There is no conception in man's mind
which hath not at first, totally or in parts,
been begotten upon by the organs of sense.
---Thomas Hobbes, Leviathan, 1651

Thursday, September 3, 2009

out of your mind (and into a computer)

Some of the graduates of our lab go on to do varied things, because sensory science has a number of different aspects to it.  At Penn State, John Hayes studies the genetic basis of food perception, and is knee deep in gene sequencing.  If luck goes his way, he'll be contracting with NASA to work on relieving food boredom for astronauts next year.  Another graduate of our lab, Frank R., works for Symrise using sensory methodology to study flavor compounds.  Jeanine D. earned tenure at Ohio State as their sensory professor before being lured away by Firmenich, where she is a Senior Scientist in R&D.  And my wife, Effie, is a Food Scientist developing new food products at The International Food Network and their resident sensory science expert.

My own research, as I've alluded to from before, involves methods of creating Perceptual Maps.  So what are these?

Well, perceptual maps are a visual way to show relationships between sensory stimuli, which may be flavor samples, products, odors, pictures, or even texture samples.  The maps show the mental process of categorization.  Products that are mentally similar are represented close to one another, and products that are mentally dissimilar are further apart. There are many sensory methods to create these maps, and one I've studies quite a bit is called projective mapping.  Lets look at the apple picture again.

Ignore the arrows, for a moment, and focus on the labeled points.  At the bottom, you see the GS (Granny Smith) and AM (Acey Mac) located close to each other.  In a blind taste test, these apples were represented very similar.  Because they are pretty far away from the Red Delicious (RD) we conclude that these are more dissimilar. 

The arrows represent attributes of the apples and are statistically correlated with  the juices in the general direction of the arrow.  While we may have been able to guess that the Granny Smith was tart, I had no idea that a Pink Lady and Acey Mac were related on that attribute.  The arrows also tell us that the tart apples are negatively associated with the sweet arrow  because they are opposite of the sweet arrow.  Thus, there is a sweet/tart opposing axis.  Along the horizontal, there is also a crisp/mushy opposing axis.  Our conclusion is that texture and sweetness/tartness were the primary sensory dimensions for apples based on this test and these samples! 

For this test, I peeled and cut up the apples and treated all of them with citric acid to prevent browning ("Fruit Fresh" is the commercial name).  They were placed in plastic cups, which were labeled with random three digit codes.  We used red tinted lights in the room to even out any minor color variations in the apples, so that people could focus more on taste and wouldn't use color as criteria in their categorization.   The subjects sat in front of a 60cm square sheet of white paper, and tasted the apples and physically placed them on the paper such that apples that were close to each other were similar, and apples that were further were different.   In essence, each person is building their own personal perceptual map.  Then via the magic of fancy statistics (via R, the best statistics software out there) we combine the maps into an average map for the entire set of subjects.  The attribute arrows are added by asking the subjects to write down words next to the apples that help distinguish that apple.

The uses of these maps are numerous.  Food companies use them to see where their products sits compared to competitors.  They are used in cognitive research to evaluate peoples perception before and after some sort of intervening test condition, such as behavioral therapy treatment, to see if it was effective.  Anthropologists use perceptual maps to see how different cultures view relationships among kin. 

While I did not invent this method, I do hope to come up with new methods of creating perceptual maps.  I've been reading up on Graph Theory and have some ideas already that I'll be testing in the next week or two.


  1. I feel sorry for the Ida Red. Soft, mushy, and bland. Not very descriptively enticing.

    What is your favorite??

  2. I like Pink Lady and Honeycrisps, the latter of which was unavailable when I picked the apples up for the study :(

    I did the test in early August, and I think that some of the apples were new-season apples and some, like the Ida Red, were probably a year old. The Ida Red was terrible.