Wednesday, August 31, 2016

The Triune Model

 The Reptile Brain, Illustration by Joe Scordo for the book Hidden Scents.

In my book Hidden Scents, I use the triune model to organize the explanation of a neurological model. The brain is complex, and I am not a scientist.

 Christel Quek via Paul MacLean: The "nose-brain" is everything besides the neocortex - it is both the limbic and the reptile brain. (And for disambiguation purposes - sometimes the term "reptile brain" includes the limbic system.)

However, there are ways of organizing the many inter-related parts of the brain that are technically inaccurate, yet hold explanatory power. The most well-known of these is the 'left-brain' 'right-brain' model. The 'left brain' is logical and the 'right brain' is emotional, one is lines and forms, the other shape and color. One is acute attention and the other a passive radar.

Although this model is useful in describing some of the most basic features of the brain as they relate to neuroanatomy, it only goes so far, and must be altered, edited and rearranged many times to reveal the intricate details of brain function.

Regarding olfaction, the triune brain model is useful in that it juxtaposes the critical characteristics of olfactory neuroanatomy against all other senses - that being its intimacy with the limbic system, the seat of our emotions and unconscious behavior.

Triune Brain

The triune model of the mammalian brain is seen as an oversimplified organizing theme by some in the field of comparative neuroscience*

It continues to hold public interest because of its simplicity. While technically inaccurate as an explanation for brain activity, it remains one of very few approximations of the truth we have to work with:

---the "neocortex" represents that cluster of brain structures involved in advanced cognition, including planning, modeling and simulation;

---the "limbic brain" refers to those brain structures, wherever located, associated with social and nurturing behaviors, mutual reciprocity, and other behaviors and affects that arose during the age of the mammals;

---and the "reptilian brain" refers to those brain structures related to territoriality, ritual behavior and other "reptile" behaviors.

The broad explanatory value makes this approximation very engaging and is a useful level of complexity for high school students to begin engaging with brain research.

*Smith CU., 2010, The triune brain in antiquity: Plato, Aristotle, Erasistratus. Journal of the History of the Neurosciences, 19:1-14. doi:10.1080/09647040802601605

Thursday, August 25, 2016

Getting At the Collapsed Dimension

aka Disembodied Navigation of High Dimensional Reality

Alex Grey – Sacred Mirrors

Oops. I just dropped a clear thumbtack on a white floor. I used to hate that, until I learned the trick. You have to collapse the dimension.

Without hesitation, I drop to the ground, lowering my plane of vision until I’m sideways, eyesight scooting across the surface of the floor, until – there it is; I found my thumbtack.

From above, the visual information of the tack is similar to that of the floor. There’s no edges to be distinguished. Viewing the floor as a one-dimensional line (once your head is level with the floor) instead of a two-dimensional plane (as seen from above) makes the thumbtack plain-as-day. The thumbtack is then the only thing that penetrates into the second dimension – it is the only thing that raises above the floor.

If what you’re trying to do is disentangle the information of the tack from the information of the floor, then reducing the information of one make the remaining asymmetry obvious. A floor is essentially 2-D, and a tack is 3-D. And by “collapsing” your visual field to 1-D (by now seeing it as a line stretching above and below you, with the floor on one side and everything else on the other), the “leftover dimension” of the tack becomes immediately apparent. This is not so much about science-fiction space-folding dimensionality, but about information in general.

In the highly abstract, “disembodied” discipline of information science, dimensionality is not about space, or what I like to call body space. Body space is the three dimensions we are so familiar with, it is a volume and we are in it. But this is only one way of using dimensionality. (The imaginings of Flatland will only get us so far.) In essence, a dimension is a line of potential measurements. It doesn’t have to be about direction like NSEW. It can be in lightness or darkness, as in the one-dimensional measurement of the rods in our eyes. The spectrum of colors is another one-dimensional measurement used for seeing. Together, these two dimensions can generate a third piece of information which is the point in-between the two (and is the difference between pink and navy blue). This is now a measurement of the two measurements. The very idea of dimensionality is for creating this information space as a way for recognizing and manipulating complex patterns in our environment.

Color uses three dimensions, which are lightness/darkness, red-blue (rainbow spectrum), and brightness/dullness (similar to but not the same as light/dark). So “color” is (according to art academia) a piece of 3-D information; it is a measure of the three measurements. And vision in general is not dependent on color alone, but other things, particularly spatial positioning. But let’s stick to color.

