adaptation Flashcards
what different attributes in the world do we adapt to
light level colour balance orientation of gratings perceived eye gaze body orientation viewpoints blur motion mean colour colour constancy pin-cushion distortions facial identity
what different effects help us understand adaptation
mechanisms of sensitivity
adjustment to light level
types of neural coding
efficient allocation of neural resources
what is the process of adaptation
the relationship between stimulus and response is not fixed, rather the response to a stimulus depends on the level of prior exposure to that stimulus
response often diminishes with extended exposure
encompasses a range of underlying that occur over many timescales in all perceptual systems
when timescales are long it is difficult to separate adaptation from neural learning and other types of plasticity
what general principles of neural coding does adaptation reveal
the way in which a response to a stimulus changes with exposure/ levels of the stimulus helps us to predict how different schemes of neural coding (e.g rate/place) are implemented
what is luminance
a measure of intensity that is weighted by the eyes’ sensitivity to different wavelengths (log cd m^2)
i.e the energy of different wavelengths weighted
how do environmental light levels change
vary over a large range in photopic luminance -6 to 8 log cd m^2
scotopic - rod vision functions alone, range of c. 10^3 (-6 to -3)
mesopic - rod and cone vision function together, range of c. 10^3 (-3 to 1)
photopic - cone vision functions alone, range>10^6 (1 to 8)
stimulus changes by 10 log units
neurons have a limited range of responses (2 log units, factor of 100)
how can we investigate light adaptation in rods
choose an area of the retina dense with rod cells
choose the stimulus wavelength that maximally exploits the difference in rod and cone spectral sensitivity
e.g green light flashing against red background (which suppresses L and M cones preventing them from being sensitive, revealing rod sensitivity to light)
the intensity of the flash is measured at the point where the flash is just visible
record increment threshold (the amount of extra light needed to just see the test light)
how do we plot a single sensitivity curve
plot increment threshold as a function of the intensity of the background to produce a TvI curve
what is the relationship between background intensity and incremental threshold
as the intensity of the background increases, the increment threshold increases
the amount of light needed to detect an increment divided by the background intensity is a constant
k= I/ΔI
what is the gradient of a single sensitivity curve
if weber’s law holds,
rearranging to ΔI = KI and taking the logarithm of both sides
shows logΔI= log I + c = 1
such that the gradient = 1
what is weber’s law
the change in a stimulus that will be just noticeable is a constant ratio of the original stimulus
holds for 4-5 log units but not for extremes of stimulation
what is rod saturation a failure of adaptation
when the background light becomes to intense, the rods are no longer able to signal additional light - this is the point of rod saturation
this demonstrates a failure to manage to the range of external intensities and to maintain the neural system
how is response saturation avoided
spatial and temporal vision changes as mean light level changes
the visual system uses this to deal with saturation
makes highly effective adjustments to the mean light level and to the relative activations of the cone classes
what is spatial frequency
the variation in light intensity (number of cycles of variation per degree of visual angle)
what is temporal frequency
variation in intensity over time (cycles of flicker in a second, Hz)
what do we find from looking at full data sets
sensitivity is the reciprocal of threshold (power of -1)
what is the contrast sensitivity function (CSF)
sensitivity to contrast at different frequencies
the ease with which people are able to detect objects of various sizes and perceive the structural detail, such as texture, of those objects.
