started with some artificial neurons, of the same amount found in a fruit fly. They programmed the neurons to identify odors, and to assign valence (pleasant or unpleasant) to odors. They didn't give the neurons any structure, no information about how to talk to each other, no blueprint on how to process information. Just pre-programmed neurons, thrown into a simulated universe of chemosensation.

In minutes, and in silico, a network emerged to look just like the nose-brain of a fruit fly. All on its own, "an initially homogeneous population of neurons segregated into two populations with distinct input and output connections, resembling learned and innate pathways." 

In other words, it learned how to smell. It took evolution some billions of years to get the fruit fly olfactory system just right. The artificial network did it in minutes. Extrapolations from the study abstract: "This implies that convergent evolution reflects an underlying logic rather than shared developmental principles."


*Their structure used expansion and compression layers, similar to the pyramidal-structure of the nose-brain, and the number of inter-connections was also the exact same number as in the fruit fly, and it worked with both feedforward and recurrent network models, and the networks are plastic, meaning they can learn new odor associations over time.  

via Massachusetts Institute of Technology's McGovern Institute for Brain Research: Peter Y. Wang et al, Evolving the olfactory system with machine learning, Neuron (2021). DOI: 10.1016/j.neuron.2021.09.010

Post Script:
The book Hidden Scents talks about how olfaction is an ideal model for understanding an artificial brain, for an artificial human. 

Although artificial neural networks resemble natural neural activity patterns, like those used by the visual cortex for object recognition, we still don't understand how the visual cortex, or most mammalian neural circuits, are inter-connected. This time, we see how they connect, a connectome of the olfactory cortex.

Post Post Script:
And this study should be kept alongside this other one, where a physics-based computational neuroscientist came up with a pretty simple way to mimic the olfactory cortex, via University of Pennsylvania: Gaia Tavoni et al, Cortical feedback and gating in odor discrimination and generalization, PLOS Computational Biology (2021). DOI: 10.1371/journal.pcbi.1009479

Thursday, October 21, 2021

Nobody Is an Expert at Everyday Smells, Not Even the Smell Experts


If you find the language of smell interesting, Asifa Majid is the scientist to follow. This is a study from 2016, but it never hurts to be reminded that your ability to perform overt olfactory exploration on your environment is likely way worse than it could be. 

We don't talk about smells, and that's part of the reason why we can't identify them. A forgotten milk box in a student's locker over summer vacation may just as well be a poor rodent who happened to die under your staff refrigerator. These are pretty different smells, but they're both "bad," and that's about as much as we need to know. (They're both protein-based, so those two descriptions are not too far apart.)

Now you might think that a wine expert would be an ideal model for a chemosensing humanoid, but no. Experts are only good at smelling the things they've been trained to smell. They are only good at smelling what they already have words for. Granted, they have lots more words than your average person, but only words for odorants specific to their discipline.
Experts only have a limited, domain-specific advantage when communicating about smells and flavors:
Neither expert group was any more accurate at identifying everyday smells or tastes. Interestingly, both wine and coffee experts tended to use more source-based terms (e.g., vanilla) in descriptions of their own area of expertise whereas novices tended to use more evaluative terms (e.g., nice).
-Not All Flavor Expertise Is Equal: The Language of Wine and Coffee Experts. Ilja Croijmans, Asifa Majid. June 2016,  PLoS ONE 11(6): e0155845. https://doi.org/10.1371/journal.pone.0155845
Putting things in perspective, it should be noted that a master perfumer has a vocabulary of **thousands** of words for smells, which is ten times more than the wine or coffee expert. (You want to know some of the words are in the smell lexicon for coffee experts? Check out the World Coffee Research Sensory Lexicon and scroll down to "Publications;" and then you can compare that to the 1,000+ descriptors in the Sigma Aldrich catalog of fragrance chemicals which are purchased by perfumers to make perfume.)

If you want to be an "Everyday Smell Expert" all you have to do is start to come up with a vocabulary of your own for the smells around you, and then start paying attention. Soon you'll be able to recognize odors in your environment at really low thresholds, and maybe even be able to accurately  associate them with their source. 


Post Script:
Later in the paper, the idea of cultural difference as inhibiting our olfactory prowess:
"The difficulty people have in naming smells and flavors could be a WEIRD (Western, Educated, Industrialized, Rich, Democratic) affair. -source

So another approach to becoming a better everyday-smeller would be to try growing up in a non-WEIRD country.


Thursday, October 14, 2021

Neuromorphic Odor Translator Helps Robots Express Their Feelings


Neural network trained to properly name organic molecules
Aug 2021, phys.org

Do you ever have a problem naming that smell? It's not just you. Science has this problem too, but maybe not for long. 

Smells are volatile organic compounds that have evaporated and entered your nose. Although they almost always occur in combination with others and not in isolation, us humans want to reduce smells to their individual components, and then name them. After all, in order to think about something, you have to know it's name. (Is that true?) 