What happens when we compare color to smell? Things get different. Smells only have two dimensions, “good” and “bad”. Then again, and just as valid, smells can have infinite dimensions. Recognizing a smell as “good” or “bad” is the highly subjective alternative, and shows more about culture and the individual than it does the smell itself. There must be another way.

To date, there is no comprehensive, universal organizing principle for odors. Some attempts use a multidimensional odorspace, such as Henning’s prism. Theoretically, any smell can be categorized, or identified, as a point in a prism. The vertices of the prism are Flowery, Foul, Fruity, Spicy, Burnt, and Resinous. Something called “Fruity” will sit right in the corner of the prism. Other odors can be in-between two or more odors, so that “Citrus” might be somewhere between Fruity and Flowery (and just a bit towards Foul?).  This is how odors start to drift away from the vertices, and then the edges, and then from the planes themselves into the middle space of the prism. Again, each point, or odor, is now in-between all the odor-classifying vertices. Technically this is a hybrid model, because there are 6 primaries, but the corresponding odor-points are represented inside a 3-D form. (Please note that the original Henning’s odor prism was meant to deliver information on the edges and between the vertices; the smells were not meant to be read inside the prism but on its surface, making it only a 3- or 4-dimensional odor space.) [By the way, check out this Cabinet article on other ways of organizing scents.]

Hennings Odor Prism

There is another way one can imagine a wheel with 6 spokes, where the points at the end of the spokes are primary odors (popcorn, mint, lemon, etc.).  Every identifiable smell looks like a misshapen, spiderwebbed splat reaching further outwards towards the smells it is similar to, and sinking down to the middle where it isn’t. This is a hybrid also, for the final piece of information is a 2-D shape, based on 6 dimensions.

Let us try one more. Imagine a ball with spikes sticking out in hundreds of directions, each one a primary odor. Any given odor is understood as being “between” these hundreds of primaries, as a measurement of the hundreds of measurements.

Here’s an example of a Radar Chart.

This is the idea of a high dimensional information space. It’s a pretty alien idea to us body-users, but not to the algorithms that run our lives, and apparently not to the nose on our face. Thinking about how we categorize smells is perhaps a step in the direction of making ourselves more at home in high-d reality.

Wednesday, August 24, 2016

People Smell

I am not a mouthbreather

Smell is a symbol for confusion and fantasy, from the time when the world was different. Yet it remains among us in a way that has not changed much in hundreds of years, at least not for the general population.

The general population doesn’t interact with smell in the same way as the previous entities. In brief, if the most concise form of knowledge aggregation (science) struggles with smells as distinctive from organic chemistry, one can be certain that the average person will be no less fortunate. Smell, in the everyday sense, does not lend itself to contemplations, and tends to avoid language altogether. People do not talk about smells themselves – they may talk about what they mean, or where they came from, but they do not talk about the smell itself.

People do not converse about the constituents of an odor – in the vast majority of cases, a smell would be referred to as a source, not as a molecular profile of potentially interchangeable parts. Beyond that, people do not notice most smells in the first place. Olfaction operates automatically, actively de-noising and investigating every one of the myriad odorants surrounding our fluid vaporsphere, and yet, we rarely notice.

Furthermore, the process of olfaction as a perception is so riddled with byzantine circuits that it leaves itself highly susceptible to cognitive override and subjective distortion. By the time a smell makes it to a person’s conscious awareness, it is far-removed from its universal essence (although some might suggest that such a thing doesn’t exist in the first place).

Smell is, after all, the animal inside us, and so being, spends very little time in the realm of collective discourse (and in spite of enervating our emotional lives to the maximum). It is somewhat of an overstatement, but not by much, if at all, to say that the general population barely knows what smell is, or how it works, and any discussion on the topic proper shifts immediately to its effects on memory and emotions, leaving the objective analysis of odorants a fallow field.

[x] In fact, Don Wilson and Richard Stevenson in their book, Learning to Smell (2006), make the case for human olfaction as a purely learned phenomenon, thus maintaining that universality in smells cannot exist.

Friday, August 19, 2016

Adventure Scents

screenshot from 

Ah, the rewards of a random internet perusal. I found here a place where you can buy smells to enhance your game-playing. This unique business was founded by a former special education teacher who was looking to add a multisensory experience to game playing. It's hard to find scents for Dragon's Lair and Abandoned Castle on the market, so she started hand-crafting her own. Now players can enhance their board games, RPGs and LARPs.

I’ll take this directly from the site:

“The Adventure Scents line contains funky fragrances based on common adventure locations. We've got your fandom covered with themes from fantasy, science fiction, historical, modern action, and horror.”

