Sensory Analysis of Foods
How humans experience their food
© Copyright 2010 Umami Information Center
Olfaction
Image courtesy of www.pacsci.org/public/education/champions/ smell2.html
Olfaction Olfactory bulb
Olfactory membrane
Olfaction
X
Image courtesy of www.umds.ac.uk/physiology/jim/tasteolf.htm
Odor • Requirements for odor – Molecule must be volatile – Molecule must be adsorbed onto the olfactory hair (receptor)
Amoore Odors • • • • • • •
Ethereal Camphoraceous Musky Floral Minty Pungent Putrid
Basic Tastes (APK) • Sour -- from acids (hydrogen ion) • Salt -- metal salts (NaCl) • Sweet -- sugars, sugar alcohols, synthetics • Bitter -- alkaloids (caffeine) • Umami -- savory, meaty tastes
Tongue surface
Image courtesy of www.tcm.hut.fi/~mpaasiva/test/39/lhminen/ kopf/kieli2.html
Circumvallate papilla
Image courtesy of Anne LeMaistre (dpalm2.med.uth.tmc.edu/ edprog/00000661.htm)
Circuvallate papillae • 1 = Circumvallate papillae • 2 = Von Ebner’s glands
Taste buds • Arrow marks the taste bud
Taste buds
Image courtesy of Anne Lemaistre (dpalm2.med.uth.tmc.edu/ edprog/00000661.htm)
Taste buds
Image courtesy of Anne LeMaistre (dpalm2.med.uth.tmc.edu/ edprog/00000661.htm)
Taste buds X
Image courtesy of Tim Jacob (www.cf.ac.uk/uwcc/momed/jacob/ teaching/sensory/taste.html#Anatomy)
Taste Requirements • Molecule must be soluble in water
• Molecule must bind to taste receptor (protein)
Threshold concentrations Salt
0.02 M
Sweet
0.02 M
Sour
0.005 M
Bitter
0.002 M
Flavor potentiators • Monosodium glutamate (MSG) O
O
OH
HO
Glutamic acid
NH2
• Nucleotide monophosphates – IMP, GMP, and XMP Many of these flavor potentiators are manufactured by Ajinomoto, a Japanese company whose name means “the source of flavor”
Flavor potentiators O
H N
N
H H
O
X
N
N
O
X = H, inosine monophosphate, IMP X = NH2, guanosine monophosphate, GMP
O O O
P O-
X = OH, xanthosine monoO phosphate, XMP
Types of Sensory Tests • • • • • •
Difference tests Rank order Rating differences Descriptive analysis Threshold Affective tests Note: See pages 11-13 in your lab manual for more information about this topic.
Panel Type • • • •
Trained -- 3-10 people Semi-trained -- 8-25 people Untrained or consumer -- 100 plus Within these panels it is necessary to consider – Selection criteria (target audience) – Composition (age, sex, etc.)
Material Evaluated • Preparation – Method – Carrier (if used)
• Presentation – Coding -- random 3 digit numbers – Order of serving -- randomization – Sample size – Temperature and method of control
Presentation (cont.) – Sample container and utensils used – Time of day – Special conditions (time interval between samples, mouth rinsing, etc.)
Statistical Design • Type of experiment – Randomized block
– Factorial
Environmental Conditions • Setting (controlled sensory booth, store, State Fair, etc.) • Lighting (color)
Paired Comparison Which sample has more of a particular characteristic? 512
314
The probability of guessing the sample is 0.50.
Paired Comparison Form Judge:
Date:
Circle the sample you prefer
Pair A B C D
Sample No.
Sample No.
Paired Comparison Form Judge: Circle the sample you prefer
Pair A B C D
Date: Sample No. 555 312 498 332
Sample No. 467 778 087 714
Null hypothesis • This is the “no effect” or “no difference” hypothesis which states that there is no difference between the test and control samples • Usually referred to as Ho • The hypothesis that states that there is a difference is called the alternative hypothesis (Ha )
Level of significance • In order to determine significant differences, a level of significance is set up, usually called p, as in p < 0.05 • What this means is that there is a less than 5% probability of getting the experimental results that you got and Ho being true at the same time
Example of null hypothesis and p< 0.05 • Suppose we were doing an experiment in which we were examining the effect of steaming time on the tenderness of broccoli. The broccolis are numbered 317 (steamed 5 minutes) and 512 (steamed 10 minutes). The alternative hypothesis is: Longer steaming time produces a more tender broccoli.
Example of null hypothesis and p < 0.05
• The null hypothesis is: Steaming time has no effect on the tenderness of broccoli • If we have 20 panelists and we are doing a paired comparison we could get any number of responses if we asked them to choose the more tender broccoli
Example of null hypothesis and p< 0.05
317
512
Panel 1
10
10
Panel 2
8
12
Panel 3
6
14
Panel 4
5
15
Panel 5
4
16
Example of null hypothesis and p < 0.05 • If the null hypothesis is really true, we might expect the results from Panel 1 • But if it is not, we might get more selections of 512 than 317. How do we tell what a significant difference is? • How do we reject the null hypothesis?
