Contrast sensitivity

Question 1

What is contrast sensitivity (CS) ?

Answer 1

-is the ability to distinguish differences in luminance (e.g., ‘shades of grey’). Crudely speaking:
-While visual acuity (VA) measures the smallest thing you can see [at 100% contrast]
CS measures the dimmest thing you can see

Question 2

how is CS measured ?

Answer 2

with Pelli-Robson letter charts. Though other tests exist.

Question 3

Why is CS important ?

Answer 3

• because in real life need to be able to see large, faint objects (i.e., CS), as well as fine spatial detail at high contrast (i.e., acuity) e.g car and road signs
• also many conditions affecting eyes and visual pathways can reduce CS, independent of VA. E.g., amblyopia, cerebral visual impairment, cataract, optic neuritis, multiple sclerosis, diabetic macular oedema, glaucoma

Question 4

What can CS also be helpful in ?

Answer 4

for screening for people with Parkinson’s or Alzheimer’s disease

Question 5

What can CS be defined as ?

Answer 5

CS is the smallest difference in luminance between a target and a background that a person can detect reliably (e.g., on 90% of trials)

In a high contrast target- there is a large difference in luminance between target and background
In low contrast target- there is a small difference in luminance between target and background

Question 6

What is CS known as in perimetry/visual field testing ?

Answer 6

Differential Light sensitivity (DLS)

Question 7

Why is CS different to measures such as in VA?

Answer 7

-because unlike measures relating to stimulus size (e.g VA) contrast is a relative measure: which is the difference in luminance between the target and it surrounding background

Question 8

What can CS targets be like ?

Answer 8

can have the exact same target
(e.g same spotted light)- can be high or low contrast in another context but same light
- or can have 2 targets of different luminance , that can produce the same level of contrast stimulus
-

Question 9

Why can CS get confusing- ?

Answer 9

because we can sometimes lapse into talking about a target being ‘high contrast’ without any reference to the background. This is often because, by convention, the background luminance is fixed at a particular value for a particular test (e.g., 10 cd/m2 in perimetry).

Question 10

What is CS ?

Answer 10

luminance difference / luminance average

-depends and varies on the stimulus you are dealing with

Question 11

When is webers contrast used ?

Answer 11

• if a simple stimulus (where you have a large tand background) - you would express In terms of weber contrast.
• Difference between target & background luminance, relative to background
• Suitable for uniform targets (blobs, letters)

Question 12

When is michealson contrast used ?

Answer 12

• if you are dealing with something with black and white stripes - like a grating stimulus
• Difference between Min and Max luminance in image, relative to the mean (difference in luminance of black and white stripes )
• Suitable for two-tone targets (stripes)

EQUATION
Difference between Min and Max luminance in image of black and white stripes / average luminance of black and white stripes

Question 13

What is RMS contrast ?

Answer 13

strange luminance all across the image

• Luminance variability relative to mean luminance
• Suitable for complex images with lots of different luminance levels

Question 14

What does CS vary with ?

Answer 14

spatial frequencies

Question 15

What are the CS in relation to spatial frequencies ? (check slide 9)

Answer 15

different spatial frequencies (large bands in low and narrow bands in high)

• the more sensitive you are the more you are able to see higher up in the image.
• you are more sensitive if you are able to see the black and white stripes further up the Y axis and middle regions than in the far left or right
• more sensitive to middling spatial frequencies than to very low or high spatial frequency

Question 16

What is cs in mathematical terms ?

Answer 16

line which maps a set of inputs (spatial frequency) to a set of outputs (CS)- call it the The Contrast Sensitivity Function (CSF)

Question 17

How is the CSF graph?

Answer 17

• they don’t always plot the whole range- not perfect U shaped- can be like a tail
• The x axis - is spatial frequency – cycles per degree (cpd)

Question 18

What is 1 degree in CSF graph ?

Answer 18

unit of size- width of your small finger nail at arm’s length

Question 19

What is 1 cycle in CSF graph ?

Answer 19

1 complete black bar + 1 complete white bar

-

Question 20

What is the peak of CSF in healthy adults?

Answer 20

2-4 cpd (cycles per degree)

Question 21

What does the CSF contain ?

Answer 21

-contains your VA- it is just one point on the CSF

VA is – the smallest thing you can see at maximum contrast- it is in the bottom right of the CSF- where the CSF approaches the x axis-
if you have really good VA it goes more to the right if not then more to the left
slide 12

Question 22

Why is the CSF an inverted U shape ?

Answer 22

• want to be sensitive to specific frequency

- The envelope of multiple independent spatially-tuned channels

Question 23

What does a channel mean?

Answer 23

neural circuit responsible for encoiding a certain spatial frequency band- can imagine green dots being photoreceptors – spaced out green spots

Question 24

What happens if you have a low spatial frequency stimulus?

Answer 24

then might have one receptive field falling on the white path and one of the black etc etc.

Question 25

What happens if the spacing of your receptive field match the spacing of your stimulus-?

Answer 25

you will see a big signal- big difference between one photoreceptor in a low luminance part and one photoreceptor (spatial receptive field ) in the high luminance path

Question 26

What happens if the spacing of your circuit- doesn’t match the spacing of the stimulus ?

Answer 26

if you have high spatial freq stimulus- all the receptives field are falling on black parts.- low luminance parts of the signal- going to report no contrast change- big uniform black background

Question 27

What happens if there is a case where there is lack of correlation and a little bit of a signal?

Answer 27

this population is more sensitive to low spatial frequency

Question 28

What happens if there is a case where all recpeptive fields are packed closely together ?

Answer 28

more sensitive to high spatial frequency -slide 17-

Question 29

What happens if you have a low spatial freq stimulus in high spatial frequency wher receptive fields are packed closely together ?