You can order Dusty Attic, Ancient Library, Desert Caravan, or Fairgrounds: “Behind the cheerful façade of kettle corn, cotton candy and funnel cakes, lies a dark and twisted secret.”

What I find most interesting is the way they allow you to search for your product – they offer you a handful of "search strategies" such as by adventure location or by a costume. These are useful ways of categorizing smells, but you would never see this in any other market.

And then, they list other ways to enhance your games, including Jelly Belly. It's funny to see, in a world where virtual reality is no longer around the corner but right in front of our face, that the ways to heighten our fantasy experiences are multifaceted and multisensory, not to mention – pretty low-tech, considering.

Wednesday, August 17, 2016

In The Age of Approximation

Please note that this article is one year old…and that in the meantime Microsoft was sued for the way they rolled out their new omniscience machine

Lots of talk about Windows 10; I like this article from Ars Technica about the necessity for (what may seem like) excessive data collection integrated into the platform. But therein, we see the changing face of computing as one leveraging approximation over precision (a la Olfaction).

Olfaction as a model for the future of software development? How about the present.

Siri needs to know the names of your contacts to be able to set up calls or send messages. Cortana needs to know when and where your appointments are to tell you when you need to leave the home or office to get to them.

But there's a deeper reason: the software powering these capabilities is fundamentally heuristic, using approximation and guesswork to generate its results. Traditionally this wasn't the case; a hardware keyboard with no autocompletion doesn't need any fancy heuristics, it just needs to directly map key presses to characters. But speech recognition, software keyboards of all kinds, and handwriting recognition don't have this precision. The software driving these things has to construct and evaluate a range of different possible interpretations and then pick a most likely option among those interpretations.

This is the way of olfaction. Within the impossibly complex chemosphere, the nose-brain must approximate in order to make sense. The ultimate need for flexibility is reflected in the design of the receptor patch that receives vaporous molecules - some receptors code for specific odor molecules, some for many, and some for nothing at all. Even at the outset, olfaction is a game of ambiguity.

Wednesday, August 10, 2016

Smelling Thoughts

“The act of smelling a thing, anything, is remarkably like the act of thinking itself.”
–Lewis Thomas, Late Night Thoughts on Listening to Mahler’s Ninth Symphony

Another late summer night, I am staring into my computer screen, half-conscious. I sniff my hands again; though it is not me that samples the chemical composition of the air, but something more primitive inside of me. My hands smell like rosemary. The bush out front is fragrant at this time of year and I occasionally grab at the leaves as I walk past. I sniff my hands again, deeper. It smells so good I want more of it, I rub my hands together to heat them up, to release more of the fragrance, and I cover my face like a catnip-crazed cat. A form of virtual ingestion – I gorge myself.

Another late summer’s night, I am again vacantly staring into my computer screen. My chin rests in my hand. Unsolicited, whispering into my perceptive field, a thought surfaces. It is not so much a thought as an alert: Mothballs. Why do I smell mothballs? My fingers, it’s on my fingers. But when –? Think – what was I doing? No, there were no mothballs, there’s no way. It must be something else. But what?

My mind performs a wonder of calculations. My memory, twitching, morphs and re-forms. A monstrous machine simultaneously scans autobiographical spatio-temporal datamaps while performing some backwards kind of virtual organic chemistry experiment. It matches every episodic memory of the past four hours with every potential combination and degradation of chemicals that can be imagined. In reverse, it matches the smell I perceive with every possible episode in which I could have come in contact with it.

Cleaning product? No, no cleaning today. Where was I today? Mothballs – something about my grandmother? (Scanning: “Grandma”) No, not that. I sniff my hands again. It’s changing – smell is always changing – something medicinal? No, herb-like; what did I eat for dinner? (Scanning: “dinner”, “food”, “refrigerator” – the breadth and depth of olfactive memory-search is awe-inspiring.) No, not that. What could it –?

Rosemary! It’s rosemary! No, how could it be? Rosemary smells like mothballs? How could that be? One more time, I sniff. Attenuation, I can’t smell it anymore but for the slightest trace. It has to be, that’s it. But how…? I search the internet: “Rosemary” – Eucalyptol/Cineole, Pinene, Borneol, Camphene, Limonene. “Mothballs” – Naphthalene, 1,4-dichlorobenzene, Paracide. Nothing overlaps. Let’s try again, a more informal search. Nothing, except that they’re both repellants. But what is the connection, which molecule is it?

It doesn’t work like that. Smell doesn’t work that way, and neither does the internet. You can’t search for smells on the internet. The internet is anosmic.

