Sensory Encoding Flashcards
Receptive fields, Tuning curves and Response Functions
What is transduction?
Transduction is the conversion of a physical source of energy (light, sound, touch,…) into spikes.
Sensory receptor neurons convert a type of stimulus energy (e.g. photon, sound wave) into an electrical impulse that can be interpreted by the brain.
What is a receptive field?
A neuron has a receptive field for a given input when that input modulates the neuron.
What is a tuning curve?
A tuning-curve shows the average response of a neuron (in spike frequency) for some feature of input (orientation of a bar, e.g.).
Note: the more stimulus dimensions are presented (e.g., color and orientation), the more dimensions can figure in the tuning curve (or tuning ‘surface’).
What is a feature?
in the context of sensory processing
A “feature” is an individual measurable property or characteristic of the input that a neural network extracts to make sense of patterns and information.
What is a topographic map?
A topographic map is a map of something that retains spatial relationships (neighborhoods).
The sensation is topographically organized in the brain, meaning that neural activity appears as a map. For example, the neurons that respond to stimuli on your fingers are grouped and not randomly peppered throughout the brain.
For example, the brain map of the sensory surface is topographical.
A topographic map is to be contrasted with a topological map, where relationships are kept, but the spatial arrangement may not be.
How is somatosensation organized in the brain?
Topographically
Describe the neural pathway involved in sound perception.
Receptors > Spiral ganglion > brainstem and midbrain nuclei > Thalamus > Primary Auditory Cortex
Is a tuning curve always a gaussian (a normal curve)?
NO
Tuning curves often look like Gaussians or Sigmoids, but they can be quite complex and have many possible shapes.
Describe the pathway for most forms of sensory perception?
Receptors > Sensory Ganglia > Brainstem > Thalamus > Cortex
What is a Reichardt detector?
A hypothetical neural circuit postulated to explain how neurons perform motion selectivity.
Give examples of low-level features for which a neuron can be selective.
Edges
Contrast
Orientation
Color
How does a neural network “extract features” from input data?
A neural network extracts features through layers of processing, where each layer identifies and refines certain aspects of the input (like edges, textures, or shapes) that are crucial for tasks like classification or recognition.
What are these types of receptive fields called?
On/Off-center receptive fields
(Retinal Ganglion Cells)
Left - an on-center off-surround receptive field. Increases the firing rate when the light covers the center of the field.
Right - an off-center on-surround receptive field. Increases the firing rate when the light covers the surrounding of the field.
What is a mechanoreceptor?
A sensory cell that responds to mechanical pressure/distortion.
There are four mechanoreceptors in the skin of most mammals:
(1) Meissner corpuscle
(2) Merkel cell
(3) Pacinian corpuscle
(4) Ruffini endings
A tuning curve for retinal disparity has a form of…
a sigmoid (logistic) function
What can you say about neuronal activity according to this inter-spike interval?
Graph from: Kostal, lansky, & Pokora, 2011
The neuronal firing depicted in this histogram appears to be regular, there is one clear peak of preferred interspike interval.
