Search Articles

Please Share

FacebookTwitterDiggDeliciousStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Subscribe to Blog via RSS

Enter your email address:

Delivered by FeedBurner

Sponsors

Search for Events

Sponsors

Most Read Articles

New Research: How Our Brains Categorize Aromas

You smell. I do too. Scientists can’t tell us precisely how though.

The physical structures that allow us to sense aromas are well-known, but the way we parse and categorize olfactory information is much less clear. Our sense of smell isn’t connected to the brain through the same pathways as our visual, auditory and tactile sensors. It’s not even connected to the same part of the brain.

A trio of researchers has just published a paper that may improve our understanding slightly. The study is called Categorical Dimensions of Human Odor Descriptor Space Revealed by Non-Negative Matrix Factorization. It’s authors are Jason B. Castro, Arvind Ramanthan and Chakra S. Chennubhotla—a neuroscientist, a computational scientist and an expert in computational and systems biology.

As you may have gathered from the title, it’s written for researchers and experts not the general interest reader. The article assumes considerable prior knowledge, is replete with specialized terminology and features several pages of text such as “we first examined the structure of H, the matrix of odor weights obtained from NMF (recall that H corresponds to an odor), and defines a point in 10-dimensional descriptor spaced spanned by W.” The paper is freely available online though, so you might read it skim it look at the colored charts.

Here’s the gist of it:
As we pick up an aroma, our brain rapidly identifies a number of attributes for the scent. Unlike our sense of taste which we know to resolve just five flavor dimensions (sour, sweet, salty, bitter and umami), there are many potential attributes for smell. They have not been fully identified, prioritized or explained. And the number of these dimensions for any particular aroma is very large which makes analyzing or categorizing the totality of an aroma extremely complex.

The authors applied mathematical algorithms, used in other areas of science to simplify complex, multi-dimensional data, to olfaction. This enabled them to do two things. First, it proved that this particular analytical approach [non-negative matrix factorization] is effective with respect to analysis of smells. Second, the researchers were able to identify what appear to be the ten principal qualities or categories of aromas, each category linked to a number of molecules that generate such aromatics.

These basic categories each have many incarnations but, for simple identification, might be labeled by their most prominent members: fragrant/floral, woody/resinous, fruity (non-citrus), putrid/decayed, chemical, minty/peppermint, sweet, popcorn, pungent and lemon/citrus. Here’s a table showing those categories and the most important members of each.

If you’ve seen wine aroma wheels or aroma categories used by sommeliers or WSET, then the table will look quite familiar. There are interesting differences though. For example, citrus and non-citrus fruit are clearly delineated but non-citrus such as cherry, pineapple and banana are all in the same category. There is also overlap. Cut grass falls into woody/resinous but also citrus. Some attributes are more general characteristics than analogs to physical objects. For example, floral, citrus, sweet and fruity categories can all result in an aroma being described as “light.”

10 vectors
From Categorical Dimensions of Human Odor Descriptor Space Revealed by Non-Negative Matrix Factorization

Further studies using this mathematical technique may break a lot more ground for industries such as wine and perfume. The current analysis was based on a standard set of aromas used by researchers. It may not have included specific notes of interest to us, like oak and aspects of minerality. And it clearly wasn’t a goal to prioritize wine-centric aromas in any way.

I look forward to follow up efforts. Better understanding how our brains naturally categorize wine-related aromas will help us become better analytical tasters. It will also guide us in communicating aromas to consumers in the way they will most readily grasp.

 

Follow NorCalWine on Twitter for breaking wine news, information on events and more. Become a fan and join the NorCal Wine community on FacebookAlso check out our comprehensive Northern California winery listings. They are very useful for planning a tasting trip or just getting in touch with a winery.

This article is original to NorCalWine.com. Copyright 2013 NorCal Wine. All rights reserved.