I spent the remainder of my summer vacation researching this topic. How could it be so evasive, so confused? I set myself on a mission. I was to gather the largest inventory of smell descriptors and molecules and plot them in a self-organizing network graph. One thousand molecules, generating three thousand descriptors from an aroma compound catalog for flavor and fragrance companies. The next time I want to know the relationship between rosemary and mothballs, I can access my own graph.

As one might predict, “mothballs” is not an aroma category, and so my graph did not pass its first test.

Days later, after countless article entries, impervious organic chemistry textbooks, and scant literature, the realization occurs – This is almost impossible. Such a powerful word. To think about the impossible can lead one to obsession; but this is what the imagination was made for – maintaining an extended state of cognitive dissonance long enough to facilitate a new configuration of one’s mindset. 

The Ongee - A Nose-Wise Society

The Adamanese people, along with the Ongee, are from the Bay of Bengal and have a very rich vocabulary for smells.

"There is no language for smells, because it bypasses the language centers of the brain.”

This is one of the basic ideas behind Hidden Scents, and is taken from a recent interview I did. I would like to welcome public commentary of the above quote as an opportunity to mention a very special culture, The Ongee:

public commentary:
“Not true of all cultures, some have a rich language to describe smells.”

The commenter is correct in that there are cultures using a much richer vocabulary for their olfactory experience than the one used in any Western culture.

an excerpt from a concluding section in my book:
In the Bay of Bengal live the Ongee people. A “nose-wise” society, they treat olfaction with as much importance as Western vision. When the West once asked the Ongee for help in making a map of their land, the Ongee man responded: “All the places in space are constantly changing. The creek is never the same; …. Your map tells lies. Places change. Does your map say that?” (Pandya 1991). Smell is like this. Whether through the meanderings of history, or the chimerical configurations of post-modernity, smell is always changing. The Ongee are right; there are no maps, no categories, and no lexicon to show that.
Pandya V (1990). Movement and Space: Andamanese Cartography. American Ethnologist 17:775–797.

For those interested, I should also mention Asifa Majid, who currently does work on this precise subject. She studies the odor lexicon of the The Aslian (Austroasiatic) languages of the Malay Peninsula, Southeast Asia.

Odors are expressible in language, as long as you speak the right language. Asifa Majid and Niclas Burenhult. CognitionNovember 2013.

Post Script:
Currently, there are only 94 native speakers of Ongee, confined to a single settlement in the northeast of Little Andaman island, making it an endangered language.
Lewis, M. Paul, Gary F. Simons, and Charles D. Fennig (eds.). 2015. Ethnologue: Languages of the World, Eighteenth edition. Dallas, Texas: SIL International.

Saturday, August 6, 2016

Revisiting the Motor Cortex

All of the images here represent your somatosensory cortex, the part of your brain that senses and controls the corresponding parts of your body. The goofy image that is forever paired with this concept was illustrated by Dr. Wilder Penfield, and never changed thereafter, until, of course, Joe Scordo did it for Hidden Scents, see below.

The 'Homunculus' illustrated by Joe Scordo for Hidden Scents

Penfield's image is perhaps the most well known in representing the motor cortex. Because of the success of the image, he is confused as having discovered it. It was, in fact, discovered decades earlier.

The Penfield map has changed little in the years since its release. Whether or not scientific diagrams should be updated to withstand current aesthetic considerations is beside the point. The 'Homunculus' illustrated by Joe Scordo for Hidden Scents is another step in the sequence of representation and seriality.

Thursday, August 4, 2016

Multimodal Crosstalk

Make it stop.

Got an article here in Wired where we learn about how we taste, and about how we sense everything, really. The brain doesn’t see sensory information as sense-specific, it processes everything together. In the article, psychologist Charles Spence relates it to signal processing – the brain has to process all the signals to figure out what’s good and what’s not. And what we get back is not raw sensory input; instead all the signals interfere with each other, distorting and reshaping each other. Changing the color of white wine to red makes people “taste” red wine. Changing the lights in a room makes you “hear” differently. Senses even distort themselves. White noise makes other noises seem further away; it distorts the aural space of the listener (see Beckerman below).

I bring up all of this because, why, our sense of smell is the most obvious candidate for proving the multi-modal sensory processing of our brains. Smells can be Sharp, Sour, or Green. In fact, smell is one sense that we do not experience as its own. All of the words we use to describe smells come from other senses; and it can be argued that we only experience smell by proxy of the other senses. Perhaps it is too primitive of a phenomenon to translate to the cognizant, self-reflective human – it is the first sense, after all, and made its appearance on the Tree of Life with the Vertebrates, and hence with brains (the two go together).