how can we measure CSF of visual system
using grating stimuli that vary in spatial frequency
adjust contrast until a threshold is found
change mean luminance to test the effect of change in mean level of sensitivity to contrast at different frequencies
what is the troland
units of the amount of luminance that reaches the retina
what is the troland candela per meter squared
measure of luminance where light energy is weighed by the sensitivity of the eye to different wavelengths
what are the parameters of the CSF curves used
CSF measured at different light levels
measured in the troland candela per meter squared
modified to account for the size of the pupil
what do spatial CSF show about adaptation
systematic change with light level
as we adapt to light with increasing light level, CSF becomes more sensitive to contrast
the shape of the CSF changes so that the particular spatial frequencies we are most sensitive to change with light adaptation/ mean light level
At high light levels, the function is bandpass (peaked shape), at low light levels it is low pass (most sensitive to lower spatial frequencies)
this change in shape indicates a shift from dominance of spaitally opponent processing to simple summation
visual system summing light over increasingly smaller areas
what does the temporal CSF show about light adaptation
similar pattern whereby the system sums light over shorter and shorter periods
analogous to how changing the shutter speed of a camera manages exposure to prevent saturation
how do we adapt to orientation
bias in perception after exposure to a tilted stimulus
vertical bars no longer appear straight
tilt after-effect: adaptation to the orientation of a grating
repulsion after effect - in opposite direction to what we are adapted to
how does coding by place explain orientation adaptation
spatial frequency channels have a peak sensitivity to a particular SF
we have a bank of many overlapping spatially tuned SF
a grating with a specific SF will produce a characteristic distribution of activity across SF channels
when we have adapted to a SF, the sensitivity of these channels is reduced by the amount they have been activated i.e the inverse proportion to the stimulus
subsequent presentations of reference stimuli post adaptation shows reduced channel output from most adapted channels so perceived SF is shifted
what is place coding
coding by the relative activation across the population of SF channels
what is the distance paradox
there must be some difference between the adapting frequency and the test frequency to get an effect
when the effect is symmetrical there is no perceived bias in the spatial frequency that we see
this would be produced when we adapt to a low spatial frequency and then test with low spatial frequency
the effect of adaptation is most noticeable at SFs removed from the adaptation frequency
these effects are common in many perceptual attributes such as perceived eye gaze, body orientation, and viewpoint
what is MONUC
common principle in perception whereby many dimensions in our perceptual experience are coded by multiple overlapping narrowly-tuned, univariant channels
what implicit assumptions do we make about individual channels
narrowly-tuned compared to the range of stimuli in the world
follow the principle of univariance (changes in stimulus affect the magnitude of the response but not form of response)
what are perceptual norms
some stimulus dimensions have special null points
these are norms which are not accounted for by MONUC
motion: a balance between left and right appears stationary
blur: a balance between blurred and too sharp appears focused
colour: a balance between yellow/blue and red/green appears achromatic or white
adaptation can change the norm (lens?)
how do we adapt to blur (Webster, 2011)
pre-adaptation norm: participants set their perceived perfect focus
adapt to a stimulus which is either blurred or sharpened relative to their norm
post adaptation must readjust the stimulus to appear focussed
norm is pulled towards the adaptation stimulus
how does webster (2011) explain perceptual norms with two broadly tuned channels
channel A and B with opposite sensitivity across the stimulus dimensions
cross at the point of the perceptual norms
sensitivity adjustments are inversely proportionate to the amount the two channels have been activated
this has a stronger effect of reducing sensitivity on one of the two channels depending on the position of the stimulus dimension
adaptation shifts the norm towards the adaptation level
what is opponent coding
difference between two channels
neurons explicitly code in terms of the difference between these two channels
adaptation at pre-opponent site changes the balance
shifts the norm towards the adaptation level
adaptation at post-opponent site (contrast) changes sensitivity without shifting the norm
what opponent-like processes are observed for encoding faces (MacLeod & Webster, 2011)
stimulus dimensions have a clear norm
implies that the visual system encodes relative responses
may be explicit opponent-like processes
however there are inconsistencies e.g contrast-adaptation for faces has little effect (what is the norm for faces? symmetry?)
general configure information seems to be adapted but not a low level attribute like SF channels
what are the benefits of adaptation
the visual system is able to integrate light over increasingly small time scales and areas as light intensity increases
adaptation maintains our neurons in an appropriate operating range so they do not saturate from the variation in the environment
natural signals in the environment have peaked distributions of likely stimulus levels
the mean and variance of these distributions can vary widely
must be encoded by neurons with limited dynamic range
adaptation fits neurosn into the dynamic range available
able to adjust to the mean and contrast, matching the variation in the stimulus
enables us to separate colours out in the scene to better identify salient objects
prevents response saturation
maximises the neural resources that can respond to the signals that are most common
summary
adaptation is a ubiquitous characteristic of sensory and perceptual systems
crucial in maintaining efficient coding of the rangs of environmental stimuli
light adaptation is not rescaling - to maintain sensitivity the visual systems changes the spatial areas and temporal durations over which it sums light
sensitivity changes that depend on the spatial and temporal frequencies in the stimulus
patterns of adaptation can provide information about the underlying encoding of a stimulus dimension
different models of encoding make different predictions about the effects of adaptation
A consequence of adaptation is that neural resources are directed to the most prevalent stimuli in the environment.
