Gain control Flashcards
(23 cards)
1
Q
Neurons
A
- Input comes from dendrite and out of axon terminals
- Have limited signalling range
- Limited range of inputs distinguishable
- Saturation = don’t get anymore output no matter how much input is received
- S curve
- ## Steep part = lots of output for little input
2
Q
Gain control intro
A
- As we move around world we are exposed to change
- Change in brightness is too much for neurons
- Neurons unable to change their sensitivity
- Brain can change sensitivity to adapt
- Neurons have to adapt their operating range + sensitivity
- Adjust over time and space
- Stablish baseline then signal changes around the baseline
3
Q
Gain control acts across time
A
- Instantaneous adaptation to env
- How neurons control sensitivity because they have to code things with big ranges
- Way of adjusting to local average
- Every neuron adapts + gains control
4
Q
Layers
A
- Depth profiles through brain
- Different layers have different sizes but the order stays the same
- Maybe if we understand 2 columns of the brain we can understand the whole brain
5
Q
Columns
A
- Cortical column
- Some represent input layers, some putput
- ## Once understand 1 column can understand them all
6
Q
Computations - Microcircuits
A
- Cut through columns + reconstruct entire column
- Replicated
- Human brain project isn’t way to study the brain
7
Q
Using vision to study the brain
A
- Eye is made up of the same stuff as the brain
- Neurons operate the same in the eye and PFC and etc and etc. everythwere in the brain
8
Q
Vision as a model system
A
- Vision is a window into the brain
- Visual stimlui are cheap, easy
- People are good at accessing their visual brain consciously
9
Q
The visual system
A
- Optiv chiasm sorts left side and right side of world - left VF sent to right hemisphere visa versa
- Passes through lateral geniculate neucleus + sends signal to PVc
- Retina –> LGN –> Cortex
10
Q
Retina turns luminance patterns into contrast
A
- Luminance = how much light there is somewhere (photons)
- Contrast = how much light changes across space
- Lumiance to contrast computed with receptive field
11
Q
The Lateral Geniculate Nucleus
A
- Gather signals from right and left visual field
- Only sees contrast
- Contrast is the currency of early visual system (Mante et al 2005)
- Record path from LGN + angle neurons sensitive to contrast
- Neuron doesn’t change when luminance is changed so mustn’t respond to luminance
- When contrast chanegs goes up slightly
- Neurons in LGN only see contrast
- Each LGN receive signals from both eyes but layers only receive signal from one eye
- Cells are tuned for eye origin but not for orientation
12
Q
Cortex
A
- V1 (striate cortex) - central visual field
- Occupies large fraction of brain
- Neurons in cortex have tuning
- Place images on screen + can see if neuron responds to it to understand its preferences
- Hunting for optimal stimulus
- Hubet + Wiesel
- Neuron have tuning for different things in cortex
- Neurons have larger receptive fields - each driven by smaller receptive field
13
Q
Gain control in cortex
A
- Change contrast of stimulus = get big response
- Ask ppts to match stimulus
- Neurons desensitise self becasue they are surrounded by patches of high contrast
- Gain control changes when desensitise neurons
- Bonds, 1989
- As you cahnge env of neuron sensitivity of neuron changes
14
Q
Gain control in visual cortex
A
- Gain control continues through visual cortex
- Neural activity moderated by local activity of group - averaged over space + time
- Neurons sensitive to env + past history
- RF in cortex are larger so normalisation happens over wider regions
15
Q
Retinal gain control
A
- Short range
- Untuned
- Depends on eye - adapt independently
- Shows adaptation hapens in the eye
16
Q
Cortical gain control
A
- Long range
- Have tuning (orientation)
- Complex features (facial expressions, blur, gender)
17
Q
When gain control goes bad
A
- Abnormal gain control observed in epilepsy, autism, SZ, parkinsons as well as normal aging
- Sensitive indicator of neural dysfunction
- Causes some neurological disease
18
Q
SZ + Gain control
A
- Delusions, hallucinations, reduced speech
- Dakin et al (2005)
- Controls match circle to 20% but its actually 40% because neurons have been desensitised
- SZ match correctly –> unable to manover in visual cortex + can’t regulate neural activity
19
Q
Epilepsy + gain control
A
- Porciatti et al (2000)
- Photosensitive epilepsy unable to reg activity leading to seizures
- Some form of epilepsy may be due to poor gain control
- Tsai, Norcia + Wade - Non-photo epilensy suffer from failed gain control when showed flashing screen
20
Q
Normal aging + gain control
A
- Get worse at detectign motion as disks get bigger
- Becayse of long-range gain control
- Tadinm 2003
- Older people are better at identifying the circle martching task because visual system doesn’t dampen down motion as much
- Older subjects able to see motion more because gain control doesn’t dampen down system
21
Q
Parkinsons disease + gain control
A
- Mutation in LRRK2 gene
- If you have it more likely to get PD
- Can take LRRK2 from humans + insert into fruit flies and they get PD
22
Q
Measuring gain control in Parkinsons fruit flies
A
- Place them in measuring tubes + put EEG cap on them
- Record brain
- Get S shaoe response
- If you put it in flies visual system have increased response - failure to gain control = very abnormal vision
- Compare with humans
- In tunisia, 40% of PD carry the gene so can test gain control
- Gain control happens early in life –> later in life neurons wear out leading to reduced sensitivity
23
Q
Summary
A
- All neurons have tuning to things they prefer
- All neurons do gain control to match sensitivity in their env
- Gain control affects visual perception
- Abnormal gain control is a signature of neurological disease in humans + animals