Example of null hypothesis and p< 0.05
Panel 1
317
512
Ho OK?
10
10
Yes
Example of null hypothesis and p< 0.05 317
512
Ho OK?
Panel 1
10
10
Yes
Panel 2
8
12
Maybe?
Example of null hypothesis and p< 0.05 317
512
Ho OK?
Panel 1
10
10
Yes
Panel 2
8
12
Maybe?
Panel 3
6
14
?
Example of null hypothesis and p< 0.05 317
512
Ho OK?
Panel 1
10
10
Yes
Panel 2
8
12
Maybe?
Panel 3
6
14
?
Panel 4
5
15
Seems unlikely
Example of null hypothesis and p< 0.05 317
512
Ho OK?
Panel 1
10
10
Yes
Panel 2
8
12
Maybe?
Panel 3
6
14
?
Panel 4
5
15
Seems unlikely
Panel 5
4
16
Are you kidding me?
Example of null hypothesis and p < 0.05
• Set levels of significance • Set at p < 0.05 • Examine statistical table for paired comparison tests at a level of p < 0.05 and utilizing 20 panelists (Lecture notes, page 28)
Use statistical tables
Example of null hypothesis and p < 0.05
• We see that we need 15 out of 20 selections of a sample to establish that the two samples are significantly different • When we have 15 selections out of 20 we are saying that there is less than a 5% chance of having that result and have Ho be true at the same time
Example of null hypothesis and p < 0.05 • Note that if we make the statistical test more rigorous by setting p < 0.01, it takes more choices of one sample over another (16 out of 20) to establish significant differences • When we have 16 selections out of 20 we are saying that there is less than a 1% chance of having that result and have Ho be true at the same time
Duo-Trio Which of the samples is the same as the reference sample? 512
Ref.
314
Probability of guessing the right answer is 0.50.
Triangle Test
512
Find the odd sample, or find the two samples that are identical. 711
314
Probability of guessing the right answer is 0.33. Thus, this test has more statistical power than the paired comparison or duo-trio tests.
Triangle Test Form Judge:
Date
Sample No.
Duplicate Samples (indicate with an x)
546 790 243
Two of these samples are identical and the other is different. Please enter all sample numbers and check the duplicate samples in the right-hand column.
Triangle Test Form Judge:
Date
Sample No. 546
Duplicate Samples (indicate with an x) x
790 243
x
Problems with the Triangle Test Suppose the triangle test was presented as shown here.
512
311
771
1
2
3
Further imagine that even though the panelists were told that one sample was different, they were, in fact, all the same!
Middle Sample Bias
in the Triangle Test Test no. 1 and 2 1 3 2 1 3 4 4 2 5 4 6 2 16
1 and 3 5 6 4 6 4 5 30
2 and 3 2 3 2 2 2 3 14
Ranking/Rating • Structured
• Unstructured
Structured Rating
Moderately tough
Extremely tough
Slightly tough
Moderately tender
Slightly tender
Extremely tender
Note that each point on the scale has a word anchor. To use, simply make a mark where you believe the sample falls.
Unstructured Rating
No Acidity
High Acidity
Note here that only the ends of the scale are anchored. Again to use, simply make a make where you think the sample falls.
Consumer Preference • Form styles vary. • The important thing about consumer panels is that you need large numbers of panelists in order to make the statistics work out.
Hedonic Ranking – – – – – – – – –
Essentially a measure of how much the panelist __Like extremely __Like very much likes a sample. Results are X __Like moderately independent of other panelist rankings. __Like slightly __Neither like nor dislike __Dislike slightly __Dislike moderately To use, simply make a mark beside the __Dislike very much statement that you __Dislike extremely
agree with.
Texture Evaluation • Mechanical characteristics • Geometrical characteristics • Other characteristics Szczesniak, A. S. (1963) Classification of textural characteristics. J. Food Sci., 28, 385-389
Mechanical Characteristics Primary parameters
Secondary parameters
Hardness Cohesiveness Brittleness Chewiness Gumminess Viscosity Elasticity Adhesiveness
Popular terms Soft, hard Crunchy Chewy,tough Pasty,gummy Thin,viscous Plastic,elastic Sticky,gooey
Geometrical Characteristics Class
Examples
Particle size, shape
Gritty,grainy, coarse
Particle shape, orientation
Fibrous, cellular, crystalline
Other Characteristics Primary Secondary Popular parameters parameters terms Dry, moist, wet,watery Oily Fat content Oiliness Greasiness Greasy Moisture