The problem is that organic molecules are big, with lots of chemicals joined together in lots of ways, so coming up with a naming convention for all these permutations is hard. IUPAC, the International Union of Pure and Applied Chemistry, sets the convention for naming molecules. And boy is it complicated.

Take sugar, a common molecule known to us by its simple name "sucrose;" in IUPAC, it's called (2R,3R,4S,5S,6R)-2-[(2S,3S,4S, 5R)-3,4-dihydroxy-2,5-bis(hydroxymethyl)oxolan-2-yl]oxy-6-(hydroxymethyl)oxane-3,4,5-triol.

Since we do live in the computer age, folks want to automate this naming process for when they discover new molecules. But as you can imagine by looking at the IUPAC name for sucrose, the algorithm at the core of that naming convention is really hard to write. So they decided to use a neural network instead.*

*I'm casually calling a neural network "neuromorphic," but in the past few years, real neuromorphic computers have forced a distinction here that I'm ignoring here for the sake of a more clickable title. 

This new artificially intelligent chemical translator is not a magical structure-to-name translator that can just look at a chemical and give it a name; that's still out of reach. It does, however, translate between IUPAC and another naming convention called SMILES.

Trained on PubChem's 100 million molecules, this translator ultimately shows how the utility of the new approach of using neural networks to help us write algorithms from the bottom up instead of the top down, which really is a revolution in computing. 

And if you think it would be cool to have robots that can smell, or to ensure that future humans maintain their sense of smell as they evolve into hyperdimensional algorithms, then making odors machine-readable is how you do that.

via Skolkovo Institute of Science and Technology, Lomonosov Moscow State University and start-up Syntelly: Lev Krasnov et al, Transformer-based artificial neural networks for the conversion between chemical notations, Scientific Reports (2021). DOI: 10.1038/s41598-021-94082-y

Thursday, October 7, 2021

On Olfactory Navigation


I purposely read this book Supernavigators (2019) hoping to get some snippets on using our sense of smell to find things, and I wasn't disappointed. 

Humans were led to a random location within a room diffused with two odors. After brief sampling and spatial disorientation, they had to return to this location. Humans located the target with higher accuracy in the olfaction-only condition than in the control condition and showed higher accuracy than chance. 
-Jacobs, L.F.; Arter, J.; Cook, A.; and Sulloway, FJ. (2015). "Olfactory orientation and navigation in humans," PLOS 'One, 10(6), e0129387. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4470656/

Note there are two different versions of olfactory navigation -- one where you track an odor to its source (this relies heavily on bilateral input, aka stereo-olfaction) and the other, much more common for modern-day humans, is when you identify a place by its odor. We usually have our eyes open, and being the ocularcentric creatures that we are, we are likely to use visual cues and not even realize the odor-identity of a place. 

But it doesn't stop here, the rabbit hole continues, and this one goes all the way back to the golden days of behavioral science, when rats told us everything we wanted to know about ourselves: 
This report is ultimately based on rat experiments, with the "men" part being only conjecture by the researcher; and he concedes, "My argument will be brief, cavalier, and dogmatic. For I am not myself a clinician or a social psychologist. What I am going to say must be considered, therefore, simply as in the nature of a rat psychologist's ratiocinations offered free.
*Ratiocinations are another word for thoughts that also happens to remind the reader that we're talking about rats (he italicized the rat in ratiocinations).

The "mapmaking" happens during what they call "Vicarious Trial and Error" or "VTE'" and described as "the hesitating, looking-back-and-forth, sort of behavior which rats can often be observed to indulge in at a choice-point before actually going one way or the other." If you're not a scientist, you can probably just call it "thinking."

via Berkeley Labs: Tolman, E.C. (1948). "Cognitive maps in rats and men," Psychological Review, 55(4), p.189.
And with that, let us not forget that olfaction is the first sense. Before all the other ways we sense our environment, bacteria and fungi were using chemotaxis, detecting and navigating their way through a world of chemical gradients. The essay at the end of Hidden Scents, called "Olfactory Space and n-Dimensionality" tells the story of the primordial eukaryote as it chemo-taxis its way through evolution, past the multi-cellular organism, the chordata (animals with vertebrate), and eventually to the big-brained, smooth-skinned monkeys that we are today.

The neocortex is an outgrowth of the nose-brain, and not the other way around, and therefore olfaction can be a useful model for understanding the n-dimensional information network in which our brains operate. The world is typically understood as a 3-dimensional space, but in fact, from the perspective of the brain, we are navigating and interacting with an infinitely-dimensional information space. 

 

Thursday, September 30, 2021

Temporal Patterns in Olfactory Perception


Temporality in olfactory perception is understudied. But as our ability to measure spike activity over small timescales improves, so will our understanding of olfaction, and human behavior in general.

On Time

Facebook knows what's up; they're making Time Cards for computers and networks so they can perform better, and integrate information better. They understand the mismatch between the size of today's clocks vs the speed of information. 