Lost in Translation

Brendan Cole, Wired, July 2016

Professor Charles Spence, at Oxford, studies applied cognitive psychology, consumer psychology, sensory marketing, and multisensory perception. And that would make him a man of interest here at Limbic Signal. He also deals a lot with the future of food.

The Sonic Boom: How Sound Transforms the Way We Think, Feel, and Buy. Joel Beckerman. 2014.

Wednesday, August 3, 2016

On Space and Place

Confused John Travolta

The hypothalamus is called the seat of social behavior. The hippocampus is the thing that knows where you are, your internal GPS. Together, these brain parts, along with your olfactory system, of course, influence all of your decisions. Everything we do is dependent on where we are (or where we think we are) and who we’re with (or who we think…). Olfactory data informs these things. Places and the smell of those places are centrally located both in the memory and in the motivating, motility actuating parts of us. The emotions that drive us are informed by the smells of the places we are in.

This new report shows that there may be very specific cells that can tell who’s house you’re in. (The experiment is on male mice and other male’s habitats.) They’re looking to use this to help people with social disorders, like autism, schizophrenia, depression, and social anxiety. Anyway, this just points to the use of olfactory studies to help with larger social problems., Jun 2016

Tuesday, August 2, 2016

Bed Bug Signal

The shed skin of bed bugs time-releases smell compounds that signal to other bed bugs where a good spot to bed down is located. They call it pheromones, but to me, that just makes the whole thing sound even funnier.

July 2016,

Monday, August 1, 2016

Deep AI Making Strides

Inceptionism Iteration

Born from my dual interest in both building systems and neural networks, this post is a bit off-track for Limbic Signal, but not really – we’re looking here at neural networks, the ones mentioned in Hidden Scents. The neural networks used to run Google’s Deep Mind are similar to the workings of the olfactory bulb in the way they use layers of feedback systems to recognize patterns.

Deep Mind is in the news because it cut the electricity bills at one of Google’s buildings by a lot. The thing about these self-learning algorithms, if you will, is that the way they work, or how they work, is really unknown to us. They are using an optimization algorithm to generate their results, which means making microtweaks on hundreds of variables and in realtime. It’s the opposite of a silver bullet approach to energy efficiency (and it's also the way we learn to smell, and why smells mean different things to different people). I’ll let the researchers themselves talk about it; this is from their blog:

20 JULY 2016, Rich Evans, Research Engineer, DeepMind and Jim Gao, Data Centre Engineer, Google

“Each data centre has a unique architecture and environment. A custom-tuned model for one system may not be applicable to another. Therefore, a general intelligence framework is needed to understand the data centre’s interactions.

“We accomplished this by taking the historical data that had already been collected by thousands of sensors within the data centre -- data such as temperatures, power, pump speeds, setpoints, etc. -- and using it to train an ensemble of deep neural networks. Since our objective was to improve data centre energy efficiency, we trained the neural networks on the average future PUE (Power Usage Effectiveness), which is defined as the ratio of the total building energy usage to the IT energy usage. We then trained two additional ensembles of deep neural networks to predict the future temperature and pressure of the data centre over the next hour. The purpose of these predictions is to simulate the recommended actions from the PUE model, to ensure that we do not go beyond any operating constraints.

“Our machine learning system was able to consistently achieve a 40 percent reduction in the amount of energy used for cooling, which equates to a 15 percent reduction in overall PUE overhead after accounting for electrical losses and other non-cooling inefficiencies. It also produced the lowest PUE the site had ever seen.


And furthermore, I’ve taken a piece from another one of their posts:

17TH JUNE 2016, David Silver, Google DeepMind

“However, deep Q-networks are only one way to solve the deep RL problem. We recently introduced an even more practical and effective method based on asynchronous RL. This approach exploits the multithreading capabilities of standard CPUs. The idea is to execute many instances of our agent in parallel, but using a shared model. This provides a viable alternative to experience replay, since parallelisation also diversifies and decorrelates the data. Our asynchronous actor-critic algorithm, A3C, combines a deep Q-network with a deep policy network for selecting actions. It achieves state-of-the-art results, using a fraction of the training time of DQN and a fraction of the resource consumption of Gorila. By building novel approaches to intrinsic motivation andtemporally abstract planning, we have also achieved breakthrough results in the most notoriously challenging Atari games, such as Montezuma’s Revenge.”

Post Script
I can’t talk about Deep Mind without mentioning Deep Dream though: check out what it looks like for a computer to dream, it’s basically a new artform called Inceptionism, and it comes from these neural networks.