Answer 29

they will see no contrast changes- a uniform black background is seen

Question 30

Where does the evidence from this come from for these ^^ ?

Answer 30

1. Adaptation studies: Repeat exposure to a grating of a particular spatial frequency makes you less sensitive at detecting gratings around that spatial frequency, but not others (a ‘notch’)- if you stimulate neuron it has a refractory period- if you over stimulate it it will struggle to make a.p – so if you keep giving stimulis after a while you become selectively less sensitivie to that stimulus.
2. Masking studies: A surrounding ‘masker’ of a particular spatial frequency makes you less sensitive at detecting gratings around that spatial frequency, but not others (a ‘notch’)

Question 31

What does CSF represent and why is that important ?

Answer 31

multiple distinct channels is clinically important, because it implies that different regions of the CSF can be affected, independent of acuity

Question 32

What part of the CSF is measured by the Pelli-Robson chart?

Answer 32

Crude answer: The peak of the CSF
Long answer: depends how close you stand!
~1.5 cpd at 1m
~0.5 cpd at 33cm
~2.6 cpd  at 3m

Question 33

What is the peak of the CSF is in healthy adults?

Answer 33

3 cpd in healthy adults, so at 3 m it gives a reasonable estimate of peak CS

Question 34

What can the CSF be reduced down to ?

Answer 34

one number by computing the Area Under the Curve (AUC).

Though again you are throwing away potentially important information about which specific spatial frequencies are affected

Question 35

What are the different tests for measuring CS?

Answer 35

1. Ohio Contrast Cards (2017)
〈no instruction〉
(clinician judges if infant saw grating)
2. Hiding Heidi (2003)
〈no instruction〉
(clinician judges if infant saw face)
3. Peek Contrast (2019)
“Swipe Up/Down/Left/Right”
4. Vistech Contrast Test (1984)
“Left, Right, Up, or Blank?”
5. SpotChecks (2020)
“Mark each circle with an X”
6. Pelli-Robson Chart (1988)
“Read the letters”
7. qCSF (2015)
“Press the matching letter”
8. F.A.C.T. (1983)
“Left, Right, Up?”

Question 36

What is an interesting thing about these tests?

Answer 36

They do the same thing

All operate in more or less the same way: Measuring contrast thresholds using Method of Limits (re: earlier lecture)

Question 37

What tests measure the whole CSF?

Answer 37

qCSF, Ohio, Vistech, FACT

Question 38

What tests are only intended to measure one threshold (Peak CS), though in principle test could be repeated at different viewing distances?

Answer 38

e.g., Pelli-Robson, Spotcheck, Peek, Heidi

Question 39

What are some other tests to do to measure peripheral CS?

Answer 39

Visual field tests.

Question 40

What do most of the tests test for ?

Answer 40

are specially measuring central (foveal) vision

Question 41

What are some of the new tests that are faster?

Answer 41

Technically some of the newest tests (e.g., qCSF) don’t use Method of Limits, but newer ‘statistical’ methods that are much faster.

Question 42

Which test is most used in measuring CS?

Answer 42

peli robson chart

Question 43

How do you measure using peli robson chart ? in detail ?

Answer 43

The clinical Gold Standard for adults
Typically viewed at 1m or 3m
Contrast decreases every 3 letters
Typically scored at +0.05 per letter (or +0.15 per triplet)
A score of 1.95+ indicates healthy vision. 1.5 or less is consistent with poor vision
As detailed previously, at 3m it is designed to test the part of the CSF corresponding to ‘peak CS’ in healthy adults

Question 44

What are some limitations of the peli Robson chart ?

Answer 44

Requires literacy
Somewhat bulky (re: school screening)
Cannot measure CS at different/specific spatial freqs
Primarily aimed at screening. Perhaps not ideal for monitoring very small changes after time/treatment(?)

Question 45

what adult conditions associated with reduced foveal CS?

Answer 45

-cataract
for various kinds fo cataract - right across the spectrum/spatial frequency spectrum
- CS is much lower

-Age-related Macular Degeneration (AMD)

-Diabetic Macular Oedema (DMO/DME)
Also appears to be sensitive to changes in central retinal thickness following anti-VEGF injections for recalcitrant diabetic macular edema

-Glaucoma
“30 of the eyes had reduced CSF(S) in the absence of marked field or acuity defects.”

-Multiple Sclerosis
Instances of low and/or mid-frequency loss, with acuity spared

-Parkinson’s
“Patients with below average contrast sensitivity function at the time of diagnosis showed higher risk of cognitive decline requiring anti-dementia drugs (adjusted odds ratio = 4.68, p = 0.04) and of visual hallucinations (adjusted odds ratio = 12.54, p = 0.04) than those above average function during the follow-up”

Question 46

What are some childhood conditions associated with reduced (foveal) CS
?

Answer 46

amblyopia
optic neuritis
congenital hypothyroidism
retinitis pigmentosa
cataract
corneal edema
cerebral lesions

Question 47

What does the CSF vary with ?

Answer 47

-Varies with eccentricity
-Varies with age
-Varies with lighting level
-Varies with stimulus size
-Varies with species
Etc.

Question 48

What else can CS be?

Answer 48

not just spatial can be temporal.
for e.g rather than making stripes narrower, we can vary the rate at which we flash the stimulus on-off at
-Just as we are selective tuned to be most sensitive to a particularly spatial range (~4 cpd), we are most sensitive to a particular temporal frequency (~4 Hz)

Question 49

What else can the CSF measure ?

Answer 49

• The spatial CSF also doesn’t have to involve differences in luminance.
• Can also measure sensitivity to colour contrast