Today's clocks go down to the picosecond (trillionth of a second), and some things are being measured at the femtoscond (quadrillionth of a second). There's even some zepto-sensitive devices out there. The smaller the time-slice, the better your pattern recognition. Even back as far as Barabási's Bursts (2010) we knew that high resolution temporal patterns could be really informative of network behavior. More people in your study is also helpful, but networks like Facebook are running out of people* (but bots are infinite!), so they have to find another way of mining user data, hence an interest in temporal activity. 

If you think the fine-grained timing of patterns isn't that big a deal, why don't you ask the New York Stock Exchange, which is actually not located on Wall Street, but a few miles away in Carteret, New Jersey, because high frequency trading algorithms compete with each other at such high speeds (the speed of light) that they need the extra light-seconds provided by placing their servers a few miles closer to the international trunk connecting the Internet to the United States (it's called a latency advantage; and that trunk actually lands in Manasquan, a town on the Jersey Shore; still, closer to Carteret than New York).

The point here is that light travels fast, and if we want to understand it, then we will need to look at smaller and smaller slivers of the clock, and add this to the growing body of knowledge about how.

On Olfactory Timing

Think about it -- the brain is a network of billions of blinking neurons. Consciousness, or any neural activity, is an intelligent organization of these blinking neurons into meaningful patterns. But when you have billions of lights blinking on and off at the same time, you need to start slicing up the "time," so it's not all happening "at once." 

In the world of olfactory perception, a new look at the temporal side of things should supplement the intense interest we're seeing in this subject (check out the books Smellosophy or Nose Dive, and stutter at the thought of a $2 million grant in olfactory artifact research).

The author of Smellosophy, Barwich, actually gets into this aspect of olfaction. For her, it's a major distinction in the way we categorize olfactory information - between the historical approach to "mapping" the receptors of the two-dimensional olfactory bulb to "measuring" the timing of receptor activity, which includes the temporal dimension.

She argues that the measurement approach can reveal a key to the coding of the olfactory network. Olfaction is not just about identifying the different molecules present and their combinatorial patterns, but about when they appear and at what concentrations, and how these concentrations change over time. After all, olfaction is about detecting change, which implies time.

She says, "Rather than molecules, your brain depicts transient information patterns" (p246)
-Smellosophy: What the Nose Tells the Mind, A.S. Barwich, 2020, Harvard University Press
http://www.smellosophy.com/

And here's a recent article on the topic:

Fast changing smells can teach mice about space
May 2021, phys.org
Mice can sense extremely fast and subtle changes in the structure of odors and use this to guide their behavior. The findings, published in Nature today, alter the current view on how odors are detected and processed in the mammalian brain.

via The Francis Crick Institute: Ackels, T., Erskine, A., Dasgupta, D. et al. Fast odour dynamics are encoded in the olfactory system and guide behaviour. Nature (2021). doi.org/10.1038/s41586-021-03514-2
Post Script:
Start with the 2014 article from Princeton (cited below) that said Facebook is losing its users like a susceptible-infected-recovered model of disease transmission, and that it would be done by 2018. Granted the article came from the unlikely Department of Mechanical and Aerospace Engineering in Princeton, and despite all the bad press, we can now fast forward to 2018 when Facebook is suspected to be more bots than people. Or is it that the people are acting more bot-like as they co-evolve with engagement algorithms in the artificial arena of natural selection. Bottom line is, social media networks, and all technologies adopted by individuals in a society, will follow a similar disease model. 

Not to mention, the bad press (possibly influenced by the same entity that took a full-page ad in the New York Times btw, and somewhere around 100K in case you were wondering) didn't age well.

But I'll let you be the judge of that:
"Facebook may be a massive drain on our attention that some people get sick of, but that doesn’t mean it actually operates like a virus." -TechCrunch, 2014
Epidemiological modeling of online social network dynamics, John Cannarell, Joshua A. Spechler. Department of Mechanical and Aerospace Engineering, Princeton University. arxiv:1401.4208v1 [cs.SI] 17 Jan 2014. https://arxiv.org/pdf/1401.4208v1.pdf

Post Post Script:
Zeptoseconds - New world record in short time measurement
Oct 2020, phys.org

Thursday, September 23, 2021

Olfacotry Camouflage


Scientists used 'fake news' to stop predators from killing endangered birds—and the result was remarkable
Mar 2021, phys.org

Amazing; you just don't hear about olfactory camouflage much:

Odors emanating from the shorebirds' feathers and eggs attract these scent-hunting mammals, which easily find the nests.

Five weeks before the shorebirds arrived for their breeding season in 2016, we mixed the odors with Vaseline and smeared the concoction on hundreds of rocks over two 1,000-hectare study sites. We did this every three days, for three months.

The predators were initially attracted to the odors. But within days, after realizing the scent would not lead to food, they lost interest and stopped visiting the site.

via: Grant L. Norbury et al. Misinformation tactics protect rare birds from problem predators, Science Advances (2021). DOI: 10.1126/sciadv.abe4164

Image credit: Lord of the Rings, scene where the Black Rider sniffs and misses.